From 8a409b0118c2d78f84f740f60fe03abda1fe3333 Mon Sep 17 00:00:00 2001 From: Vojtech Pavlik Date: Mon, 5 Sep 2005 00:08:08 -0500 Subject: Input: HID - add more consumer usages Extend mapping of the consumer usage page in hid-input.c to handle more cases appearing on new USB keyboards. Signed-off-by: Vojtech Pavlik Signed-off-by: Dmitry Torokhov --- include/linux/input.h | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'include/linux') diff --git a/include/linux/input.h b/include/linux/input.h index bdc53c6cc962..227a497c0678 100644 --- a/include/linux/input.h +++ b/include/linux/input.h @@ -287,6 +287,8 @@ struct input_absinfo { #define KEY_SCROLLDOWN 178 #define KEY_KPLEFTPAREN 179 #define KEY_KPRIGHTPAREN 180 +#define KEY_NEW 181 +#define KEY_REDO 182 #define KEY_F13 183 #define KEY_F14 184 @@ -333,6 +335,12 @@ struct input_absinfo { #define KEY_KBDILLUMDOWN 229 #define KEY_KBDILLUMUP 230 +#define KEY_SEND 231 +#define KEY_REPLY 232 +#define KEY_FORWARDMAIL 233 +#define KEY_SAVE 234 +#define KEY_DOCUMENTS 235 + #define KEY_UNKNOWN 240 #define BTN_MISC 0x100 -- cgit v1.2.3 From 507d37cf269ebbd1b32bcc435fe577e411f73151 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 9 Sep 2005 02:30:07 -0300 Subject: [CCID] Only call the HC insert_options methods when requested Signed-off-by: Arnaldo Carvalho de Melo --- include/linux/dccp.h | 2 ++ net/dccp/ccids/ccid3.c | 5 ++++- net/dccp/options.c | 11 ++++++++--- 3 files changed, 14 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/dccp.h b/include/linux/dccp.h index 007c290f74d4..5e0af0d08a93 100644 --- a/include/linux/dccp.h +++ b/include/linux/dccp.h @@ -433,6 +433,8 @@ struct dccp_sock { struct ccid *dccps_hc_tx_ccid; struct dccp_options_received dccps_options_received; enum dccp_role dccps_role:2; + __u8 dccps_hc_rx_insert_options:1; + __u8 dccps_hc_tx_insert_options:1; }; static inline struct dccp_sock *dccp_sk(const struct sock *sk) diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index ae0500c79d07..37fd9eb8daaf 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -358,10 +358,12 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, } /* Can we send? if so add options and add to packet history */ - if (rc == 0) + if (rc == 0) { + dp->dccps_hc_tx_insert_options = 1; new_packet->dccphtx_ccval = DCCP_SKB_CB(skb)->dccpd_ccval = hctx->ccid3hctx_last_win_count; + } out: return rc; } @@ -811,6 +813,7 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk) hcrx->ccid3hcrx_pinv = ~0; else hcrx->ccid3hcrx_pinv = 1000000 / hcrx->ccid3hcrx_p; + dp->dccps_hc_rx_insert_options = 1; dccp_send_ack(sk); } diff --git a/net/dccp/options.c b/net/dccp/options.c index 382c5894acb2..7ad2f4266ff9 100644 --- a/net/dccp/options.c +++ b/net/dccp/options.c @@ -505,13 +505,18 @@ void dccp_insert_options(struct sock *sk, struct sk_buff *skb) (dp->dccps_hc_rx_ackpkts->dccpap_buf_ackno != DCCP_MAX_SEQNO + 1)) dccp_insert_option_ack_vector(sk, skb); - if (dp->dccps_timestamp_echo != 0) dccp_insert_option_timestamp_echo(sk, skb); } - ccid_hc_rx_insert_options(dp->dccps_hc_rx_ccid, sk, skb); - ccid_hc_tx_insert_options(dp->dccps_hc_tx_ccid, sk, skb); + if (dp->dccps_hc_rx_insert_options) { + ccid_hc_rx_insert_options(dp->dccps_hc_rx_ccid, sk, skb); + dp->dccps_hc_rx_insert_options = 0; + } + if (dp->dccps_hc_tx_insert_options) { + ccid_hc_tx_insert_options(dp->dccps_hc_tx_ccid, sk, skb); + dp->dccps_hc_tx_insert_options = 0; + } /* XXX: insert other options when appropriate */ -- cgit v1.2.3 From b0e567806d16586629468c824dfb2e71155df7da Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 9 Sep 2005 02:38:35 -0300 Subject: [DCCP] Introduce dccp_timestamp To start the timestamps with 0.0ms, easing the integer maths in the CCIDs, this probably will be reworked to use the to be introduced struct timeval_offset infrastructure out of skb_get_timestamp, etc. Signed-off-by: Arnaldo Carvalho de Melo --- include/linux/dccp.h | 1 + net/dccp/ccids/ccid3.c | 27 +++++++++++----------- net/dccp/ccids/ccid3.h | 2 +- net/dccp/ccids/lib/packet_history.h | 3 ++- net/dccp/dccp.h | 16 ++++--------- net/dccp/input.c | 4 ++-- net/dccp/ipv4.c | 1 + net/dccp/minisocks.c | 1 + net/dccp/options.c | 45 ++++++++++++++++++++++++++++--------- 9 files changed, 61 insertions(+), 39 deletions(-) (limited to 'include/linux') diff --git a/include/linux/dccp.h b/include/linux/dccp.h index 5e0af0d08a93..8bf4bacb5051 100644 --- a/include/linux/dccp.h +++ b/include/linux/dccp.h @@ -432,6 +432,7 @@ struct dccp_sock { struct ccid *dccps_hc_rx_ccid; struct ccid *dccps_hc_tx_ccid; struct dccp_options_received dccps_options_received; + struct timeval dccps_epoch; enum dccp_role dccps_role:2; __u8 dccps_hc_rx_insert_options:1; __u8 dccps_hc_tx_insert_options:1; diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 145aafafe4e2..348e6fb262c3 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -169,7 +169,7 @@ static void ccid3_hc_tx_update_x(struct sock *sk) } else { struct timeval now; - do_gettimeofday(&now); + dccp_timestamp(sk, &now); if (timeval_delta(&now, &hctx->ccid3hctx_t_ld) >= hctx->ccid3hctx_rtt) { hctx->ccid3hctx_x = max_t(u32, min_t(u32, hctx->ccid3hctx_x_recv, @@ -317,7 +317,7 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, dccp_tx_hist_add_entry(&hctx->ccid3hctx_hist, new_packet); } - do_gettimeofday(&now); + dccp_timestamp(sk, &now); switch (hctx->ccid3hctx_state) { case TFRC_SSTATE_NO_SENT: @@ -382,7 +382,7 @@ static void ccid3_hc_tx_packet_sent(struct sock *sk, int more, int len) return; } - do_gettimeofday(&now); + dccp_timestamp(sk, &now); /* check if we have sent a data packet */ if (len > 0) { @@ -461,6 +461,7 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) struct ccid3_hc_tx_sock *hctx = dp->dccps_hc_tx_ccid_private; struct ccid3_options_received *opt_recv; struct dccp_tx_hist_entry *packet; + struct timeval now; unsigned long next_tmout; u32 t_elapsed; u32 pinv; @@ -508,7 +509,8 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb) } /* Update RTT */ - r_sample = timeval_now_delta(&packet->dccphtx_tstamp); + dccp_timestamp(sk, &now); + r_sample = timeval_delta(&now, &packet->dccphtx_tstamp); if (unlikely(r_sample <= t_elapsed)) LIMIT_NETDEBUG(KERN_WARNING "%s: r_sample=%uus, t_elapsed=%uus\n", @@ -774,7 +776,7 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk) ccid3_pr_debug("%s, sk=%p\n", dccp_role(sk), sk); - do_gettimeofday(&now); + dccp_timestamp(sk, &now); switch (hcrx->ccid3hcrx_state) { case TFRC_RSTATE_NO_DATA: @@ -903,10 +905,9 @@ found: if (rtt == 0) rtt = 1; - delta = timeval_now_delta(&hcrx->ccid3hcrx_tstamp_last_feedback); - x_recv = hcrx->ccid3hcrx_bytes_recv * USEC_PER_SEC; - if (likely(delta > 1)) - x_recv /= delta; + dccp_timestamp(sk, &tstamp); + delta = timeval_delta(&tstamp, &hcrx->ccid3hcrx_tstamp_last_feedback); + x_recv = usecs_div(hcrx->ccid3hcrx_bytes_recv, delta); tmp1 = (u64)x_recv * (u64)rtt; do_div(tmp1,10000000); @@ -981,7 +982,7 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) if (opt_recv->dccpor_timestamp_echo == 0) break; p_prev = hcrx->ccid3hcrx_rtt; - do_gettimeofday(&now); + dccp_timestamp(sk, &now); timeval_sub_usecs(&now, opt_recv->dccpor_timestamp_echo * 10); r_sample = timeval_usecs(&now); t_elapsed = opt_recv->dccpor_elapsed_time * 10; @@ -1013,7 +1014,7 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) return; } - packet = dccp_rx_hist_entry_new(ccid3_rx_hist, opt_recv->dccpor_ndp, + packet = dccp_rx_hist_entry_new(ccid3_rx_hist, sk, opt_recv->dccpor_ndp, skb, SLAB_ATOMIC); if (packet == NULL) { ccid3_pr_debug("%s, sk=%p, Not enough mem to add rx packet " @@ -1045,7 +1046,7 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) if (ins != 0) break; - do_gettimeofday(&now); + dccp_timestamp(sk, &now); if (timeval_delta(&now, &hcrx->ccid3hcrx_tstamp_last_ack) >= hcrx->ccid3hcrx_rtt) { hcrx->ccid3hcrx_tstamp_last_ack = now; @@ -1100,7 +1101,7 @@ static int ccid3_hc_rx_init(struct sock *sk) hcrx->ccid3hcrx_state = TFRC_RSTATE_NO_DATA; INIT_LIST_HEAD(&hcrx->ccid3hcrx_hist); INIT_LIST_HEAD(&hcrx->ccid3hcrx_li_hist); - do_gettimeofday(&hcrx->ccid3hcrx_tstamp_last_ack); + dccp_timestamp(sk, &hcrx->ccid3hcrx_tstamp_last_ack); hcrx->ccid3hcrx_tstamp_last_feedback = hcrx->ccid3hcrx_tstamp_last_ack; hcrx->ccid3hcrx_rtt = 5000; /* XXX 5ms for now... */ return 0; diff --git a/net/dccp/ccids/ccid3.h b/net/dccp/ccids/ccid3.h index ee8cbace6630..58be6125b695 100644 --- a/net/dccp/ccids/ccid3.h +++ b/net/dccp/ccids/ccid3.h @@ -115,7 +115,7 @@ struct ccid3_hc_rx_sock { u64 ccid3hcrx_seqno_last_counter:48, ccid3hcrx_state:8, ccid3hcrx_last_counter:4; - unsigned long ccid3hcrx_rtt; + u32 ccid3hcrx_rtt; u32 ccid3hcrx_p; u32 ccid3hcrx_bytes_recv; struct timeval ccid3hcrx_tstamp_last_feedback; diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h index fb90a91aa93d..b375ebdb7dcf 100644 --- a/net/dccp/ccids/lib/packet_history.h +++ b/net/dccp/ccids/lib/packet_history.h @@ -134,6 +134,7 @@ static inline struct dccp_tx_hist_entry * static inline struct dccp_rx_hist_entry * dccp_rx_hist_entry_new(struct dccp_rx_hist *hist, + const struct sock *sk, const u32 ndp, const struct sk_buff *skb, const unsigned int __nocast prio) @@ -148,7 +149,7 @@ static inline struct dccp_rx_hist_entry * entry->dccphrx_ccval = dh->dccph_ccval; entry->dccphrx_type = dh->dccph_type; entry->dccphrx_ndp = ndp; - do_gettimeofday(&(entry->dccphrx_tstamp)); + dccp_timestamp(sk, &entry->dccphrx_tstamp); } return entry; diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 33456c0d5937..95c4630b3b18 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -426,10 +426,13 @@ extern struct dccp_ackpkts * dccp_ackpkts_alloc(unsigned int len, const unsigned int __nocast priority); extern void dccp_ackpkts_free(struct dccp_ackpkts *ap); -extern int dccp_ackpkts_add(struct dccp_ackpkts *ap, u64 ackno, u8 state); +extern int dccp_ackpkts_add(struct dccp_ackpkts *ap, const struct sock *sk, + u64 ackno, u8 state); extern void dccp_ackpkts_check_rcv_ackno(struct dccp_ackpkts *ap, struct sock *sk, u64 ackno); +extern void dccp_timestamp(const struct sock *sk, struct timeval *tv); + static inline suseconds_t timeval_usecs(const struct timeval *tv) { return tv->tv_sec * USEC_PER_SEC + tv->tv_usec; @@ -468,17 +471,6 @@ static inline void timeval_sub_usecs(struct timeval *tv, } } -/* - * Returns the difference in usecs between timeval - * passed in and current time - */ -static inline suseconds_t timeval_now_delta(const struct timeval *tv) -{ - struct timeval now; - do_gettimeofday(&now); - return timeval_delta(&now, tv); -} - #ifdef CONFIG_IP_DCCP_DEBUG extern void dccp_ackvector_print(const u64 ackno, const unsigned char *vector, int len); diff --git a/net/dccp/input.c b/net/dccp/input.c index ef29cef1dafe..c60bc3433f5e 100644 --- a/net/dccp/input.c +++ b/net/dccp/input.c @@ -170,7 +170,7 @@ int dccp_rcv_established(struct sock *sk, struct sk_buff *skb, if (dp->dccps_options.dccpo_send_ack_vector) { struct dccp_ackpkts *ap = dp->dccps_hc_rx_ackpkts; - if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, + if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, sk, DCCP_SKB_CB(skb)->dccpd_seq, DCCP_ACKPKTS_STATE_RECEIVED)) { LIMIT_NETDEBUG(KERN_WARNING "DCCP: acknowledgeable " @@ -498,7 +498,7 @@ int dccp_rcv_state_process(struct sock *sk, struct sk_buff *skb, * DCCP_ACKPKTS_STATE_ECN_MARKED */ if (dp->dccps_options.dccpo_send_ack_vector) { - if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, + if (dccp_ackpkts_add(dp->dccps_hc_rx_ackpkts, sk, DCCP_SKB_CB(skb)->dccpd_seq, DCCP_ACKPKTS_STATE_RECEIVED)) goto discard; diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 3fc75dbee4b8..fee9a8c3777b 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -1243,6 +1243,7 @@ static int dccp_v4_init_sock(struct sock *sk) static int dccp_ctl_socket_init = 1; dccp_options_init(&dp->dccps_options); + do_gettimeofday(&dp->dccps_epoch); if (dp->dccps_options.dccpo_send_ack_vector) { dp->dccps_hc_rx_ackpkts = diff --git a/net/dccp/minisocks.c b/net/dccp/minisocks.c index ce5dff4ac22e..18461bc04cbe 100644 --- a/net/dccp/minisocks.c +++ b/net/dccp/minisocks.c @@ -96,6 +96,7 @@ struct sock *dccp_create_openreq_child(struct sock *sk, newdp->dccps_hc_rx_ackpkts = NULL; newdp->dccps_role = DCCP_ROLE_SERVER; newicsk->icsk_rto = DCCP_TIMEOUT_INIT; + do_gettimeofday(&newdp->dccps_epoch); if (newdp->dccps_options.dccpo_send_ack_vector) { newdp->dccps_hc_rx_ackpkts = diff --git a/net/dccp/options.c b/net/dccp/options.c index 34b230a00875..d4c4242d8dd7 100644 --- a/net/dccp/options.c +++ b/net/dccp/options.c @@ -140,7 +140,7 @@ int dccp_parse_options(struct sock *sk, struct sk_buff *skb) opt_recv->dccpor_timestamp = ntohl(*(u32 *)value); dp->dccps_timestamp_echo = opt_recv->dccpor_timestamp; - do_gettimeofday(&dp->dccps_timestamp_time); + dccp_timestamp(sk, &dp->dccps_timestamp_time); dccp_pr_debug("%sTIMESTAMP=%u, ackno=%llu\n", debug_prefix, opt_recv->dccpor_timestamp, @@ -361,9 +361,13 @@ static void dccp_insert_option_ack_vector(struct sock *sk, struct sk_buff *skb) #endif struct dccp_ackpkts *ap = dp->dccps_hc_rx_ackpkts; int len = ap->dccpap_buf_vector_len + 2; - const u32 elapsed_time = timeval_now_delta(&ap->dccpap_time) / 10; + struct timeval now; + u32 elapsed_time; unsigned char *to, *from; + dccp_timestamp(sk, &now); + elapsed_time = timeval_delta(&now, &ap->dccpap_time) / 10; + if (elapsed_time != 0) dccp_insert_option_elapsed_time(sk, skb, elapsed_time); @@ -428,13 +432,29 @@ static void dccp_insert_option_ack_vector(struct sock *sk, struct sk_buff *skb) (unsigned long long) ap->dccpap_ack_ackno); } +void dccp_timestamp(const struct sock *sk, struct timeval *tv) +{ + const struct dccp_sock *dp = dccp_sk(sk); + + do_gettimeofday(tv); + tv->tv_sec -= dp->dccps_epoch.tv_sec; + tv->tv_usec -= dp->dccps_epoch.tv_usec; + + while (tv->tv_usec < 0) { + tv->tv_sec--; + tv->tv_usec += USEC_PER_SEC; + } +} + +EXPORT_SYMBOL_GPL(dccp_timestamp); + void dccp_insert_option_timestamp(struct sock *sk, struct sk_buff *skb) { struct timeval tv; u32 now; - do_gettimeofday(&tv); - now = (tv.tv_sec * USEC_PER_SEC + tv.tv_usec) / 10; + dccp_timestamp(sk, &tv); + now = timeval_usecs(&tv) / 10; /* yes this will overflow but that is the point as we want a * 10 usec 32 bit timer which mean it wraps every 11.9 hours */ @@ -452,13 +472,17 @@ static void dccp_insert_option_timestamp_echo(struct sock *sk, const char *debug_prefix = dp->dccps_role == DCCP_ROLE_CLIENT ? "CLIENT TX opt: " : "server TX opt: "; #endif + struct timeval now; u32 tstamp_echo; - const u32 elapsed_time = - timeval_now_delta(&dp->dccps_timestamp_time) / 10; - const int elapsed_time_len = dccp_elapsed_time_len(elapsed_time); - const int len = 6 + elapsed_time_len; + u32 elapsed_time; + int len, elapsed_time_len; unsigned char *to; + dccp_timestamp(sk, &now); + elapsed_time = timeval_delta(&now, &dp->dccps_timestamp_time) / 10; + elapsed_time_len = dccp_elapsed_time_len(elapsed_time); + len = 6 + elapsed_time_len; + if (DCCP_SKB_CB(skb)->dccpd_opt_len + len > DCCP_MAX_OPT_LEN) { LIMIT_NETDEBUG(KERN_INFO "DCCP: packet too small to insert " "timestamp echo!\n"); @@ -623,7 +647,8 @@ static inline int dccp_ackpkts_set_buf_head_state(struct dccp_ackpkts *ap, /* * Implements the draft-ietf-dccp-spec-11.txt Appendix A */ -int dccp_ackpkts_add(struct dccp_ackpkts *ap, u64 ackno, u8 state) +int dccp_ackpkts_add(struct dccp_ackpkts *ap, const struct sock *sk, + u64 ackno, u8 state) { /* * Check at the right places if the buffer is full, if it is, tell the @@ -704,7 +729,7 @@ int dccp_ackpkts_add(struct dccp_ackpkts *ap, u64 ackno, u8 state) } ap->dccpap_buf_ackno = ackno; - do_gettimeofday(&ap->dccpap_time); + dccp_timestamp(sk, &ap->dccpap_time); out: dccp_pr_debug(""); dccp_ackpkts_print(ap); -- cgit v1.2.3 From 144a50ea5e1487b8b8e722289b4177713354448a Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Tue, 9 Aug 2005 00:20:10 -0400 Subject: [PATCH] must_check attributes for PCI layer. Self explanatory really. Some newer gcc's print a warning if a function is used and we don't check its result. We do this for a bunch of things in the kernel already, this extends that to the PCI layer. Based on a patch originally from Arjan van de Ven. Signed-off-by: Dave Jones Signed-off-by: Arjan van de Ven Signed-off-by: Greg Kroah-Hartman --- include/linux/pci.h | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/pci.h b/include/linux/pci.h index 6caaba0af469..609499356e07 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -376,32 +376,32 @@ static inline int pci_write_config_dword(struct pci_dev *dev, int where, u32 val return pci_bus_write_config_dword (dev->bus, dev->devfn, where, val); } -int pci_enable_device(struct pci_dev *dev); -int pci_enable_device_bars(struct pci_dev *dev, int mask); +int __must_check pci_enable_device(struct pci_dev *dev); +int __must_check pci_enable_device_bars(struct pci_dev *dev, int mask); void pci_disable_device(struct pci_dev *dev); void pci_set_master(struct pci_dev *dev); #define HAVE_PCI_SET_MWI -int pci_set_mwi(struct pci_dev *dev); +int __must_check pci_set_mwi(struct pci_dev *dev); void pci_clear_mwi(struct pci_dev *dev); void pci_intx(struct pci_dev *dev, int enable); -int pci_set_dma_mask(struct pci_dev *dev, u64 mask); -int pci_set_consistent_dma_mask(struct pci_dev *dev, u64 mask); +int __must_check pci_set_dma_mask(struct pci_dev *dev, u64 mask); +int __must_check pci_set_consistent_dma_mask(struct pci_dev *dev, u64 mask); void pci_update_resource(struct pci_dev *dev, struct resource *res, int resno); int pci_assign_resource(struct pci_dev *dev, int i); void pci_restore_bars(struct pci_dev *dev); /* ROM control related routines */ -void __iomem *pci_map_rom(struct pci_dev *pdev, size_t *size); -void __iomem *pci_map_rom_copy(struct pci_dev *pdev, size_t *size); +void __iomem __must_check *pci_map_rom(struct pci_dev *pdev, size_t *size); +void __iomem __must_check *pci_map_rom_copy(struct pci_dev *pdev, size_t *size); void pci_unmap_rom(struct pci_dev *pdev, void __iomem *rom); void pci_remove_rom(struct pci_dev *pdev); /* Power management related routines */ int pci_save_state(struct pci_dev *dev); int pci_restore_state(struct pci_dev *dev); -int pci_set_power_state(struct pci_dev *dev, pci_power_t state); -pci_power_t pci_choose_state(struct pci_dev *dev, pm_message_t state); -int pci_enable_wake(struct pci_dev *dev, pci_power_t state, int enable); +int __must_check pci_set_power_state(struct pci_dev *dev, pci_power_t state); +pci_power_t __must_check pci_choose_state(struct pci_dev *dev, pm_message_t state); +int __must_check pci_enable_wake(struct pci_dev *dev, pci_power_t state, int enable); /* Helper functions for low-level code (drivers/pci/setup-[bus,res].c) */ void pci_bus_assign_resources(struct pci_bus *bus); -- cgit v1.2.3 From 5e41ff9e0650f327a6c819841fa412da95d57319 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Fri, 9 Sep 2005 13:01:35 -0700 Subject: [PATCH] security: enable atomic inode security labeling The following patch set enables atomic security labeling of newly created inodes by altering the fs code to invoke a new LSM hook to obtain the security attribute to apply to a newly created inode and to set up the incore inode security state during the inode creation transaction. This parallels the existing processing for setting ACLs on newly created inodes. Otherwise, it is possible for new inodes to be accessed by another thread via the dcache prior to complete security setup (presently handled by the post_create/mkdir/... LSM hooks in the VFS) and a newly created inode may be left unlabeled on the disk in the event of a crash. SELinux presently works around the issue by ensuring that the incore inode security label is initialized to a special SID that is inaccessible to unprivileged processes (in accordance with policy), thereby preventing inappropriate access but potentially causing false denials on legitimate accesses. A simple test program demonstrates such false denials on SELinux, and the patch solves the problem. Similar such false denials have been encountered in real applications. This patch defines a new inode_init_security LSM hook to obtain the security attribute to apply to a newly created inode and to set up the incore inode security state for it, and adds a corresponding hook function implementation to SELinux. Signed-off-by: Stephen Smalley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/security.h | 41 +++++++++++++++++++++++++++ security/dummy.c | 7 +++++ security/selinux/hooks.c | 59 +++++++++++++++++++++++++++++++++++++++ security/selinux/include/objsec.h | 1 + 4 files changed, 108 insertions(+) (limited to 'include/linux') diff --git a/include/linux/security.h b/include/linux/security.h index 7aab6ab7c57f..d4f3b7a94ea6 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -250,6 +250,25 @@ struct swap_info_struct; * @inode contains the inode structure. * Deallocate the inode security structure and set @inode->i_security to * NULL. + * @inode_init_security: + * Obtain the security attribute name suffix and value to set on a newly + * created inode and set up the incore security field for the new inode. + * This hook is called by the fs code as part of the inode creation + * transaction and provides for atomic labeling of the inode, unlike + * the post_create/mkdir/... hooks called by the VFS. The hook function + * is expected to allocate the name and value via kmalloc, with the caller + * being responsible for calling kfree after using them. + * If the security module does not use security attributes or does + * not wish to put a security attribute on this particular inode, + * then it should return -EOPNOTSUPP to skip this processing. + * @inode contains the inode structure of the newly created inode. + * @dir contains the inode structure of the parent directory. + * @name will be set to the allocated name suffix (e.g. selinux). + * @value will be set to the allocated attribute value. + * @len will be set to the length of the value. + * Returns 0 if @name and @value have been successfully set, + * -EOPNOTSUPP if no security attribute is needed, or + * -ENOMEM on memory allocation failure. * @inode_create: * Check permission to create a regular file. * @dir contains inode structure of the parent of the new file. @@ -1080,6 +1099,8 @@ struct security_operations { int (*inode_alloc_security) (struct inode *inode); void (*inode_free_security) (struct inode *inode); + int (*inode_init_security) (struct inode *inode, struct inode *dir, + char **name, void **value, size_t *len); int (*inode_create) (struct inode *dir, struct dentry *dentry, int mode); void (*inode_post_create) (struct inode *dir, @@ -1442,6 +1463,17 @@ static inline void security_inode_free (struct inode *inode) return; security_ops->inode_free_security (inode); } + +static inline int security_inode_init_security (struct inode *inode, + struct inode *dir, + char **name, + void **value, + size_t *len) +{ + if (unlikely (IS_PRIVATE (inode))) + return -EOPNOTSUPP; + return security_ops->inode_init_security (inode, dir, name, value, len); +} static inline int security_inode_create (struct inode *dir, struct dentry *dentry, @@ -2171,6 +2203,15 @@ static inline int security_inode_alloc (struct inode *inode) static inline void security_inode_free (struct inode *inode) { } + +static inline int security_inode_init_security (struct inode *inode, + struct inode *dir, + char **name, + void **value, + size_t *len) +{ + return -EOPNOTSUPP; +} static inline int security_inode_create (struct inode *dir, struct dentry *dentry, diff --git a/security/dummy.c b/security/dummy.c index 6ff887586479..e8a00fa80469 100644 --- a/security/dummy.c +++ b/security/dummy.c @@ -258,6 +258,12 @@ static void dummy_inode_free_security (struct inode *inode) return; } +static int dummy_inode_init_security (struct inode *inode, struct inode *dir, + char **name, void **value, size_t *len) +{ + return -EOPNOTSUPP; +} + static int dummy_inode_create (struct inode *inode, struct dentry *dentry, int mask) { @@ -886,6 +892,7 @@ void security_fixup_ops (struct security_operations *ops) set_to_dummy_if_null(ops, sb_post_pivotroot); set_to_dummy_if_null(ops, inode_alloc_security); set_to_dummy_if_null(ops, inode_free_security); + set_to_dummy_if_null(ops, inode_init_security); set_to_dummy_if_null(ops, inode_create); set_to_dummy_if_null(ops, inode_post_create); set_to_dummy_if_null(ops, inode_link); diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 8641f8894b4c..63701fe0e1ad 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1274,6 +1274,7 @@ static int post_create(struct inode *dir, struct inode *inode; struct inode_security_struct *dsec; struct superblock_security_struct *sbsec; + struct inode_security_struct *isec; u32 newsid; char *context; unsigned int len; @@ -1293,6 +1294,11 @@ static int post_create(struct inode *dir, return 0; } + isec = inode->i_security; + + if (isec->security_attr_init) + return 0; + if (tsec->create_sid && sbsec->behavior != SECURITY_FS_USE_MNTPOINT) { newsid = tsec->create_sid; } else { @@ -2018,6 +2024,58 @@ static void selinux_inode_free_security(struct inode *inode) inode_free_security(inode); } +static int selinux_inode_init_security(struct inode *inode, struct inode *dir, + char **name, void **value, + size_t *len) +{ + struct task_security_struct *tsec; + struct inode_security_struct *dsec; + struct superblock_security_struct *sbsec; + struct inode_security_struct *isec; + u32 newsid; + int rc; + char *namep, *context; + + tsec = current->security; + dsec = dir->i_security; + sbsec = dir->i_sb->s_security; + isec = inode->i_security; + + if (tsec->create_sid && sbsec->behavior != SECURITY_FS_USE_MNTPOINT) { + newsid = tsec->create_sid; + } else { + rc = security_transition_sid(tsec->sid, dsec->sid, + inode_mode_to_security_class(inode->i_mode), + &newsid); + if (rc) { + printk(KERN_WARNING "%s: " + "security_transition_sid failed, rc=%d (dev=%s " + "ino=%ld)\n", + __FUNCTION__, + -rc, inode->i_sb->s_id, inode->i_ino); + return rc; + } + } + + inode_security_set_sid(inode, newsid); + + namep = kstrdup(XATTR_SELINUX_SUFFIX, GFP_KERNEL); + if (!namep) + return -ENOMEM; + *name = namep; + + rc = security_sid_to_context(newsid, &context, len); + if (rc) { + kfree(namep); + return rc; + } + *value = context; + + isec->security_attr_init = 1; + + return 0; +} + static int selinux_inode_create(struct inode *dir, struct dentry *dentry, int mask) { return may_create(dir, dentry, SECCLASS_FILE); @@ -4298,6 +4356,7 @@ static struct security_operations selinux_ops = { .inode_alloc_security = selinux_inode_alloc_security, .inode_free_security = selinux_inode_free_security, + .inode_init_security = selinux_inode_init_security, .inode_create = selinux_inode_create, .inode_post_create = selinux_inode_post_create, .inode_link = selinux_inode_link, diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h index 887937c8134a..c515bc0b58a1 100644 --- a/security/selinux/include/objsec.h +++ b/security/selinux/include/objsec.h @@ -46,6 +46,7 @@ struct inode_security_struct { unsigned char initialized; /* initialization flag */ struct semaphore sem; unsigned char inherit; /* inherit SID from parent entry */ + unsigned char security_attr_init; /* security attributes init flag */ }; struct file_security_struct { -- cgit v1.2.3 From a74574aafea3a63add3251047601611111f44562 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Fri, 9 Sep 2005 13:01:44 -0700 Subject: [PATCH] Remove security_inode_post_create/mkdir/symlink/mknod hooks This patch removes the inode_post_create/mkdir/mknod/symlink LSM hooks as they are obsoleted by the new inode_init_security hook that enables atomic inode security labeling. If anyone sees any reason to retain these hooks, please speak now. Also, is anyone using the post_rename/link hooks; if not, those could also be removed. Signed-off-by: Stephen Smalley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/namei.c | 16 ++---- include/linux/security.h | 90 ------------------------------- security/dummy.c | 28 ---------- security/selinux/hooks.c | 111 -------------------------------------- security/selinux/include/objsec.h | 1 - 5 files changed, 4 insertions(+), 242 deletions(-) (limited to 'include/linux') diff --git a/fs/namei.c b/fs/namei.c index 145e852c4bd0..993a65a7d570 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1316,10 +1316,8 @@ int vfs_create(struct inode *dir, struct dentry *dentry, int mode, return error; DQUOT_INIT(dir); error = dir->i_op->create(dir, dentry, mode, nd); - if (!error) { + if (!error) fsnotify_create(dir, dentry->d_name.name); - security_inode_post_create(dir, dentry, mode); - } return error; } @@ -1635,10 +1633,8 @@ int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) DQUOT_INIT(dir); error = dir->i_op->mknod(dir, dentry, mode, dev); - if (!error) { + if (!error) fsnotify_create(dir, dentry->d_name.name); - security_inode_post_mknod(dir, dentry, mode, dev); - } return error; } @@ -1708,10 +1704,8 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) DQUOT_INIT(dir); error = dir->i_op->mkdir(dir, dentry, mode); - if (!error) { + if (!error) fsnotify_mkdir(dir, dentry->d_name.name); - security_inode_post_mkdir(dir,dentry, mode); - } return error; } @@ -1947,10 +1941,8 @@ int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname, i DQUOT_INIT(dir); error = dir->i_op->symlink(dir, dentry, oldname); - if (!error) { + if (!error) fsnotify_create(dir, dentry->d_name.name); - security_inode_post_symlink(dir, dentry, oldname); - } return error; } diff --git a/include/linux/security.h b/include/linux/security.h index d4f3b7a94ea6..875225bf8986 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -275,12 +275,6 @@ struct swap_info_struct; * @dentry contains the dentry structure for the file to be created. * @mode contains the file mode of the file to be created. * Return 0 if permission is granted. - * @inode_post_create: - * Set the security attributes on a newly created regular file. This hook - * is called after a file has been successfully created. - * @dir contains the inode structure of the parent directory of the new file. - * @dentry contains the the dentry structure for the newly created file. - * @mode contains the file mode. * @inode_link: * Check permission before creating a new hard link to a file. * @old_dentry contains the dentry structure for an existing link to the file. @@ -303,13 +297,6 @@ struct swap_info_struct; * @dentry contains the dentry structure of the symbolic link. * @old_name contains the pathname of file. * Return 0 if permission is granted. - * @inode_post_symlink: - * @dir contains the inode structure of the parent directory of the new link. - * @dentry contains the dentry structure of new symbolic link. - * @old_name contains the pathname of file. - * Set security attributes for a newly created symbolic link. Note that - * @dentry->d_inode may be NULL, since the filesystem might not - * instantiate the dentry (e.g. NFS). * @inode_mkdir: * Check permissions to create a new directory in the existing directory * associated with inode strcture @dir. @@ -317,11 +304,6 @@ struct swap_info_struct; * @dentry contains the dentry structure of new directory. * @mode contains the mode of new directory. * Return 0 if permission is granted. - * @inode_post_mkdir: - * Set security attributes on a newly created directory. - * @dir contains the inode structure of parent of the directory to be created. - * @dentry contains the dentry structure of new directory. - * @mode contains the mode of new directory. * @inode_rmdir: * Check the permission to remove a directory. * @dir contains the inode structure of parent of the directory to be removed. @@ -337,13 +319,6 @@ struct swap_info_struct; * @mode contains the mode of the new file. * @dev contains the the device number. * Return 0 if permission is granted. - * @inode_post_mknod: - * Set security attributes on a newly created special file (or socket or - * fifo file created via the mknod system call). - * @dir contains the inode structure of parent of the new node. - * @dentry contains the dentry structure of the new node. - * @mode contains the mode of the new node. - * @dev contains the the device number. * @inode_rename: * Check for permission to rename a file or directory. * @old_dir contains the inode structure for parent of the old link. @@ -1103,8 +1078,6 @@ struct security_operations { char **name, void **value, size_t *len); int (*inode_create) (struct inode *dir, struct dentry *dentry, int mode); - void (*inode_post_create) (struct inode *dir, - struct dentry *dentry, int mode); int (*inode_link) (struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry); void (*inode_post_link) (struct dentry *old_dentry, @@ -1112,17 +1085,10 @@ struct security_operations { int (*inode_unlink) (struct inode *dir, struct dentry *dentry); int (*inode_symlink) (struct inode *dir, struct dentry *dentry, const char *old_name); - void (*inode_post_symlink) (struct inode *dir, - struct dentry *dentry, - const char *old_name); int (*inode_mkdir) (struct inode *dir, struct dentry *dentry, int mode); - void (*inode_post_mkdir) (struct inode *dir, struct dentry *dentry, - int mode); int (*inode_rmdir) (struct inode *dir, struct dentry *dentry); int (*inode_mknod) (struct inode *dir, struct dentry *dentry, int mode, dev_t dev); - void (*inode_post_mknod) (struct inode *dir, struct dentry *dentry, - int mode, dev_t dev); int (*inode_rename) (struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry); void (*inode_post_rename) (struct inode *old_dir, @@ -1484,15 +1450,6 @@ static inline int security_inode_create (struct inode *dir, return security_ops->inode_create (dir, dentry, mode); } -static inline void security_inode_post_create (struct inode *dir, - struct dentry *dentry, - int mode) -{ - if (dentry->d_inode && unlikely (IS_PRIVATE (dentry->d_inode))) - return; - security_ops->inode_post_create (dir, dentry, mode); -} - static inline int security_inode_link (struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry) @@ -1528,15 +1485,6 @@ static inline int security_inode_symlink (struct inode *dir, return security_ops->inode_symlink (dir, dentry, old_name); } -static inline void security_inode_post_symlink (struct inode *dir, - struct dentry *dentry, - const char *old_name) -{ - if (dentry->d_inode && unlikely (IS_PRIVATE (dentry->d_inode))) - return; - security_ops->inode_post_symlink (dir, dentry, old_name); -} - static inline int security_inode_mkdir (struct inode *dir, struct dentry *dentry, int mode) @@ -1546,15 +1494,6 @@ static inline int security_inode_mkdir (struct inode *dir, return security_ops->inode_mkdir (dir, dentry, mode); } -static inline void security_inode_post_mkdir (struct inode *dir, - struct dentry *dentry, - int mode) -{ - if (dentry->d_inode && unlikely (IS_PRIVATE (dentry->d_inode))) - return; - security_ops->inode_post_mkdir (dir, dentry, mode); -} - static inline int security_inode_rmdir (struct inode *dir, struct dentry *dentry) { @@ -1572,15 +1511,6 @@ static inline int security_inode_mknod (struct inode *dir, return security_ops->inode_mknod (dir, dentry, mode, dev); } -static inline void security_inode_post_mknod (struct inode *dir, - struct dentry *dentry, - int mode, dev_t dev) -{ - if (dentry->d_inode && unlikely (IS_PRIVATE (dentry->d_inode))) - return; - security_ops->inode_post_mknod (dir, dentry, mode, dev); -} - static inline int security_inode_rename (struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, @@ -2220,11 +2150,6 @@ static inline int security_inode_create (struct inode *dir, return 0; } -static inline void security_inode_post_create (struct inode *dir, - struct dentry *dentry, - int mode) -{ } - static inline int security_inode_link (struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry) @@ -2250,11 +2175,6 @@ static inline int security_inode_symlink (struct inode *dir, return 0; } -static inline void security_inode_post_symlink (struct inode *dir, - struct dentry *dentry, - const char *old_name) -{ } - static inline int security_inode_mkdir (struct inode *dir, struct dentry *dentry, int mode) @@ -2262,11 +2182,6 @@ static inline int security_inode_mkdir (struct inode *dir, return 0; } -static inline void security_inode_post_mkdir (struct inode *dir, - struct dentry *dentry, - int mode) -{ } - static inline int security_inode_rmdir (struct inode *dir, struct dentry *dentry) { @@ -2280,11 +2195,6 @@ static inline int security_inode_mknod (struct inode *dir, return 0; } -static inline void security_inode_post_mknod (struct inode *dir, - struct dentry *dentry, - int mode, dev_t dev) -{ } - static inline int security_inode_rename (struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, diff --git a/security/dummy.c b/security/dummy.c index e8a00fa80469..5083314e14b1 100644 --- a/security/dummy.c +++ b/security/dummy.c @@ -270,12 +270,6 @@ static int dummy_inode_create (struct inode *inode, struct dentry *dentry, return 0; } -static void dummy_inode_post_create (struct inode *inode, struct dentry *dentry, - int mask) -{ - return; -} - static int dummy_inode_link (struct dentry *old_dentry, struct inode *inode, struct dentry *new_dentry) { @@ -300,24 +294,12 @@ static int dummy_inode_symlink (struct inode *inode, struct dentry *dentry, return 0; } -static void dummy_inode_post_symlink (struct inode *inode, - struct dentry *dentry, const char *name) -{ - return; -} - static int dummy_inode_mkdir (struct inode *inode, struct dentry *dentry, int mask) { return 0; } -static void dummy_inode_post_mkdir (struct inode *inode, struct dentry *dentry, - int mask) -{ - return; -} - static int dummy_inode_rmdir (struct inode *inode, struct dentry *dentry) { return 0; @@ -329,12 +311,6 @@ static int dummy_inode_mknod (struct inode *inode, struct dentry *dentry, return 0; } -static void dummy_inode_post_mknod (struct inode *inode, struct dentry *dentry, - int mode, dev_t dev) -{ - return; -} - static int dummy_inode_rename (struct inode *old_inode, struct dentry *old_dentry, struct inode *new_inode, @@ -894,17 +870,13 @@ void security_fixup_ops (struct security_operations *ops) set_to_dummy_if_null(ops, inode_free_security); set_to_dummy_if_null(ops, inode_init_security); set_to_dummy_if_null(ops, inode_create); - set_to_dummy_if_null(ops, inode_post_create); set_to_dummy_if_null(ops, inode_link); set_to_dummy_if_null(ops, inode_post_link); set_to_dummy_if_null(ops, inode_unlink); set_to_dummy_if_null(ops, inode_symlink); - set_to_dummy_if_null(ops, inode_post_symlink); set_to_dummy_if_null(ops, inode_mkdir); - set_to_dummy_if_null(ops, inode_post_mkdir); set_to_dummy_if_null(ops, inode_rmdir); set_to_dummy_if_null(ops, inode_mknod); - set_to_dummy_if_null(ops, inode_post_mknod); set_to_dummy_if_null(ops, inode_rename); set_to_dummy_if_null(ops, inode_post_rename); set_to_dummy_if_null(ops, inode_readlink); diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 265f33d3af9b..c9c20828be79 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1265,91 +1265,6 @@ static int inode_security_set_sid(struct inode *inode, u32 sid) return 0; } -/* Set the security attributes on a newly created file. */ -static int post_create(struct inode *dir, - struct dentry *dentry) -{ - - struct task_security_struct *tsec; - struct inode *inode; - struct inode_security_struct *dsec; - struct superblock_security_struct *sbsec; - struct inode_security_struct *isec; - u32 newsid; - char *context; - unsigned int len; - int rc; - - tsec = current->security; - dsec = dir->i_security; - sbsec = dir->i_sb->s_security; - - inode = dentry->d_inode; - if (!inode) { - /* Some file system types (e.g. NFS) may not instantiate - a dentry for all create operations (e.g. symlink), - so we have to check to see if the inode is non-NULL. */ - printk(KERN_WARNING "post_create: no inode, dir (dev=%s, " - "ino=%ld)\n", dir->i_sb->s_id, dir->i_ino); - return 0; - } - - isec = inode->i_security; - - if (isec->security_attr_init) - return 0; - - if (tsec->create_sid && sbsec->behavior != SECURITY_FS_USE_MNTPOINT) { - newsid = tsec->create_sid; - } else { - rc = security_transition_sid(tsec->sid, dsec->sid, - inode_mode_to_security_class(inode->i_mode), - &newsid); - if (rc) { - printk(KERN_WARNING "post_create: " - "security_transition_sid failed, rc=%d (dev=%s " - "ino=%ld)\n", - -rc, inode->i_sb->s_id, inode->i_ino); - return rc; - } - } - - rc = inode_security_set_sid(inode, newsid); - if (rc) { - printk(KERN_WARNING "post_create: inode_security_set_sid " - "failed, rc=%d (dev=%s ino=%ld)\n", - -rc, inode->i_sb->s_id, inode->i_ino); - return rc; - } - - if (sbsec->behavior == SECURITY_FS_USE_XATTR && - inode->i_op->setxattr) { - /* Use extended attributes. */ - rc = security_sid_to_context(newsid, &context, &len); - if (rc) { - printk(KERN_WARNING "post_create: sid_to_context " - "failed, rc=%d (dev=%s ino=%ld)\n", - -rc, inode->i_sb->s_id, inode->i_ino); - return rc; - } - down(&inode->i_sem); - rc = inode->i_op->setxattr(dentry, - XATTR_NAME_SELINUX, - context, len, 0); - up(&inode->i_sem); - kfree(context); - if (rc < 0) { - printk(KERN_WARNING "post_create: setxattr failed, " - "rc=%d (dev=%s ino=%ld)\n", - -rc, inode->i_sb->s_id, inode->i_ino); - return rc; - } - } - - return 0; -} - - /* Hook functions begin here. */ static int selinux_ptrace(struct task_struct *parent, struct task_struct *child) @@ -2076,8 +1991,6 @@ static int selinux_inode_init_security(struct inode *inode, struct inode *dir, *len = clen; } - isec->security_attr_init = 1; - return 0; } @@ -2086,11 +1999,6 @@ static int selinux_inode_create(struct inode *dir, struct dentry *dentry, int ma return may_create(dir, dentry, SECCLASS_FILE); } -static void selinux_inode_post_create(struct inode *dir, struct dentry *dentry, int mask) -{ - post_create(dir, dentry); -} - static int selinux_inode_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry) { int rc; @@ -2121,21 +2029,11 @@ static int selinux_inode_symlink(struct inode *dir, struct dentry *dentry, const return may_create(dir, dentry, SECCLASS_LNK_FILE); } -static void selinux_inode_post_symlink(struct inode *dir, struct dentry *dentry, const char *name) -{ - post_create(dir, dentry); -} - static int selinux_inode_mkdir(struct inode *dir, struct dentry *dentry, int mask) { return may_create(dir, dentry, SECCLASS_DIR); } -static void selinux_inode_post_mkdir(struct inode *dir, struct dentry *dentry, int mask) -{ - post_create(dir, dentry); -} - static int selinux_inode_rmdir(struct inode *dir, struct dentry *dentry) { return may_link(dir, dentry, MAY_RMDIR); @@ -2152,11 +2050,6 @@ static int selinux_inode_mknod(struct inode *dir, struct dentry *dentry, int mod return may_create(dir, dentry, inode_mode_to_security_class(mode)); } -static void selinux_inode_post_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) -{ - post_create(dir, dentry); -} - static int selinux_inode_rename(struct inode *old_inode, struct dentry *old_dentry, struct inode *new_inode, struct dentry *new_dentry) { @@ -4363,17 +4256,13 @@ static struct security_operations selinux_ops = { .inode_free_security = selinux_inode_free_security, .inode_init_security = selinux_inode_init_security, .inode_create = selinux_inode_create, - .inode_post_create = selinux_inode_post_create, .inode_link = selinux_inode_link, .inode_post_link = selinux_inode_post_link, .inode_unlink = selinux_inode_unlink, .inode_symlink = selinux_inode_symlink, - .inode_post_symlink = selinux_inode_post_symlink, .inode_mkdir = selinux_inode_mkdir, - .inode_post_mkdir = selinux_inode_post_mkdir, .inode_rmdir = selinux_inode_rmdir, .inode_mknod = selinux_inode_mknod, - .inode_post_mknod = selinux_inode_post_mknod, .inode_rename = selinux_inode_rename, .inode_post_rename = selinux_inode_post_rename, .inode_readlink = selinux_inode_readlink, diff --git a/security/selinux/include/objsec.h b/security/selinux/include/objsec.h index c515bc0b58a1..887937c8134a 100644 --- a/security/selinux/include/objsec.h +++ b/security/selinux/include/objsec.h @@ -46,7 +46,6 @@ struct inode_security_struct { unsigned char initialized; /* initialization flag */ struct semaphore sem; unsigned char inherit; /* inherit SID from parent entry */ - unsigned char security_attr_init; /* security attributes init flag */ }; struct file_security_struct { -- cgit v1.2.3 From e31e14ec356f36b131576be5bc31d8fef7e95483 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Fri, 9 Sep 2005 13:01:45 -0700 Subject: [PATCH] remove the inode_post_link and inode_post_rename LSM hooks This patch removes the inode_post_link and inode_post_rename LSM hooks as they are unused (and likely useless). Signed-off-by: Stephen Smalley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/namei.c | 10 ++-------- include/linux/security.h | 49 ------------------------------------------------ security/dummy.c | 17 ----------------- security/selinux/hooks.c | 13 ------------- 4 files changed, 2 insertions(+), 87 deletions(-) (limited to 'include/linux') diff --git a/fs/namei.c b/fs/namei.c index 993a65a7d570..21d85f1ac839 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2012,10 +2012,8 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de DQUOT_INIT(dir); error = dir->i_op->link(old_dentry, dir, new_dentry); up(&old_dentry->d_inode->i_sem); - if (!error) { + if (!error) fsnotify_create(dir, new_dentry->d_name.name); - security_inode_post_link(old_dentry, dir, new_dentry); - } return error; } @@ -2134,11 +2132,8 @@ static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry, d_rehash(new_dentry); dput(new_dentry); } - if (!error) { + if (!error) d_move(old_dentry,new_dentry); - security_inode_post_rename(old_dir, old_dentry, - new_dir, new_dentry); - } return error; } @@ -2164,7 +2159,6 @@ static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry, /* The following d_move() should become unconditional */ if (!(old_dir->i_sb->s_type->fs_flags & FS_ODD_RENAME)) d_move(old_dentry, new_dentry); - security_inode_post_rename(old_dir, old_dentry, new_dir, new_dentry); } if (target) up(&target->i_sem); diff --git a/include/linux/security.h b/include/linux/security.h index 875225bf8986..55b02e1c73f4 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -281,11 +281,6 @@ struct swap_info_struct; * @dir contains the inode structure of the parent directory of the new link. * @new_dentry contains the dentry structure for the new link. * Return 0 if permission is granted. - * @inode_post_link: - * Set security attributes for a new hard link to a file. - * @old_dentry contains the dentry structure for the existing link. - * @dir contains the inode structure of the parent directory of the new file. - * @new_dentry contains the dentry structure for the new file link. * @inode_unlink: * Check the permission to remove a hard link to a file. * @dir contains the inode structure of parent directory of the file. @@ -326,12 +321,6 @@ struct swap_info_struct; * @new_dir contains the inode structure for parent of the new link. * @new_dentry contains the dentry structure of the new link. * Return 0 if permission is granted. - * @inode_post_rename: - * Set security attributes on a renamed file or directory. - * @old_dir contains the inode structure for parent of the old link. - * @old_dentry contains the dentry structure of the old link. - * @new_dir contains the inode structure for parent of the new link. - * @new_dentry contains the dentry structure of the new link. * @inode_readlink: * Check the permission to read the symbolic link. * @dentry contains the dentry structure for the file link. @@ -1080,8 +1069,6 @@ struct security_operations { struct dentry *dentry, int mode); int (*inode_link) (struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry); - void (*inode_post_link) (struct dentry *old_dentry, - struct inode *dir, struct dentry *new_dentry); int (*inode_unlink) (struct inode *dir, struct dentry *dentry); int (*inode_symlink) (struct inode *dir, struct dentry *dentry, const char *old_name); @@ -1091,10 +1078,6 @@ struct security_operations { int mode, dev_t dev); int (*inode_rename) (struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry); - void (*inode_post_rename) (struct inode *old_dir, - struct dentry *old_dentry, - struct inode *new_dir, - struct dentry *new_dentry); int (*inode_readlink) (struct dentry *dentry); int (*inode_follow_link) (struct dentry *dentry, struct nameidata *nd); int (*inode_permission) (struct inode *inode, int mask, struct nameidata *nd); @@ -1459,15 +1442,6 @@ static inline int security_inode_link (struct dentry *old_dentry, return security_ops->inode_link (old_dentry, dir, new_dentry); } -static inline void security_inode_post_link (struct dentry *old_dentry, - struct inode *dir, - struct dentry *new_dentry) -{ - if (new_dentry->d_inode && unlikely (IS_PRIVATE (new_dentry->d_inode))) - return; - security_ops->inode_post_link (old_dentry, dir, new_dentry); -} - static inline int security_inode_unlink (struct inode *dir, struct dentry *dentry) { @@ -1523,18 +1497,6 @@ static inline int security_inode_rename (struct inode *old_dir, new_dir, new_dentry); } -static inline void security_inode_post_rename (struct inode *old_dir, - struct dentry *old_dentry, - struct inode *new_dir, - struct dentry *new_dentry) -{ - if (unlikely (IS_PRIVATE (old_dentry->d_inode) || - (new_dentry->d_inode && IS_PRIVATE (new_dentry->d_inode)))) - return; - security_ops->inode_post_rename (old_dir, old_dentry, - new_dir, new_dentry); -} - static inline int security_inode_readlink (struct dentry *dentry) { if (unlikely (IS_PRIVATE (dentry->d_inode))) @@ -2157,11 +2119,6 @@ static inline int security_inode_link (struct dentry *old_dentry, return 0; } -static inline void security_inode_post_link (struct dentry *old_dentry, - struct inode *dir, - struct dentry *new_dentry) -{ } - static inline int security_inode_unlink (struct inode *dir, struct dentry *dentry) { @@ -2203,12 +2160,6 @@ static inline int security_inode_rename (struct inode *old_dir, return 0; } -static inline void security_inode_post_rename (struct inode *old_dir, - struct dentry *old_dentry, - struct inode *new_dir, - struct dentry *new_dentry) -{ } - static inline int security_inode_readlink (struct dentry *dentry) { return 0; diff --git a/security/dummy.c b/security/dummy.c index 5083314e14b1..9623a61dfc76 100644 --- a/security/dummy.c +++ b/security/dummy.c @@ -276,13 +276,6 @@ static int dummy_inode_link (struct dentry *old_dentry, struct inode *inode, return 0; } -static void dummy_inode_post_link (struct dentry *old_dentry, - struct inode *inode, - struct dentry *new_dentry) -{ - return; -} - static int dummy_inode_unlink (struct inode *inode, struct dentry *dentry) { return 0; @@ -319,14 +312,6 @@ static int dummy_inode_rename (struct inode *old_inode, return 0; } -static void dummy_inode_post_rename (struct inode *old_inode, - struct dentry *old_dentry, - struct inode *new_inode, - struct dentry *new_dentry) -{ - return; -} - static int dummy_inode_readlink (struct dentry *dentry) { return 0; @@ -871,14 +856,12 @@ void security_fixup_ops (struct security_operations *ops) set_to_dummy_if_null(ops, inode_init_security); set_to_dummy_if_null(ops, inode_create); set_to_dummy_if_null(ops, inode_link); - set_to_dummy_if_null(ops, inode_post_link); set_to_dummy_if_null(ops, inode_unlink); set_to_dummy_if_null(ops, inode_symlink); set_to_dummy_if_null(ops, inode_mkdir); set_to_dummy_if_null(ops, inode_rmdir); set_to_dummy_if_null(ops, inode_mknod); set_to_dummy_if_null(ops, inode_rename); - set_to_dummy_if_null(ops, inode_post_rename); set_to_dummy_if_null(ops, inode_readlink); set_to_dummy_if_null(ops, inode_follow_link); set_to_dummy_if_null(ops, inode_permission); diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index c9c20828be79..3f0b533be92c 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -2009,11 +2009,6 @@ static int selinux_inode_link(struct dentry *old_dentry, struct inode *dir, stru return may_link(dir, old_dentry, MAY_LINK); } -static void selinux_inode_post_link(struct dentry *old_dentry, struct inode *inode, struct dentry *new_dentry) -{ - return; -} - static int selinux_inode_unlink(struct inode *dir, struct dentry *dentry) { int rc; @@ -2056,12 +2051,6 @@ static int selinux_inode_rename(struct inode *old_inode, struct dentry *old_dent return may_rename(old_inode, old_dentry, new_inode, new_dentry); } -static void selinux_inode_post_rename(struct inode *old_inode, struct dentry *old_dentry, - struct inode *new_inode, struct dentry *new_dentry) -{ - return; -} - static int selinux_inode_readlink(struct dentry *dentry) { return dentry_has_perm(current, NULL, dentry, FILE__READ); @@ -4257,14 +4246,12 @@ static struct security_operations selinux_ops = { .inode_init_security = selinux_inode_init_security, .inode_create = selinux_inode_create, .inode_link = selinux_inode_link, - .inode_post_link = selinux_inode_post_link, .inode_unlink = selinux_inode_unlink, .inode_symlink = selinux_inode_symlink, .inode_mkdir = selinux_inode_mkdir, .inode_rmdir = selinux_inode_rmdir, .inode_mknod = selinux_inode_mknod, .inode_rename = selinux_inode_rename, - .inode_post_rename = selinux_inode_post_rename, .inode_readlink = selinux_inode_readlink, .inode_follow_link = selinux_inode_follow_link, .inode_permission = selinux_inode_permission, -- cgit v1.2.3 From 383f2835eb9afb723af71850037b2f074ac9db60 Mon Sep 17 00:00:00 2001 From: "Chen, Kenneth W" Date: Fri, 9 Sep 2005 13:02:02 -0700 Subject: [PATCH] Prefetch kernel stacks to speed up context switch For architecture like ia64, the switch stack structure is fairly large (currently 528 bytes). For context switch intensive application, we found that significant amount of cache misses occurs in switch_to() function. The following patch adds a hook in the schedule() function to prefetch switch stack structure as soon as 'next' task is determined. This allows maximum overlap in prefetch cache lines for that structure. Signed-off-by: Ken Chen Cc: Ingo Molnar Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/ia64/kernel/entry.S | 23 +++++++++++++++++++++++ include/asm-ia64/system.h | 1 + include/linux/sched.h | 5 +++++ kernel/sched.c | 1 + 4 files changed, 30 insertions(+) (limited to 'include/linux') diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index 3c8821024509..915e12791836 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -470,6 +470,29 @@ ENTRY(load_switch_stack) br.cond.sptk.many b7 END(load_switch_stack) +GLOBAL_ENTRY(prefetch_stack) + add r14 = -IA64_SWITCH_STACK_SIZE, sp + add r15 = IA64_TASK_THREAD_KSP_OFFSET, in0 + ;; + ld8 r16 = [r15] // load next's stack pointer + lfetch.fault.excl [r14], 128 + ;; + lfetch.fault.excl [r14], 128 + lfetch.fault [r16], 128 + ;; + lfetch.fault.excl [r14], 128 + lfetch.fault [r16], 128 + ;; + lfetch.fault.excl [r14], 128 + lfetch.fault [r16], 128 + ;; + lfetch.fault.excl [r14], 128 + lfetch.fault [r16], 128 + ;; + lfetch.fault [r16], 128 + br.ret.sptk.many rp +END(prefetch_switch_stack) + GLOBAL_ENTRY(execve) mov r15=__NR_execve // put syscall number in place break __BREAK_SYSCALL diff --git a/include/asm-ia64/system.h b/include/asm-ia64/system.h index 33256db4a7cf..635235fa1e32 100644 --- a/include/asm-ia64/system.h +++ b/include/asm-ia64/system.h @@ -275,6 +275,7 @@ extern void ia64_load_extra (struct task_struct *task); */ #define __ARCH_WANT_UNLOCKED_CTXSW +#define ARCH_HAS_PREFETCH_SWITCH_STACK #define ia64_platform_is(x) (strcmp(x, platform_name) == 0) void cpu_idle_wait(void); diff --git a/include/linux/sched.h b/include/linux/sched.h index ea1b5f32ec5c..c551e6a1447e 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -604,6 +604,11 @@ extern int groups_search(struct group_info *group_info, gid_t grp); #define GROUP_AT(gi, i) \ ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK]) +#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK +extern void prefetch_stack(struct task_struct*); +#else +static inline void prefetch_stack(struct task_struct *t) { } +#endif struct audit_context; /* See audit.c */ struct mempolicy; diff --git a/kernel/sched.c b/kernel/sched.c index 18b95520a2e2..2632b812cf24 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2888,6 +2888,7 @@ switch_tasks: if (next == rq->idle) schedstat_inc(rq, sched_goidle); prefetch(next); + prefetch_stack(next); clear_tsk_need_resched(prev); rcu_qsctr_inc(task_cpu(prev)); -- cgit v1.2.3 From fac92becdaecff64dd91daab0292c5131de92f0d Mon Sep 17 00:00:00 2001 From: Andrew Stribblehill Date: Fri, 9 Sep 2005 13:02:04 -0700 Subject: [PATCH] bfs: fix endianness, signedness; add trivial bugfix * Makes BFS code endianness-clean. * Fixes some signedness warnings. * Fixes a problem in fs/bfs/inode.c:164 where inodes not synced to disk don't get fully marked as clean. Here's how to reproduce it: # mount -o loop -t bfs /bfs.img /mnt # df -i /mnt Filesystem Inodes IUsed IFree IUse% Mounted on /bfs.img 48 1 47 3% /mnt # df -k /mnt Filesystem 1K-blocks Used Available Use% Mounted on /bfs.img 512 5 508 1% /mnt # cp 60k-archive.zip /mnt/mt.zip # df -k /mnt Filesystem 1K-blocks Used Available Use% Mounted on /bfs.img 512 65 447 13% /mnt # df -i /mnt Filesystem Inodes IUsed IFree IUse% Mounted on /bfs.img 48 2 46 5% /mnt # rm /mnt/mt.zip # echo $? 0 [If the unlink happens before the buffers flush, the following happens:] # df -i /mnt Filesystem Inodes IUsed IFree IUse% Mounted on /bfs.img 48 2 46 5% /mnt # df -k /mnt Filesystem 1K-blocks Used Available Use% Mounted on /bfs.img 512 65 447 13% /mnt fs/bfs/bfs.h | 1 Signed-off-by: Andrew Stribblehill Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/bfs/bfs.h | 1 - fs/bfs/dir.c | 25 ++++++------ fs/bfs/file.c | 23 ++++++----- fs/bfs/inode.c | 102 ++++++++++++++++++++++++++----------------------- include/linux/bfs_fs.h | 23 +++++------ 5 files changed, 93 insertions(+), 81 deletions(-) (limited to 'include/linux') diff --git a/fs/bfs/bfs.h b/fs/bfs/bfs.h index 1020dbc88bec..1fbc53f14aba 100644 --- a/fs/bfs/bfs.h +++ b/fs/bfs/bfs.h @@ -20,7 +20,6 @@ struct bfs_sb_info { unsigned long si_lasti; unsigned long * si_imap; struct buffer_head * si_sbh; /* buffer header w/superblock */ - struct bfs_super_block * si_bfs_sb; /* superblock in si_sbh->b_data */ }; /* diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c index 5a1e5ce057ff..e240c335eb23 100644 --- a/fs/bfs/dir.c +++ b/fs/bfs/dir.c @@ -2,6 +2,7 @@ * fs/bfs/dir.c * BFS directory operations. * Copyright (C) 1999,2000 Tigran Aivazian + * Made endianness-clean by Andrew Stribblehill 2005 */ #include @@ -20,9 +21,9 @@ #define dprintf(x...) #endif -static int bfs_add_entry(struct inode * dir, const char * name, int namelen, int ino); +static int bfs_add_entry(struct inode * dir, const unsigned char * name, int namelen, int ino); static struct buffer_head * bfs_find_entry(struct inode * dir, - const char * name, int namelen, struct bfs_dirent ** res_dir); + const unsigned char * name, int namelen, struct bfs_dirent ** res_dir); static int bfs_readdir(struct file * f, void * dirent, filldir_t filldir) { @@ -53,7 +54,7 @@ static int bfs_readdir(struct file * f, void * dirent, filldir_t filldir) de = (struct bfs_dirent *)(bh->b_data + offset); if (de->ino) { int size = strnlen(de->name, BFS_NAMELEN); - if (filldir(dirent, de->name, size, f->f_pos, de->ino, DT_UNKNOWN) < 0) { + if (filldir(dirent, de->name, size, f->f_pos, le16_to_cpu(de->ino), DT_UNKNOWN) < 0) { brelse(bh); unlock_kernel(); return 0; @@ -107,7 +108,7 @@ static int bfs_create(struct inode * dir, struct dentry * dentry, int mode, inode->i_mapping->a_ops = &bfs_aops; inode->i_mode = mode; inode->i_ino = ino; - BFS_I(inode)->i_dsk_ino = ino; + BFS_I(inode)->i_dsk_ino = cpu_to_le16(ino); BFS_I(inode)->i_sblock = 0; BFS_I(inode)->i_eblock = 0; insert_inode_hash(inode); @@ -139,7 +140,7 @@ static struct dentry * bfs_lookup(struct inode * dir, struct dentry * dentry, st lock_kernel(); bh = bfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len, &de); if (bh) { - unsigned long ino = le32_to_cpu(de->ino); + unsigned long ino = (unsigned long)le16_to_cpu(de->ino); brelse(bh); inode = iget(dir->i_sb, ino); if (!inode) { @@ -183,7 +184,7 @@ static int bfs_unlink(struct inode * dir, struct dentry * dentry) inode = dentry->d_inode; lock_kernel(); bh = bfs_find_entry(dir, dentry->d_name.name, dentry->d_name.len, &de); - if (!bh || de->ino != inode->i_ino) + if (!bh || le16_to_cpu(de->ino) != inode->i_ino) goto out_brelse; if (!inode->i_nlink) { @@ -224,7 +225,7 @@ static int bfs_rename(struct inode * old_dir, struct dentry * old_dentry, old_dentry->d_name.name, old_dentry->d_name.len, &old_de); - if (!old_bh || old_de->ino != old_inode->i_ino) + if (!old_bh || le16_to_cpu(old_de->ino) != old_inode->i_ino) goto end_rename; error = -EPERM; @@ -270,7 +271,7 @@ struct inode_operations bfs_dir_inops = { .rename = bfs_rename, }; -static int bfs_add_entry(struct inode * dir, const char * name, int namelen, int ino) +static int bfs_add_entry(struct inode * dir, const unsigned char * name, int namelen, int ino) { struct buffer_head * bh; struct bfs_dirent * de; @@ -304,7 +305,7 @@ static int bfs_add_entry(struct inode * dir, const char * name, int namelen, int } dir->i_mtime = CURRENT_TIME_SEC; mark_inode_dirty(dir); - de->ino = ino; + de->ino = cpu_to_le16((u16)ino); for (i=0; iname[i] = (i < namelen) ? name[i] : 0; mark_buffer_dirty(bh); @@ -317,7 +318,7 @@ static int bfs_add_entry(struct inode * dir, const char * name, int namelen, int return -ENOSPC; } -static inline int bfs_namecmp(int len, const char * name, const char * buffer) +static inline int bfs_namecmp(int len, const unsigned char * name, const char * buffer) { if (len < BFS_NAMELEN && buffer[len]) return 0; @@ -325,7 +326,7 @@ static inline int bfs_namecmp(int len, const char * name, const char * buffer) } static struct buffer_head * bfs_find_entry(struct inode * dir, - const char * name, int namelen, struct bfs_dirent ** res_dir) + const unsigned char * name, int namelen, struct bfs_dirent ** res_dir) { unsigned long block, offset; struct buffer_head * bh; @@ -346,7 +347,7 @@ static struct buffer_head * bfs_find_entry(struct inode * dir, } de = (struct bfs_dirent *)(bh->b_data + offset); offset += BFS_DIRENT_SIZE; - if (de->ino && bfs_namecmp(namelen, name, de->name)) { + if (le16_to_cpu(de->ino) && bfs_namecmp(namelen, name, de->name)) { *res_dir = de; return bh; } diff --git a/fs/bfs/file.c b/fs/bfs/file.c index 747fd1ea55e0..807723b65daf 100644 --- a/fs/bfs/file.c +++ b/fs/bfs/file.c @@ -40,8 +40,8 @@ static int bfs_move_block(unsigned long from, unsigned long to, struct super_blo return 0; } -static int bfs_move_blocks(struct super_block *sb, unsigned long start, unsigned long end, - unsigned long where) +static int bfs_move_blocks(struct super_block *sb, unsigned long start, + unsigned long end, unsigned long where) { unsigned long i; @@ -57,20 +57,21 @@ static int bfs_move_blocks(struct super_block *sb, unsigned long start, unsigned static int bfs_get_block(struct inode * inode, sector_t block, struct buffer_head * bh_result, int create) { - long phys; + unsigned long phys; int err; struct super_block *sb = inode->i_sb; struct bfs_sb_info *info = BFS_SB(sb); struct bfs_inode_info *bi = BFS_I(inode); struct buffer_head *sbh = info->si_sbh; - if (block < 0 || block > info->si_blocks) + if (block > info->si_blocks) return -EIO; phys = bi->i_sblock + block; if (!create) { if (phys <= bi->i_eblock) { - dprintf("c=%d, b=%08lx, phys=%08lx (granted)\n", create, block, phys); + dprintf("c=%d, b=%08lx, phys=%09lx (granted)\n", + create, (unsigned long)block, phys); map_bh(bh_result, sb, phys); } return 0; @@ -80,7 +81,7 @@ static int bfs_get_block(struct inode * inode, sector_t block, of blocks allocated for this file, we can grant it */ if (inode->i_size && phys <= bi->i_eblock) { dprintf("c=%d, b=%08lx, phys=%08lx (interim block granted)\n", - create, block, phys); + create, (unsigned long)block, phys); map_bh(bh_result, sb, phys); return 0; } @@ -88,11 +89,12 @@ static int bfs_get_block(struct inode * inode, sector_t block, /* the rest has to be protected against itself */ lock_kernel(); - /* if the last data block for this file is the last allocated block, we can - extend the file trivially, without moving it anywhere */ + /* if the last data block for this file is the last allocated + block, we can extend the file trivially, without moving it + anywhere */ if (bi->i_eblock == info->si_lf_eblk) { dprintf("c=%d, b=%08lx, phys=%08lx (simple extension)\n", - create, block, phys); + create, (unsigned long)block, phys); map_bh(bh_result, sb, phys); info->si_freeb -= phys - bi->i_eblock; info->si_lf_eblk = bi->i_eblock = phys; @@ -114,7 +116,8 @@ static int bfs_get_block(struct inode * inode, sector_t block, } else err = 0; - dprintf("c=%d, b=%08lx, phys=%08lx (moved)\n", create, block, phys); + dprintf("c=%d, b=%08lx, phys=%08lx (moved)\n", + create, (unsigned long)block, phys); bi->i_sblock = phys; phys += block; info->si_lf_eblk = bi->i_eblock = phys; diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c index 628c2c1a7d7e..c7b39aa279d7 100644 --- a/fs/bfs/inode.c +++ b/fs/bfs/inode.c @@ -3,6 +3,8 @@ * BFS superblock and inode operations. * Copyright (C) 1999,2000 Tigran Aivazian * From fs/minix, Copyright (C) 1991, 1992 Linus Torvalds. + * + * Made endianness-clean by Andrew Stribblehill , 2005. */ #include @@ -54,46 +56,50 @@ static void bfs_read_inode(struct inode * inode) off = (ino - BFS_ROOT_INO) % BFS_INODES_PER_BLOCK; di = (struct bfs_inode *)bh->b_data + off; - inode->i_mode = 0x0000FFFF & di->i_mode; - if (di->i_vtype == BFS_VDIR) { + inode->i_mode = 0x0000FFFF & le32_to_cpu(di->i_mode); + if (le32_to_cpu(di->i_vtype) == BFS_VDIR) { inode->i_mode |= S_IFDIR; inode->i_op = &bfs_dir_inops; inode->i_fop = &bfs_dir_operations; - } else if (di->i_vtype == BFS_VREG) { + } else if (le32_to_cpu(di->i_vtype) == BFS_VREG) { inode->i_mode |= S_IFREG; inode->i_op = &bfs_file_inops; inode->i_fop = &bfs_file_operations; inode->i_mapping->a_ops = &bfs_aops; } - inode->i_uid = di->i_uid; - inode->i_gid = di->i_gid; - inode->i_nlink = di->i_nlink; + BFS_I(inode)->i_sblock = le32_to_cpu(di->i_sblock); + BFS_I(inode)->i_eblock = le32_to_cpu(di->i_eblock); + inode->i_uid = le32_to_cpu(di->i_uid); + inode->i_gid = le32_to_cpu(di->i_gid); + inode->i_nlink = le32_to_cpu(di->i_nlink); inode->i_size = BFS_FILESIZE(di); inode->i_blocks = BFS_FILEBLOCKS(di); + if (inode->i_size || inode->i_blocks) dprintf("Registered inode with %lld size, %ld blocks\n", inode->i_size, inode->i_blocks); inode->i_blksize = PAGE_SIZE; - inode->i_atime.tv_sec = di->i_atime; - inode->i_mtime.tv_sec = di->i_mtime; - inode->i_ctime.tv_sec = di->i_ctime; + inode->i_atime.tv_sec = le32_to_cpu(di->i_atime); + inode->i_mtime.tv_sec = le32_to_cpu(di->i_mtime); + inode->i_ctime.tv_sec = le32_to_cpu(di->i_ctime); inode->i_atime.tv_nsec = 0; inode->i_mtime.tv_nsec = 0; inode->i_ctime.tv_nsec = 0; - BFS_I(inode)->i_dsk_ino = di->i_ino; /* can be 0 so we store a copy */ - BFS_I(inode)->i_sblock = di->i_sblock; - BFS_I(inode)->i_eblock = di->i_eblock; + BFS_I(inode)->i_dsk_ino = le16_to_cpu(di->i_ino); /* can be 0 so we store a copy */ brelse(bh); } static int bfs_write_inode(struct inode * inode, int unused) { - unsigned long ino = inode->i_ino; + unsigned int ino = (u16)inode->i_ino; + unsigned long i_sblock; struct bfs_inode * di; struct buffer_head * bh; int block, off; + dprintf("ino=%08x\n", ino); + if (ino < BFS_ROOT_INO || ino > BFS_SB(inode->i_sb)->si_lasti) { - printf("Bad inode number %s:%08lx\n", inode->i_sb->s_id, ino); + printf("Bad inode number %s:%08x\n", inode->i_sb->s_id, ino); return -EIO; } @@ -101,7 +107,7 @@ static int bfs_write_inode(struct inode * inode, int unused) block = (ino - BFS_ROOT_INO)/BFS_INODES_PER_BLOCK + 1; bh = sb_bread(inode->i_sb, block); if (!bh) { - printf("Unable to read inode %s:%08lx\n", inode->i_sb->s_id, ino); + printf("Unable to read inode %s:%08x\n", inode->i_sb->s_id, ino); unlock_kernel(); return -EIO; } @@ -109,24 +115,26 @@ static int bfs_write_inode(struct inode * inode, int unused) off = (ino - BFS_ROOT_INO)%BFS_INODES_PER_BLOCK; di = (struct bfs_inode *)bh->b_data + off; - if (inode->i_ino == BFS_ROOT_INO) - di->i_vtype = BFS_VDIR; + if (ino == BFS_ROOT_INO) + di->i_vtype = cpu_to_le32(BFS_VDIR); else - di->i_vtype = BFS_VREG; - - di->i_ino = inode->i_ino; - di->i_mode = inode->i_mode; - di->i_uid = inode->i_uid; - di->i_gid = inode->i_gid; - di->i_nlink = inode->i_nlink; - di->i_atime = inode->i_atime.tv_sec; - di->i_mtime = inode->i_mtime.tv_sec; - di->i_ctime = inode->i_ctime.tv_sec; - di->i_sblock = BFS_I(inode)->i_sblock; - di->i_eblock = BFS_I(inode)->i_eblock; - di->i_eoffset = di->i_sblock * BFS_BSIZE + inode->i_size - 1; + di->i_vtype = cpu_to_le32(BFS_VREG); + + di->i_ino = cpu_to_le16(ino); + di->i_mode = cpu_to_le32(inode->i_mode); + di->i_uid = cpu_to_le32(inode->i_uid); + di->i_gid = cpu_to_le32(inode->i_gid); + di->i_nlink = cpu_to_le32(inode->i_nlink); + di->i_atime = cpu_to_le32(inode->i_atime.tv_sec); + di->i_mtime = cpu_to_le32(inode->i_mtime.tv_sec); + di->i_ctime = cpu_to_le32(inode->i_ctime.tv_sec); + i_sblock = BFS_I(inode)->i_sblock; + di->i_sblock = cpu_to_le32(i_sblock); + di->i_eblock = cpu_to_le32(BFS_I(inode)->i_eblock); + di->i_eoffset = cpu_to_le32(i_sblock * BFS_BSIZE + inode->i_size - 1); mark_buffer_dirty(bh); + dprintf("Written ino=%d into %d:%d\n",le16_to_cpu(di->i_ino),block,off); brelse(bh); unlock_kernel(); return 0; @@ -140,13 +148,14 @@ static void bfs_delete_inode(struct inode * inode) int block, off; struct super_block * s = inode->i_sb; struct bfs_sb_info * info = BFS_SB(s); + struct bfs_inode_info * bi = BFS_I(inode); - dprintf("ino=%08lx\n", inode->i_ino); + dprintf("ino=%08lx\n", ino); truncate_inode_pages(&inode->i_data, 0); - if (inode->i_ino < BFS_ROOT_INO || inode->i_ino > info->si_lasti) { - printf("invalid ino=%08lx\n", inode->i_ino); + if (ino < BFS_ROOT_INO || ino > info->si_lasti) { + printf("invalid ino=%08lx\n", ino); return; } @@ -162,13 +171,13 @@ static void bfs_delete_inode(struct inode * inode) return; } off = (ino - BFS_ROOT_INO)%BFS_INODES_PER_BLOCK; - di = (struct bfs_inode *)bh->b_data + off; - if (di->i_ino) { - info->si_freeb += BFS_FILEBLOCKS(di); + di = (struct bfs_inode *) bh->b_data + off; + if (bi->i_dsk_ino) { + info->si_freeb += 1 + bi->i_eblock - bi->i_sblock; info->si_freei++; - clear_bit(di->i_ino, info->si_imap); + clear_bit(ino, info->si_imap); dump_imap("delete_inode", s); - } + } di->i_ino = 0; di->i_sblock = 0; mark_buffer_dirty(bh); @@ -274,14 +283,14 @@ static struct super_operations bfs_sops = { void dump_imap(const char *prefix, struct super_block * s) { -#if 0 +#ifdef DEBUG int i; char *tmpbuf = (char *)get_zeroed_page(GFP_KERNEL); if (!tmpbuf) return; for (i=BFS_SB(s)->si_lasti; i>=0; i--) { - if (i>PAGE_SIZE-100) break; + if (i > PAGE_SIZE-100) break; if (test_bit(i, BFS_SB(s)->si_imap)) strcat(tmpbuf, "1"); else @@ -297,7 +306,7 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) struct buffer_head * bh; struct bfs_super_block * bfs_sb; struct inode * inode; - int i, imap_len; + unsigned i, imap_len; struct bfs_sb_info * info; info = kmalloc(sizeof(*info), GFP_KERNEL); @@ -312,19 +321,18 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) if(!bh) goto out; bfs_sb = (struct bfs_super_block *)bh->b_data; - if (bfs_sb->s_magic != BFS_MAGIC) { + if (le32_to_cpu(bfs_sb->s_magic) != BFS_MAGIC) { if (!silent) printf("No BFS filesystem on %s (magic=%08x)\n", - s->s_id, bfs_sb->s_magic); + s->s_id, le32_to_cpu(bfs_sb->s_magic)); goto out; } if (BFS_UNCLEAN(bfs_sb, s) && !silent) printf("%s is unclean, continuing\n", s->s_id); s->s_magic = BFS_MAGIC; - info->si_bfs_sb = bfs_sb; info->si_sbh = bh; - info->si_lasti = (bfs_sb->s_start - BFS_BSIZE)/sizeof(struct bfs_inode) + info->si_lasti = (le32_to_cpu(bfs_sb->s_start) - BFS_BSIZE)/sizeof(struct bfs_inode) + BFS_ROOT_INO - 1; imap_len = info->si_lasti/8 + 1; @@ -348,8 +356,8 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent) goto out; } - info->si_blocks = (bfs_sb->s_end + 1)>>BFS_BSIZE_BITS; /* for statfs(2) */ - info->si_freeb = (bfs_sb->s_end + 1 - bfs_sb->s_start)>>BFS_BSIZE_BITS; + info->si_blocks = (le32_to_cpu(bfs_sb->s_end) + 1)>>BFS_BSIZE_BITS; /* for statfs(2) */ + info->si_freeb = (le32_to_cpu(bfs_sb->s_end) + 1 - cpu_to_le32(bfs_sb->s_start))>>BFS_BSIZE_BITS; info->si_freei = 0; info->si_lf_eblk = 0; info->si_lf_sblk = 0; diff --git a/include/linux/bfs_fs.h b/include/linux/bfs_fs.h index f7f0913cd110..c1237aa92e38 100644 --- a/include/linux/bfs_fs.h +++ b/include/linux/bfs_fs.h @@ -14,8 +14,9 @@ #define BFS_INODES_PER_BLOCK 8 /* SVR4 vnode type values (bfs_inode->i_vtype) */ -#define BFS_VDIR 2 -#define BFS_VREG 1 +#define BFS_VDIR 2L +#define BFS_VREG 1L + /* BFS inode layout on disk */ struct bfs_inode { @@ -58,22 +59,22 @@ struct bfs_super_block { __u32 s_padding[118]; }; -#define BFS_NZFILESIZE(ip) \ - (((ip)->i_eoffset + 1) - (ip)->i_sblock * BFS_BSIZE) - -#define BFS_FILESIZE(ip) \ - ((ip)->i_sblock == 0 ? 0 : BFS_NZFILESIZE(ip)) - -#define BFS_FILEBLOCKS(ip) \ - ((ip)->i_sblock == 0 ? 0 : ((ip)->i_eblock + 1) - (ip)->i_sblock) #define BFS_OFF2INO(offset) \ ((((offset) - BFS_BSIZE) / sizeof(struct bfs_inode)) + BFS_ROOT_INO) #define BFS_INO2OFF(ino) \ ((__u32)(((ino) - BFS_ROOT_INO) * sizeof(struct bfs_inode)) + BFS_BSIZE) +#define BFS_NZFILESIZE(ip) \ + ((cpu_to_le32((ip)->i_eoffset) + 1) - cpu_to_le32((ip)->i_sblock) * BFS_BSIZE) + +#define BFS_FILESIZE(ip) \ + ((ip)->i_sblock == 0 ? 0 : BFS_NZFILESIZE(ip)) +#define BFS_FILEBLOCKS(ip) \ + ((ip)->i_sblock == 0 ? 0 : (cpu_to_le32((ip)->i_eblock) + 1) - cpu_to_le32((ip)->i_sblock)) #define BFS_UNCLEAN(bfs_sb, sb) \ - ((bfs_sb->s_from != -1) && (bfs_sb->s_to != -1) && !(sb->s_flags & MS_RDONLY)) + ((cpu_to_le32(bfs_sb->s_from) != -1) && (cpu_to_le32(bfs_sb->s_to) != -1) && !(sb->s_flags & MS_RDONLY)) + #endif /* _LINUX_BFS_FS_H */ -- cgit v1.2.3 From b3743fa4442fc172e950ff0eaf6aa96e7d5ce9be Mon Sep 17 00:00:00 2001 From: Dominik Brodowski Date: Fri, 9 Sep 2005 13:03:23 -0700 Subject: [PATCH] yenta: share code with PCI core Share code between setup-bus.c and yenta_socket.c: use the write-out code of resources to the bridge also in yenta_socket.c, as it provides useful debug output. In addition, it fixes the bug that the CPU-centric resource view might need to be transferred to the PCI-centric view: setup-bus.c does that, while yenta-socket.c did not. Signed-off-by: Dominik Brodowski Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/pci/setup-bus.c | 4 ++-- drivers/pcmcia/yenta_socket.c | 41 ++++++++++++++++++----------------------- include/linux/pci.h | 1 + 3 files changed, 21 insertions(+), 25 deletions(-) (limited to 'include/linux') diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c index 6b0e6464eb39..657be948baf7 100644 --- a/drivers/pci/setup-bus.c +++ b/drivers/pci/setup-bus.c @@ -77,8 +77,7 @@ pbus_assign_resources_sorted(struct pci_bus *bus) } } -static void __devinit -pci_setup_cardbus(struct pci_bus *bus) +void pci_setup_cardbus(struct pci_bus *bus) { struct pci_dev *bridge = bus->self; struct pci_bus_region region; @@ -130,6 +129,7 @@ pci_setup_cardbus(struct pci_bus *bus) region.end); } } +EXPORT_SYMBOL(pci_setup_cardbus); /* Initialize bridges with base/limit values we have collected. PCI-to-PCI Bridge Architecture Specification rev. 1.1 (1998) diff --git a/drivers/pcmcia/yenta_socket.c b/drivers/pcmcia/yenta_socket.c index 0347a29f297b..271a52b7c2e6 100644 --- a/drivers/pcmcia/yenta_socket.c +++ b/drivers/pcmcia/yenta_socket.c @@ -667,7 +667,7 @@ static int yenta_search_res(struct yenta_socket *socket, struct resource *res, return 0; } -static void yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned type, int addr_start, int addr_end) +static int yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned type, int addr_start, int addr_end) { struct resource *root, *res; struct pci_bus_region region; @@ -676,7 +676,7 @@ static void yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned typ res = socket->dev->resource + PCI_BRIDGE_RESOURCES + nr; /* Already allocated? */ if (res->parent) - return; + return 0; /* The granularity of the memory limit is 4kB, on IO it's 4 bytes */ mask = ~0xfff; @@ -692,7 +692,7 @@ static void yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned typ pcibios_bus_to_resource(socket->dev, res, ®ion); root = pci_find_parent_resource(socket->dev, res); if (root && (request_resource(root, res) == 0)) - return; + return 0; printk(KERN_INFO "yenta %s: Preassigned resource %d busy or not available, reconfiguring...\n", pci_name(socket->dev), nr); } @@ -700,35 +700,27 @@ static void yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned typ if (type & IORESOURCE_IO) { if ((yenta_search_res(socket, res, BRIDGE_IO_MAX)) || (yenta_search_res(socket, res, BRIDGE_IO_ACC)) || - (yenta_search_res(socket, res, BRIDGE_IO_MIN))) { - config_writel(socket, addr_start, res->start); - config_writel(socket, addr_end, res->end); - return; - } + (yenta_search_res(socket, res, BRIDGE_IO_MIN))) + return 1; } else { if (type & IORESOURCE_PREFETCH) { if ((yenta_search_res(socket, res, BRIDGE_MEM_MAX)) || (yenta_search_res(socket, res, BRIDGE_MEM_ACC)) || - (yenta_search_res(socket, res, BRIDGE_MEM_MIN))) { - config_writel(socket, addr_start, res->start); - config_writel(socket, addr_end, res->end); - return; - } + (yenta_search_res(socket, res, BRIDGE_MEM_MIN))) + return 1; /* Approximating prefetchable by non-prefetchable */ res->flags = IORESOURCE_MEM; } if ((yenta_search_res(socket, res, BRIDGE_MEM_MAX)) || (yenta_search_res(socket, res, BRIDGE_MEM_ACC)) || - (yenta_search_res(socket, res, BRIDGE_MEM_MIN))) { - config_writel(socket, addr_start, res->start); - config_writel(socket, addr_end, res->end); - return; - } + (yenta_search_res(socket, res, BRIDGE_MEM_MIN))) + return 1; } printk(KERN_INFO "yenta %s: no resource of type %x available, trying to continue...\n", pci_name(socket->dev), type); res->start = res->end = res->flags = 0; + return 0; } /* @@ -736,14 +728,17 @@ static void yenta_allocate_res(struct yenta_socket *socket, int nr, unsigned typ */ static void yenta_allocate_resources(struct yenta_socket *socket) { - yenta_allocate_res(socket, 0, IORESOURCE_IO, + int program = 0; + program += yenta_allocate_res(socket, 0, IORESOURCE_IO, PCI_CB_IO_BASE_0, PCI_CB_IO_LIMIT_0); - yenta_allocate_res(socket, 1, IORESOURCE_IO, + program += yenta_allocate_res(socket, 1, IORESOURCE_IO, PCI_CB_IO_BASE_1, PCI_CB_IO_LIMIT_1); - yenta_allocate_res(socket, 2, IORESOURCE_MEM|IORESOURCE_PREFETCH, + program += yenta_allocate_res(socket, 2, IORESOURCE_MEM|IORESOURCE_PREFETCH, PCI_CB_MEMORY_BASE_0, PCI_CB_MEMORY_LIMIT_0); - yenta_allocate_res(socket, 3, IORESOURCE_MEM, + program += yenta_allocate_res(socket, 3, IORESOURCE_MEM, PCI_CB_MEMORY_BASE_1, PCI_CB_MEMORY_LIMIT_1); + if (program) + pci_setup_cardbus(socket->dev->subordinate); } @@ -758,7 +753,7 @@ static void yenta_free_resources(struct yenta_socket *socket) res = socket->dev->resource + PCI_BRIDGE_RESOURCES + i; if (res->start != 0 && res->end != 0) release_resource(res); - res->start = res->end = 0; + res->start = res->end = res->flags = 0; } } diff --git a/include/linux/pci.h b/include/linux/pci.h index 6caaba0af469..c62e89270237 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -326,6 +326,7 @@ extern struct pci_dev *pci_dev_get(struct pci_dev *dev); extern void pci_dev_put(struct pci_dev *dev); extern void pci_remove_bus(struct pci_bus *b); extern void pci_remove_bus_device(struct pci_dev *dev); +void pci_setup_cardbus(struct pci_bus *bus); /* Generic PCI functions exported to card drivers */ -- cgit v1.2.3 From 793cf9e6a54c698e109a599c8b8e303658fcaae6 Mon Sep 17 00:00:00 2001 From: Mauro Carvalho Chehab Date: Fri, 9 Sep 2005 13:03:37 -0700 Subject: [PATCH] v4l: common part Updates and tuner additions - Remove $Id CVS logs for V4L files - Included newer cards. - Added a new NEC protocol for ir based on pulse distance. - Enable ATSC support for DViCO FusionHDTV5 Gold. - Added tuner LG NTSC (TALN mini series). - Fixed tea5767 autodetection. - Resolve more tuner types. - Commented debug function removed from mainstream. - Remove comments from mainstream. Still on development tree. - linux/version dependencies removed. - BTSC Lang1 now is set to auto_stereo mode. - New tuner standby API. - i2c-core.c uses hexadecimal for the i2c address, so it should stay consistent. Signed-off-by: Uli Luckas Signed-off-by: Mac Michaels Signed-off-by: Michael Krufky Signed-off-by: Hermann Pitton Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/video4linux/CARDLIST.tuner | 1 + drivers/media/common/ir-common.c | 68 ++++++++++++++++++++++++++- drivers/media/video/btcx-risc.c | 1 - drivers/media/video/btcx-risc.h | 1 - drivers/media/video/ir-kbd-gpio.c | 1 - drivers/media/video/ir-kbd-i2c.c | 1 - drivers/media/video/msp3400.h | 1 - drivers/media/video/mt20xx.c | 2 +- drivers/media/video/tda8290.c | 19 +++++++- drivers/media/video/tda9887.c | 39 ++++++++++------ drivers/media/video/tea5767.c | 37 +++++++++------ drivers/media/video/tuner-core.c | 79 +++++++++++++++++++------------- drivers/media/video/tuner-simple.c | 5 +- drivers/media/video/tveeprom.c | 25 ++-------- drivers/media/video/tvmixer.c | 1 - drivers/media/video/v4l1-compat.c | 16 ------- drivers/media/video/v4l2-common.c | 18 -------- drivers/media/video/video-buf-dvb.c | 1 - drivers/media/video/video-buf.c | 1 - include/linux/videodev.h | 3 -- include/linux/videodev2.h | 4 -- include/media/audiochip.h | 1 - include/media/id.h | 1 - include/media/ir-common.h | 4 +- include/media/tuner.h | 25 +++++----- include/media/tveeprom.h | 1 - include/media/video-buf.h | 1 - 27 files changed, 203 insertions(+), 154 deletions(-) (limited to 'include/linux') diff --git a/Documentation/video4linux/CARDLIST.tuner b/Documentation/video4linux/CARDLIST.tuner index f3302e1b1b9c..f5876be658a6 100644 --- a/Documentation/video4linux/CARDLIST.tuner +++ b/Documentation/video4linux/CARDLIST.tuner @@ -64,3 +64,4 @@ tuner=62 - Philips TEA5767HN FM Radio tuner=63 - Philips FMD1216ME MK3 Hybrid Tuner tuner=64 - LG TDVS-H062F/TUA6034 tuner=65 - Ymec TVF66T5-B/DFF +tuner=66 - LG NTSC (TALN mini series) diff --git a/drivers/media/common/ir-common.c b/drivers/media/common/ir-common.c index ab7a1fba4427..a0e700d7a4a4 100644 --- a/drivers/media/common/ir-common.c +++ b/drivers/media/common/ir-common.c @@ -1,5 +1,4 @@ /* - * $Id: ir-common.c,v 1.11 2005/07/07 14:44:43 mchehab Exp $ * * some common structs and functions to handle infrared remotes via * input layer ... @@ -335,6 +334,72 @@ int ir_dump_samples(u32 *samples, int count) return 0; } +/* decode raw samples, pulse distance coding used by NEC remotes */ +int ir_decode_pulsedistance(u32 *samples, int count, int low, int high) +{ + int i,last,bit,len; + u32 curBit; + u32 value; + + /* find start burst */ + for (i = len = 0; i < count * 32; i++) { + bit = getbit(samples,i); + if (bit) { + len++; + } else { + if (len >= 29) + break; + len = 0; + } + } + + /* start burst to short */ + if (len < 29) + return 0xffffffff; + + /* find start silence */ + for (len = 0; i < count * 32; i++) { + bit = getbit(samples,i); + if (bit) { + break; + } else { + len++; + } + } + + /* silence to short */ + if (len < 7) + return 0xffffffff; + + /* go decoding */ + len = 0; + last = 1; + value = 0; curBit = 1; + for (; i < count * 32; i++) { + bit = getbit(samples,i); + if (last) { + if(bit) { + continue; + } else { + len = 1; + } + } else { + if (bit) { + if (len > (low + high) /2) + value |= curBit; + curBit <<= 1; + if (curBit == 1) + break; + } else { + len++; + } + } + last = bit; + } + + return value; +} + /* decode raw samples, biphase coding, used by rc5 for example */ int ir_decode_biphase(u32 *samples, int count, int low, int high) { @@ -383,6 +448,7 @@ EXPORT_SYMBOL_GPL(ir_input_keydown); EXPORT_SYMBOL_GPL(ir_extract_bits); EXPORT_SYMBOL_GPL(ir_dump_samples); EXPORT_SYMBOL_GPL(ir_decode_biphase); +EXPORT_SYMBOL_GPL(ir_decode_pulsedistance); /* * Local variables: diff --git a/drivers/media/video/btcx-risc.c b/drivers/media/video/btcx-risc.c index 7f2d515d2873..a48de3c0e3f0 100644 --- a/drivers/media/video/btcx-risc.c +++ b/drivers/media/video/btcx-risc.c @@ -1,5 +1,4 @@ /* - $Id: btcx-risc.c,v 1.6 2005/02/21 13:57:59 kraxel Exp $ btcx-risc.c diff --git a/drivers/media/video/btcx-risc.h b/drivers/media/video/btcx-risc.h index 41f60395a520..503e6c6d7b69 100644 --- a/drivers/media/video/btcx-risc.h +++ b/drivers/media/video/btcx-risc.h @@ -1,5 +1,4 @@ /* - * $Id: btcx-risc.h,v 1.2 2004/09/15 16:15:24 kraxel Exp $ */ struct btcx_riscmem { unsigned int size; diff --git a/drivers/media/video/ir-kbd-gpio.c b/drivers/media/video/ir-kbd-gpio.c index a565823330aa..eddadc76e11d 100644 --- a/drivers/media/video/ir-kbd-gpio.c +++ b/drivers/media/video/ir-kbd-gpio.c @@ -1,5 +1,4 @@ /* - * $Id: ir-kbd-gpio.c,v 1.13 2005/05/15 19:01:26 mchehab Exp $ * * Copyright (c) 2003 Gerd Knorr * Copyright (c) 2003 Pavel Machek diff --git a/drivers/media/video/ir-kbd-i2c.c b/drivers/media/video/ir-kbd-i2c.c index 1e273ff3f956..67105b9804a2 100644 --- a/drivers/media/video/ir-kbd-i2c.c +++ b/drivers/media/video/ir-kbd-i2c.c @@ -1,5 +1,4 @@ /* - * $Id: ir-kbd-i2c.c,v 1.11 2005/07/07 16:42:11 mchehab Exp $ * * keyboard input driver for i2c IR remote controls * diff --git a/drivers/media/video/msp3400.h b/drivers/media/video/msp3400.h index 023f33056a4f..2d9ff40f0b09 100644 --- a/drivers/media/video/msp3400.h +++ b/drivers/media/video/msp3400.h @@ -1,5 +1,4 @@ /* - * $Id: msp3400.h,v 1.3 2005/06/12 04:19:19 mchehab Exp $ */ #ifndef MSP3400_H diff --git a/drivers/media/video/mt20xx.c b/drivers/media/video/mt20xx.c index 2fb7c2d1787a..972aa5e0aeef 100644 --- a/drivers/media/video/mt20xx.c +++ b/drivers/media/video/mt20xx.c @@ -1,5 +1,4 @@ /* - * $Id: mt20xx.c,v 1.5 2005/06/16 08:29:49 nsh Exp $ * * i2c tv tuner chip device driver * controls microtune tuners, mt2032 + mt2050 at the moment. @@ -494,6 +493,7 @@ int microtune_init(struct i2c_client *c) memset(buf,0,sizeof(buf)); t->tv_freq = NULL; t->radio_freq = NULL; + t->standby = NULL; name = "unknown"; i2c_master_send(c,buf,1); diff --git a/drivers/media/video/tda8290.c b/drivers/media/video/tda8290.c index a8b6a8df5109..c65f0c7680a2 100644 --- a/drivers/media/video/tda8290.c +++ b/drivers/media/video/tda8290.c @@ -1,5 +1,4 @@ /* - * $Id: tda8290.c,v 1.15 2005/07/08 20:21:33 mchehab Exp $ * * i2c tv tuner chip device driver * controls the philips tda8290+75 tuner chip combo. @@ -9,6 +8,9 @@ #include #include +#define I2C_ADDR_TDA8290 0x4b +#define I2C_ADDR_TDA8275 0x61 + /* ---------------------------------------------------------------------- */ struct freq_entry { @@ -75,10 +77,12 @@ static unsigned char i2c_init_tda8275[14] = { 0x00, 0x00, 0x00, 0x00, static unsigned char i2c_set_VS[2] = { 0x30, 0x6F }; static unsigned char i2c_set_GP01_CF[2] = { 0x20, 0x0B }; static unsigned char i2c_tda8290_reset[2] = { 0x00, 0x00 }; +static unsigned char i2c_tda8290_standby[2] = { 0x00, 0x02 }; static unsigned char i2c_gainset_off[2] = { 0x28, 0x14 }; static unsigned char i2c_gainset_on[2] = { 0x28, 0x54 }; static unsigned char i2c_agc3_00[2] = { 0x80, 0x00 }; static unsigned char i2c_agc2_BF[2] = { 0x60, 0xBF }; +static unsigned char i2c_cb1_D0[2] = { 0x30, 0xD0 }; static unsigned char i2c_cb1_D2[2] = { 0x30, 0xD2 }; static unsigned char i2c_cb1_56[2] = { 0x30, 0x56 }; static unsigned char i2c_cb1_52[2] = { 0x30, 0x52 }; @@ -117,6 +121,13 @@ static struct i2c_msg i2c_msg_epilog[] = { { I2C_ADDR_TDA8290, 0, ARRAY_SIZE(i2c_gainset_on), i2c_gainset_on }, }; +static struct i2c_msg i2c_msg_standby[] = { + { I2C_ADDR_TDA8290, 0, ARRAY_SIZE(i2c_enable_bridge), i2c_enable_bridge }, + { I2C_ADDR_TDA8275, 0, ARRAY_SIZE(i2c_cb1_D0), i2c_cb1_D0 }, + { I2C_ADDR_TDA8290, 0, ARRAY_SIZE(i2c_disable_bridge), i2c_disable_bridge }, + { I2C_ADDR_TDA8290, 0, ARRAY_SIZE(i2c_tda8290_standby), i2c_tda8290_standby }, +}; + static int tda8290_tune(struct i2c_client *c) { struct tuner *t = i2c_get_clientdata(c); @@ -205,6 +216,11 @@ static int has_signal(struct i2c_client *c) return (afc & 0x80)? 65535:0; } +static void standby(struct i2c_client *c) +{ + i2c_transfer(c->adapter, i2c_msg_standby, ARRAY_SIZE(i2c_msg_standby)); +} + int tda8290_init(struct i2c_client *c) { struct tuner *t = i2c_get_clientdata(c); @@ -214,6 +230,7 @@ int tda8290_init(struct i2c_client *c) t->tv_freq = set_tv_freq; t->radio_freq = set_radio_freq; t->has_signal = has_signal; + t->standby = standby; i2c_master_send(c, i2c_enable_bridge, ARRAY_SIZE(i2c_enable_bridge)); i2c_transfer(c->adapter, i2c_msg_init, ARRAY_SIZE(i2c_msg_init)); diff --git a/drivers/media/video/tda9887.c b/drivers/media/video/tda9887.c index d60fc562aecd..79e0bd1aa70f 100644 --- a/drivers/media/video/tda9887.c +++ b/drivers/media/video/tda9887.c @@ -49,7 +49,7 @@ MODULE_LICENSE("GPL"); struct tda9887 { struct i2c_client client; v4l2_std_id std; - unsigned int radio; + enum tuner_mode mode; unsigned int config; unsigned int pinnacle_id; unsigned int using_v4l2; @@ -196,7 +196,7 @@ static struct tvnorm tvnorms[] = { .b = ( cNegativeFmTV | cQSS ), .c = ( cDeemphasisON | - cDeemphasis50 ), + cDeemphasis75 ), .e = ( cGating_36 | cAudioIF_4_5 | cVideoIF_45_75 ), @@ -364,7 +364,7 @@ static int tda9887_set_tvnorm(struct tda9887 *t, char *buf) struct tvnorm *norm = NULL; int i; - if (t->radio) { + if (t->mode == T_RADIO) { if (t->radio_mode == V4L2_TUNER_MODE_MONO) norm = &radio_mono; else @@ -378,7 +378,7 @@ static int tda9887_set_tvnorm(struct tda9887 *t, char *buf) } } if (NULL == norm) { - dprintk(PREFIX "Oops: no tvnorm entry found\n"); + dprintk(PREFIX "Unsupported tvnorm entry - audio muted\n"); return -1; } @@ -569,6 +569,10 @@ static int tda9887_configure(struct tda9887 *t) tda9887_set_config(t,buf); tda9887_set_insmod(t,buf); + if (t->mode == T_STANDBY) { + buf[1] |= cForcedMuteAudioON; + } + dprintk(PREFIX "writing: b=0x%02x c=0x%02x e=0x%02x\n", buf[1],buf[2],buf[3]); @@ -653,10 +657,17 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) /* --- configuration --- */ case AUDC_SET_RADIO: - t->radio = 1; + { + t->mode = T_RADIO; tda9887_configure(t); break; - + } + case TUNER_SET_STANDBY: + { + t->mode = T_STANDBY; + tda9887_configure(t); + break; + } case AUDC_CONFIG_PINNACLE: { int *i = arg; @@ -689,7 +700,7 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) struct video_channel *vc = arg; CHECK_V4L2; - t->radio = 0; + t->mode = T_ANALOG_TV; if (vc->norm < ARRAY_SIZE(map)) t->std = map[vc->norm]; tda9887_fixup_std(t); @@ -701,7 +712,7 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) v4l2_std_id *id = arg; SWITCH_V4L2; - t->radio = 0; + t->mode = T_ANALOG_TV; t->std = *id; tda9887_fixup_std(t); tda9887_configure(t); @@ -713,14 +724,14 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) SWITCH_V4L2; if (V4L2_TUNER_ANALOG_TV == f->type) { - if (t->radio == 0) + if (t->mode == T_ANALOG_TV) return 0; - t->radio = 0; + t->mode = T_ANALOG_TV; } if (V4L2_TUNER_RADIO == f->type) { - if (t->radio == 1) + if (t->mode == T_RADIO) return 0; - t->radio = 1; + t->mode = T_RADIO; } tda9887_configure(t); break; @@ -735,7 +746,7 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) }; struct v4l2_tuner* tuner = arg; - if (t->radio) { + if (t->mode == T_RADIO) { __u8 reg = 0; tuner->afc=0; if (1 == i2c_master_recv(&t->client,®,1)) @@ -747,7 +758,7 @@ tda9887_command(struct i2c_client *client, unsigned int cmd, void *arg) { struct v4l2_tuner* tuner = arg; - if (t->radio) { + if (t->mode == T_RADIO) { t->radio_mode = tuner->audmode; tda9887_configure (t); } diff --git a/drivers/media/video/tea5767.c b/drivers/media/video/tea5767.c index cebcc1fa68d1..38bf50943798 100644 --- a/drivers/media/video/tea5767.c +++ b/drivers/media/video/tea5767.c @@ -2,7 +2,6 @@ * For Philips TEA5767 FM Chip used on some TV Cards like Prolink Pixelview * I2C address is allways 0xC0. * - * $Id: tea5767.c,v 1.27 2005/07/31 12:10:56 mchehab Exp $ * * Copyright (c) 2005 Mauro Carvalho Chehab (mchehab@brturbo.com.br) * This code is placed under the terms of the GNU General Public License @@ -205,11 +204,6 @@ static void set_radio_freq(struct i2c_client *c, unsigned int frq) TEA5767_ST_NOISE_CTL | TEA5767_JAPAN_BAND; buffer[4] = 0; - if (t->mode == T_STANDBY) { - tuner_dbg("TEA5767 set to standby mode\n"); - buffer[3] |= TEA5767_STDBY; - } - if (t->audmode == V4L2_TUNER_MODE_MONO) { tuner_dbg("TEA5767 set to mono\n"); buffer[2] |= TEA5767_MONO; @@ -290,13 +284,31 @@ static int tea5767_stereo(struct i2c_client *c) return ((buffer[2] & TEA5767_STEREO_MASK) ? V4L2_TUNER_SUB_STEREO : 0); } +static void tea5767_standby(struct i2c_client *c) +{ + unsigned char buffer[5]; + struct tuner *t = i2c_get_clientdata(c); + unsigned div, rc; + + div = (87500 * 4 + 700 + 225 + 25) / 50; /* Set frequency to 87.5 MHz */ + buffer[0] = (div >> 8) & 0x3f; + buffer[1] = div & 0xff; + buffer[2] = TEA5767_PORT1_HIGH; + buffer[3] = TEA5767_PORT2_HIGH | TEA5767_HIGH_CUT_CTRL | + TEA5767_ST_NOISE_CTL | TEA5767_JAPAN_BAND | TEA5767_STDBY; + buffer[4] = 0; + + if (5 != (rc = i2c_master_send(c, buffer, 5))) + tuner_warn("i2c i/o error: rc == %d (should be 5)\n", rc); +} + int tea5767_autodetection(struct i2c_client *c) { unsigned char buffer[7] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }; int rc; struct tuner *t = i2c_get_clientdata(c); - if (7 != (rc = i2c_master_recv(c, buffer, 7))) { + if ((rc = i2c_master_recv(c, buffer, 7))< 5) { tuner_warn("It is not a TEA5767. Received %i bytes.\n", rc); return EINVAL; } @@ -313,15 +325,10 @@ int tea5767_autodetection(struct i2c_client *c) * bit 0 : internally set to 0 * Byte 5: bit 7:0 : == 0 */ - if (!((buffer[3] & 0x0f) == 0x00) && (buffer[4] == 0x00)) { + if (((buffer[3] & 0x0f) != 0x00) || (buffer[4] != 0x00)) { tuner_warn("Chip ID is not zero. It is not a TEA5767\n"); return EINVAL; } - /* It seems that tea5767 returns 0xff after the 5th byte */ - if ((buffer[5] != 0xff) || (buffer[6] != 0xff)) { - tuner_warn("Returned more than 5 bytes. It is not a TEA5767\n"); - return EINVAL; - } /* It seems that tea5767 returns 0xff after the 5th byte */ if ((buffer[5] != 0xff) || (buffer[6] != 0xff)) { @@ -337,14 +344,14 @@ int tea5767_tuner_init(struct i2c_client *c) { struct tuner *t = i2c_get_clientdata(c); - tuner_info("type set to %d (%s)\n", t->type, - "Philips TEA5767HN FM Radio"); + tuner_info("type set to %d (%s)\n", t->type, "Philips TEA5767HN FM Radio"); strlcpy(c->name, "tea5767", sizeof(c->name)); t->tv_freq = set_tv_freq; t->radio_freq = set_radio_freq; t->has_signal = tea5767_signal; t->is_stereo = tea5767_stereo; + t->standby = tea5767_standby; return (0); } diff --git a/drivers/media/video/tuner-core.c b/drivers/media/video/tuner-core.c index 3b1893c2ae3b..afc96bbb1c11 100644 --- a/drivers/media/video/tuner-core.c +++ b/drivers/media/video/tuner-core.c @@ -1,5 +1,4 @@ /* - * $Id: tuner-core.c,v 1.63 2005/07/28 18:19:55 mchehab Exp $ * * i2c tv tuner chip device driver * core core, i.e. kernel interfaces, registering and so on @@ -182,6 +181,14 @@ static void set_type(struct i2c_client *c, unsigned int type, i2c_master_send(c, buffer, 4); default_tuner_init(c); break; + case TUNER_LG_TDVS_H062F: + /* Set the Auxiliary Byte. */ + buffer[2] &= ~0x20; + buffer[2] |= 0x18; + buffer[3] = 0x20; + i2c_master_send(c, buffer, 4); + default_tuner_init(c); + break; default: default_tuner_init(c); break; @@ -208,31 +215,31 @@ static void set_addr(struct i2c_client *c, struct tuner_setup *tun_setup) { struct tuner *t = i2c_get_clientdata(c); - if (tun_setup->addr == ADDR_UNSET) { - if (t->mode_mask & tun_setup->mode_mask) + if ((tun_setup->addr == ADDR_UNSET && + (t->mode_mask & tun_setup->mode_mask)) || + tun_setup->addr == c->addr) { set_type(c, tun_setup->type, tun_setup->mode_mask); - } else if (tun_setup->addr == c->addr) { - set_type(c, tun_setup->type, tun_setup->mode_mask); } } static inline int check_mode(struct tuner *t, char *cmd) { - if (1 << t->mode & t->mode_mask) { - switch (t->mode) { - case V4L2_TUNER_RADIO: - tuner_dbg("Cmd %s accepted for radio\n", cmd); - break; - case V4L2_TUNER_ANALOG_TV: - tuner_dbg("Cmd %s accepted for analog TV\n", cmd); - break; - case V4L2_TUNER_DIGITAL_TV: - tuner_dbg("Cmd %s accepted for digital TV\n", cmd); - break; - } - return 0; + if ((1 << t->mode & t->mode_mask) == 0) { + return EINVAL; + } + + switch (t->mode) { + case V4L2_TUNER_RADIO: + tuner_dbg("Cmd %s accepted for radio\n", cmd); + break; + case V4L2_TUNER_ANALOG_TV: + tuner_dbg("Cmd %s accepted for analog TV\n", cmd); + break; + case V4L2_TUNER_DIGITAL_TV: + tuner_dbg("Cmd %s accepted for digital TV\n", cmd); + break; } - return EINVAL; + return 0; } static char pal[] = "-"; @@ -406,20 +413,18 @@ static int tuner_detach(struct i2c_client *client) static inline int set_mode(struct i2c_client *client, struct tuner *t, int mode, char *cmd) { - if (mode != t->mode) { - - t->mode = mode; - if (check_mode(t, cmd) == EINVAL) { - t->mode = T_STANDBY; - if (V4L2_TUNER_RADIO == mode) { - set_tv_freq(client, 400 * 16); - } else { - set_radio_freq(client, 87.5 * 16000); - } - return EINVAL; - } - } - return 0; + if (mode == t->mode) + return 0; + + t->mode = mode; + + if (check_mode(t, cmd) == EINVAL) { + t->mode = T_STANDBY; + if (t->standby) + t->standby (client); + return EINVAL; + } + return 0; } #define switch_v4l2() if (!t->using_v4l2) \ @@ -453,6 +458,14 @@ static int tuner_command(struct i2c_client *client, unsigned int cmd, void *arg) case AUDC_SET_RADIO: set_mode(client,t,V4L2_TUNER_RADIO, "AUDC_SET_RADIO"); break; + case TUNER_SET_STANDBY: + { + if (check_mode(t, "TUNER_SET_STANDBY") == EINVAL) + return 0; + if (t->standby) + t->standby (client); + break; + } case AUDC_CONFIG_PINNACLE: if (check_mode(t, "AUDC_CONFIG_PINNACLE") == EINVAL) return 0; diff --git a/drivers/media/video/tuner-simple.c b/drivers/media/video/tuner-simple.c index de0c93aeb75d..26034406b372 100644 --- a/drivers/media/video/tuner-simple.c +++ b/drivers/media/video/tuner-simple.c @@ -1,5 +1,4 @@ /* - * $Id: tuner-simple.c,v 1.43 2005/07/28 18:41:21 mchehab Exp $ * * i2c tv tuner chip device driver * controls all those simple 4-control-bytes style tuners. @@ -248,9 +247,10 @@ static struct tunertype tuners[] = { { "LG TDVS-H062F/TUA6034", LGINNOTEK, NTSC, 16*160.00,16*455.00,0x01,0x02,0x04,0x8e,732}, - { "Ymec TVF66T5-B/DFF", Philips, PAL, 16*160.25,16*464.25,0x01,0x02,0x08,0x8e,623}, + { "LG NTSC (TALN mini series)", LGINNOTEK, NTSC, + 16*150.00,16*425.00,0x01,0x02,0x08,0x8e,732 }, }; unsigned const int tuner_count = ARRAY_SIZE(tuners); @@ -497,6 +497,7 @@ int default_tuner_init(struct i2c_client *c) t->radio_freq = default_set_radio_freq; t->has_signal = tuner_signal; t->is_stereo = tuner_stereo; + t->standby = NULL; return 0; } diff --git a/drivers/media/video/tveeprom.c b/drivers/media/video/tveeprom.c index 3c3356a01cc6..d0a00d3a6c4f 100644 --- a/drivers/media/video/tveeprom.c +++ b/drivers/media/video/tveeprom.c @@ -155,10 +155,10 @@ hauppauge_tuner[] = { TUNER_ABSENT, "Philips FQ1216ME MK3"}, { TUNER_ABSENT, "Philips FI1236 MK3"}, { TUNER_PHILIPS_FM1216ME_MK3, "Philips FM1216 ME MK3"}, - { TUNER_ABSENT, "Philips FM1236 MK3"}, + { TUNER_PHILIPS_FM1236_MK3, "Philips FM1236 MK3"}, { TUNER_ABSENT, "Philips FM1216MP MK3"}, /* 60-69 */ - { TUNER_ABSENT, "LG S001D MK3"}, + { TUNER_PHILIPS_FM1216ME_MK3, "LG S001D MK3"}, { TUNER_ABSENT, "LG M001D MK3"}, { TUNER_ABSENT, "LG S701D MK3"}, { TUNER_ABSENT, "LG M701D MK3"}, @@ -183,8 +183,8 @@ hauppauge_tuner[] = { TUNER_ABSENT, "Philips FQ1216LME MK3"}, { TUNER_ABSENT, "LG TAPC G701D"}, { TUNER_LG_NTSC_NEW_TAPC, "LG TAPC H791F"}, - { TUNER_ABSENT, "TCL 2002MB 3"}, - { TUNER_ABSENT, "TCL 2002MI 3"}, + { TUNER_LG_PAL_NEW_TAPC, "TCL 2002MB 3"}, + { TUNER_LG_PAL_NEW_TAPC, "TCL 2002MI 3"}, { TUNER_TCL_2002N, "TCL 2002N 6A"}, { TUNER_ABSENT, "Philips FQ1236 MK3"}, { TUNER_ABSENT, "Samsung TCPN 2121P30A"}, @@ -445,23 +445,6 @@ int tveeprom_read(struct i2c_client *c, unsigned char *eedata, int len) } EXPORT_SYMBOL(tveeprom_read); -#if 0 -int tveeprom_dump(unsigned char *eedata, int len) -{ - int i; - - dprintk(1, "%s\n",__FUNCTION__); - for (i = 0; i < len; i++) { - if (0 == (i % 16)) - printk(KERN_INFO "tveeprom: %02x:",i); - printk(" %02x",eedata[i]); - if (15 == (i % 16)) - printk("\n"); - } - return 0; -} -EXPORT_SYMBOL(tveeprom_dump); -#endif /* 0 */ /* ----------------------------------------------------------------------- */ /* needed for ivtv.sf.net at the moment. Should go away in the long */ diff --git a/drivers/media/video/tvmixer.c b/drivers/media/video/tvmixer.c index a43301a154af..d86e08ebddfc 100644 --- a/drivers/media/video/tvmixer.c +++ b/drivers/media/video/tvmixer.c @@ -1,5 +1,4 @@ /* - * $Id: tvmixer.c,v 1.8 2005/06/12 04:19:19 mchehab Exp $ */ #include diff --git a/drivers/media/video/v4l1-compat.c b/drivers/media/video/v4l1-compat.c index 70ecbdb80277..59bb71381a1b 100644 --- a/drivers/media/video/v4l1-compat.c +++ b/drivers/media/video/v4l1-compat.c @@ -1,5 +1,4 @@ /* - * $Id: v4l1-compat.c,v 1.9 2005/06/12 04:19:19 mchehab Exp $ * * Video for Linux Two * Backward Compatibility Layer @@ -604,9 +603,6 @@ v4l_compat_translate_ioctl(struct inode *inode, dprintk("VIDIOCGPICT / VIDIOC_G_FMT: %d\n",err); break; } -#if 0 /* FIXME */ - pict->depth = fmt2->fmt.pix.depth; -#endif pict->palette = pixelformat_to_palette( fmt2->fmt.pix.pixelformat); break; @@ -707,13 +703,7 @@ v4l_compat_translate_ioctl(struct inode *inode, } case VIDIOCSTUNER: /* select a tuner input */ { -#if 0 /* FIXME */ - err = drv(inode, file, VIDIOC_S_INPUT, &i); - if (err < 0) - dprintk("VIDIOCSTUNER / VIDIOC_S_INPUT: %d\n",err); -#else err = 0; -#endif break; } case VIDIOCGFREQ: /* get frequency */ @@ -852,12 +842,6 @@ v4l_compat_translate_ioctl(struct inode *inode, err = 0; break; } -#if 0 - case VIDIOCGMBUF: - /* v4l2 drivers must implement that themself. The - mmap() differences can't be translated fully - transparent, thus there is no point to try that */ -#endif case VIDIOCMCAPTURE: /* capture a frame */ { struct video_mmap *mm = arg; diff --git a/drivers/media/video/v4l2-common.c b/drivers/media/video/v4l2-common.c index b5e0cf3448f4..597b8db35a13 100644 --- a/drivers/media/video/v4l2-common.c +++ b/drivers/media/video/v4l2-common.c @@ -84,20 +84,6 @@ MODULE_LICENSE("GPL"); * Video Standard Operations (contributed by Michael Schimek) */ -#if 0 /* seems to have no users */ -/* This is the recommended method to deal with the framerate fields. More - sophisticated drivers will access the fields directly. */ -unsigned int -v4l2_video_std_fps(struct v4l2_standard *vs) -{ - if (vs->frameperiod.numerator > 0) - return (((vs->frameperiod.denominator << 8) / - vs->frameperiod.numerator) + - (1 << 7)) / (1 << 8); - return 0; -} -EXPORT_SYMBOL(v4l2_video_std_fps); -#endif /* Fill in the fields of a v4l2_standard structure according to the 'id' and 'transmission' parameters. Returns negative on error. */ @@ -213,10 +199,6 @@ char *v4l2_ioctl_names[256] = { [_IOC_NR(VIDIOC_ENUM_FMT)] = "VIDIOC_ENUM_FMT", [_IOC_NR(VIDIOC_G_FMT)] = "VIDIOC_G_FMT", [_IOC_NR(VIDIOC_S_FMT)] = "VIDIOC_S_FMT", -#if 0 - [_IOC_NR(VIDIOC_G_COMP)] = "VIDIOC_G_COMP", - [_IOC_NR(VIDIOC_S_COMP)] = "VIDIOC_S_COMP", -#endif [_IOC_NR(VIDIOC_REQBUFS)] = "VIDIOC_REQBUFS", [_IOC_NR(VIDIOC_QUERYBUF)] = "VIDIOC_QUERYBUF", [_IOC_NR(VIDIOC_G_FBUF)] = "VIDIOC_G_FBUF", diff --git a/drivers/media/video/video-buf-dvb.c b/drivers/media/video/video-buf-dvb.c index 15f5bb486963..55f129e964eb 100644 --- a/drivers/media/video/video-buf-dvb.c +++ b/drivers/media/video/video-buf-dvb.c @@ -1,5 +1,4 @@ /* - * $Id: video-buf-dvb.c,v 1.7 2004/12/09 12:51:35 kraxel Exp $ * * some helper function for simple DVB cards which simply DMA the * complete transport stream and let the computer sort everything else diff --git a/drivers/media/video/video-buf.c b/drivers/media/video/video-buf.c index 5afdc7852610..97354f253a80 100644 --- a/drivers/media/video/video-buf.c +++ b/drivers/media/video/video-buf.c @@ -1,5 +1,4 @@ /* - * $Id: video-buf.c,v 1.18 2005/02/24 13:32:30 kraxel Exp $ * * generic helper functions for video4linux capture buffers, to handle * memory management and PCI DMA. Right now bttv + saa7134 use it. diff --git a/include/linux/videodev.h b/include/linux/videodev.h index 9d6fbde3d29c..1cc8c31b7988 100644 --- a/include/linux/videodev.h +++ b/include/linux/videodev.h @@ -3,7 +3,6 @@ #include #include -#include #define HAVE_V4L2 1 #include @@ -29,7 +28,6 @@ struct video_device void (*release)(struct video_device *vfd); -#if 1 /* to be removed in 2.7.x */ /* obsolete -- fops->owner is used instead */ struct module *owner; /* dev->driver_data will be used instead some day. @@ -37,7 +35,6 @@ struct video_device * so the switch over will be transparent for you. * Or use {pci|usb}_{get|set}_drvdata() directly. */ void *priv; -#endif /* for videodev.c intenal usage -- please don't touch */ int users; /* video_exclusive_{open|close} ... */ diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h index acbfc525576d..f623a33b9abe 100644 --- a/include/linux/videodev2.h +++ b/include/linux/videodev2.h @@ -270,7 +270,6 @@ struct v4l2_timecode /* The above is based on SMPTE timecodes */ -#if 1 /* * M P E G C O M P R E S S I O N P A R A M E T E R S * @@ -357,7 +356,6 @@ struct v4l2_mpeg_compression { /* I don't expect the above being perfect yet ;) */ __u32 reserved_5[8]; }; -#endif struct v4l2_jpegcompression { @@ -871,10 +869,8 @@ struct v4l2_streamparm #define VIDIOC_ENUM_FMT _IOWR ('V', 2, struct v4l2_fmtdesc) #define VIDIOC_G_FMT _IOWR ('V', 4, struct v4l2_format) #define VIDIOC_S_FMT _IOWR ('V', 5, struct v4l2_format) -#if 1 /* experimental */ #define VIDIOC_G_MPEGCOMP _IOR ('V', 6, struct v4l2_mpeg_compression) #define VIDIOC_S_MPEGCOMP _IOW ('V', 7, struct v4l2_mpeg_compression) -#endif #define VIDIOC_REQBUFS _IOWR ('V', 8, struct v4l2_requestbuffers) #define VIDIOC_QUERYBUF _IOWR ('V', 9, struct v4l2_buffer) #define VIDIOC_G_FBUF _IOR ('V', 10, struct v4l2_framebuffer) diff --git a/include/media/audiochip.h b/include/media/audiochip.h index cd831168fdc1..a7ceee9fc5e9 100644 --- a/include/media/audiochip.h +++ b/include/media/audiochip.h @@ -1,5 +1,4 @@ /* - * $Id: audiochip.h,v 1.5 2005/06/16 22:59:16 hhackmann Exp $ */ #ifndef AUDIOCHIP_H diff --git a/include/media/id.h b/include/media/id.h index 801ddef301aa..6d02c94cdc0d 100644 --- a/include/media/id.h +++ b/include/media/id.h @@ -1,5 +1,4 @@ /* - * $Id: id.h,v 1.4 2005/06/12 04:19:19 mchehab Exp $ */ /* FIXME: this temporarely, until these are included in linux/i2c-id.h */ diff --git a/include/media/ir-common.h b/include/media/ir-common.h index 698670547f16..01b56822df4d 100644 --- a/include/media/ir-common.h +++ b/include/media/ir-common.h @@ -1,5 +1,4 @@ /* - * $Id: ir-common.h,v 1.9 2005/05/15 19:01:26 mchehab Exp $ * * some common structs and functions to handle infrared remotes via * input layer ... @@ -21,11 +20,11 @@ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ -#include #include #define IR_TYPE_RC5 1 +#define IR_TYPE_PD 2 /* Pulse distance encoded IR */ #define IR_TYPE_OTHER 99 #define IR_KEYTAB_TYPE u32 @@ -60,6 +59,7 @@ void ir_input_keydown(struct input_dev *dev, struct ir_input_state *ir, u32 ir_extract_bits(u32 data, u32 mask); int ir_dump_samples(u32 *samples, int count); int ir_decode_biphase(u32 *samples, int count, int low, int high); +int ir_decode_pulsedistance(u32 *samples, int count, int low, int high); /* * Local variables: diff --git a/include/media/tuner.h b/include/media/tuner.h index eeaa15ddee85..252673bfa592 100644 --- a/include/media/tuner.h +++ b/include/media/tuner.h @@ -1,5 +1,3 @@ - -/* $Id: tuner.h,v 1.45 2005/07/28 18:41:21 mchehab Exp $ * tuner.h - definition for different tuners @@ -111,6 +109,8 @@ #define TUNER_LG_TDVS_H062F 64 /* DViCO FusionHDTV 5 */ #define TUNER_YMEC_TVF66T5_B_DFF 65 /* Acorp Y878F */ +#define TUNER_LG_NTSC_TALN_MINI 66 + #define NOTUNER 0 #define PAL 1 /* PAL_BG */ #define PAL_I 2 @@ -134,6 +134,7 @@ #define THOMSON 12 #define TUNER_SET_TYPE_ADDR _IOW('T',3,int) +#define TUNER_SET_STANDBY _IOW('T',4,int) #define TDA9887_SET_CONFIG _IOW('t',5,int) /* tv card specific */ @@ -153,9 +154,6 @@ #ifdef __KERNEL__ -#define I2C_ADDR_TDA8290 0x4b -#define I2C_ADDR_TDA8275 0x61 - enum tuner_mode { T_UNINITIALIZED = 0, T_RADIO = 1 << V4L2_TUNER_RADIO, @@ -198,6 +196,7 @@ struct tuner { void (*radio_freq)(struct i2c_client *c, unsigned int freq); int (*has_signal)(struct i2c_client *c); int (*is_stereo)(struct i2c_client *c); + void (*standby)(struct i2c_client *c); }; extern unsigned int tuner_debug; @@ -209,12 +208,16 @@ extern int tea5767_tuner_init(struct i2c_client *c); extern int default_tuner_init(struct i2c_client *c); extern int tea5767_autodetection(struct i2c_client *c); -#define tuner_warn(fmt, arg...) \ - dev_printk(KERN_WARNING , &t->i2c.dev , fmt , ## arg) -#define tuner_info(fmt, arg...) \ - dev_printk(KERN_INFO , &t->i2c.dev , fmt , ## arg) -#define tuner_dbg(fmt, arg...) \ - if (tuner_debug) dev_printk(KERN_DEBUG , &t->i2c.dev , fmt , ## arg) +#define tuner_warn(fmt, arg...) do {\ + printk(KERN_WARNING "%s %d-%04x: " fmt, t->i2c.driver->name, \ + t->i2c.adapter->nr, t->i2c.addr , ##arg); } while (0) +#define tuner_info(fmt, arg...) do {\ + printk(KERN_INFO "%s %d-%04x: " fmt, t->i2c.driver->name, \ + t->i2c.adapter->nr, t->i2c.addr , ##arg); } while (0) +#define tuner_dbg(fmt, arg...) do {\ + if (tuner_debug) \ + printk(KERN_DEBUG "%s %d-%04x: " fmt, t->i2c.driver->name, \ + t->i2c.adapter->nr, t->i2c.addr , ##arg); } while (0) #endif /* __KERNEL__ */ diff --git a/include/media/tveeprom.h b/include/media/tveeprom.h index 854a2c2f105b..e24e841c3211 100644 --- a/include/media/tveeprom.h +++ b/include/media/tveeprom.h @@ -1,5 +1,4 @@ /* - * $Id: tveeprom.h,v 1.2 2005/06/12 04:19:19 mchehab Exp $ */ struct tveeprom { diff --git a/include/media/video-buf.h b/include/media/video-buf.h index ae6da6de98de..ae8d7a000440 100644 --- a/include/media/video-buf.h +++ b/include/media/video-buf.h @@ -1,5 +1,4 @@ /* - * $Id: video-buf.h,v 1.9 2004/11/07 13:17:15 kraxel Exp $ * * generic helper functions for video4linux capture buffers, to handle * memory management and PCI DMA. Right now bttv + saa7134 use it. -- cgit v1.2.3 From 8b6490e5faafb3a16ea45654fb55f9ff086f1495 Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Fri, 9 Sep 2005 13:04:07 -0700 Subject: [PATCH] files: fix rcu initializers First of a number of files_lock scaability patches. Here are the x86 numbers - tiobench on a 4(8)-way (HT) P4 system on ramdisk : (lockfree) Test 2.6.10-vanilla Stdev 2.6.10-fd Stdev ------------------------------------------------------------- Seqread 1400.8 11.52 1465.4 34.27 Randread 1594 8.86 2397.2 29.21 Seqwrite 242.72 3.47 238.46 6.53 Randwrite 445.74 9.15 446.4 9.75 The performance improvement is very significant. We are getting killed by the cacheline bouncing of the files_struct lock here. Writes on ramdisk (ext2) seems to vary just too much to get any meaningful number. Also, With Tridge's thread_perf test on a 4(8)-way (HT) P4 xeon system : 2.6.12-rc5-vanilla : Running test 'readwrite' with 8 tasks Threads 0.34 +/- 0.01 seconds Processes 0.16 +/- 0.00 seconds 2.6.12-rc5-fd : Running test 'readwrite' with 8 tasks Threads 0.17 +/- 0.02 seconds Processes 0.17 +/- 0.02 seconds I repeated the measurements on ramfs (as opposed to ext2 on ramdisk in the earlier measurement) and I got more consistent results from tiobench : 4(8) way xeon P4 ----------------- (lock-free) Test 2.6.12-rc5 Stdev 2.6.12-rc5-fd Stdev ------------------------------------------------------------- Seqread 1282 18.59 1343.6 26.37 Randread 1517 7 2415 34.27 Seqwrite 702.2 5.27 709.46 5.9 Randwrite 846.86 15.15 919.68 21.4 4-way ppc64 ------------ (lock-free) Test 2.6.12-rc5 Stdev 2.6.12-rc5-fd Stdev ------------------------------------------------------------- Seqread 1549 91.16 1569.6 47.2 Randread 1473.6 25.11 1585.4 69.99 Seqwrite 1096.8 20.03 1136 29.61 Randwrite 1189.6 4.04 1275.2 32.96 Also running Tridge's thread_perf test on ppc64 : 2.6.12-rc5-vanilla -------------------- Running test 'readwrite' with 4 tasks Threads 0.20 +/- 0.02 seconds Processes 0.16 +/- 0.01 seconds 2.6.12-rc5-fd -------------------- Running test 'readwrite' with 4 tasks Threads 0.18 +/- 0.04 seconds Processes 0.16 +/- 0.01 seconds The benefits are huge (upto ~60%) in some cases on x86 primarily due to the atomic operations during acquisition of ->file_lock and cache line bouncing in fast path. ppc64 benefits are modest due to LL/SC based locking, but still statistically significant. This patch: RCU head initilizer no longer needs the head varible name since we don't use list.h lists anymore. Signed-off-by: Dipankar Sarma Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/rcupdate.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index fd276adf0fd5..4e65eb44adfd 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -52,8 +52,8 @@ struct rcu_head { void (*func)(struct rcu_head *head); }; -#define RCU_HEAD_INIT(head) { .next = NULL, .func = NULL } -#define RCU_HEAD(head) struct rcu_head head = RCU_HEAD_INIT(head) +#define RCU_HEAD_INIT { .next = NULL, .func = NULL } +#define RCU_HEAD(head) struct rcu_head head = RCU_HEAD_INIT #define INIT_RCU_HEAD(ptr) do { \ (ptr)->next = NULL; (ptr)->func = NULL; \ } while (0) -- cgit v1.2.3 From c0dfb2905126e9e94edebbce8d3e05001301f52d Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Fri, 9 Sep 2005 13:04:09 -0700 Subject: [PATCH] files: rcuref APIs Adds a set of primitives to do reference counting for objects that are looked up without locks using RCU. Signed-off-by: Ravikiran Thirumalai Signed-off-by: Dipankar Sarma Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/RCU/rcuref.txt | 74 +++++++++++++++ include/linux/rcuref.h | 220 +++++++++++++++++++++++++++++++++++++++++++ kernel/rcupdate.c | 14 +++ 3 files changed, 308 insertions(+) create mode 100644 Documentation/RCU/rcuref.txt create mode 100644 include/linux/rcuref.h (limited to 'include/linux') diff --git a/Documentation/RCU/rcuref.txt b/Documentation/RCU/rcuref.txt new file mode 100644 index 000000000000..a23fee66064d --- /dev/null +++ b/Documentation/RCU/rcuref.txt @@ -0,0 +1,74 @@ +Refcounter framework for elements of lists/arrays protected by +RCU. + +Refcounting on elements of lists which are protected by traditional +reader/writer spinlocks or semaphores are straight forward as in: + +1. 2. +add() search_and_reference() +{ { + alloc_object read_lock(&list_lock); + ... search_for_element + atomic_set(&el->rc, 1); atomic_inc(&el->rc); + write_lock(&list_lock); ... + add_element read_unlock(&list_lock); + ... ... + write_unlock(&list_lock); } +} + +3. 4. +release_referenced() delete() +{ { + ... write_lock(&list_lock); + atomic_dec(&el->rc, relfunc) ... + ... delete_element +} write_unlock(&list_lock); + ... + if (atomic_dec_and_test(&el->rc)) + kfree(el); + ... + } + +If this list/array is made lock free using rcu as in changing the +write_lock in add() and delete() to spin_lock and changing read_lock +in search_and_reference to rcu_read_lock(), the rcuref_get in +search_and_reference could potentially hold reference to an element which +has already been deleted from the list/array. rcuref_lf_get_rcu takes +care of this scenario. search_and_reference should look as; + +1. 2. +add() search_and_reference() +{ { + alloc_object rcu_read_lock(); + ... search_for_element + atomic_set(&el->rc, 1); if (rcuref_inc_lf(&el->rc)) { + write_lock(&list_lock); rcu_read_unlock(); + return FAIL; + add_element } + ... ... + write_unlock(&list_lock); rcu_read_unlock(); +} } +3. 4. +release_referenced() delete() +{ { + ... write_lock(&list_lock); + rcuref_dec(&el->rc, relfunc) ... + ... delete_element +} write_unlock(&list_lock); + ... + if (rcuref_dec_and_test(&el->rc)) + call_rcu(&el->head, el_free); + ... + } + +Sometimes, reference to the element need to be obtained in the +update (write) stream. In such cases, rcuref_inc_lf might be an overkill +since the spinlock serialising list updates are held. rcuref_inc +is to be used in such cases. +For arches which do not have cmpxchg rcuref_inc_lf +api uses a hashed spinlock implementation and the same hashed spinlock +is acquired in all rcuref_xxx primitives to preserve atomicity. +Note: Use rcuref_inc api only if you need to use rcuref_inc_lf on the +refcounter atleast at one place. Mixing rcuref_inc and atomic_xxx api +might lead to races. rcuref_inc_lf() must be used in lockfree +RCU critical sections only. diff --git a/include/linux/rcuref.h b/include/linux/rcuref.h new file mode 100644 index 000000000000..e1adbba14b67 --- /dev/null +++ b/include/linux/rcuref.h @@ -0,0 +1,220 @@ +/* + * rcuref.h + * + * Reference counting for elements of lists/arrays protected by + * RCU. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (C) IBM Corporation, 2005 + * + * Author: Dipankar Sarma + * Ravikiran Thirumalai + * + * See Documentation/RCU/rcuref.txt for detailed user guide. + * + */ + +#ifndef _RCUREF_H_ +#define _RCUREF_H_ + +#ifdef __KERNEL__ + +#include +#include +#include +#include + +/* + * These APIs work on traditional atomic_t counters used in the + * kernel for reference counting. Under special circumstances + * where a lock-free get() operation races with a put() operation + * these APIs can be used. See Documentation/RCU/rcuref.txt. + */ + +#ifdef __HAVE_ARCH_CMPXCHG + +/** + * rcuref_inc - increment refcount for object. + * @rcuref: reference counter in the object in question. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference + * in a lock-free reader-side critical section. + */ +static inline void rcuref_inc(atomic_t *rcuref) +{ + atomic_inc(rcuref); +} + +/** + * rcuref_dec - decrement refcount for object. + * @rcuref: reference counter in the object in question. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference + * in a lock-free reader-side critical section. + */ +static inline void rcuref_dec(atomic_t *rcuref) +{ + atomic_dec(rcuref); +} + +/** + * rcuref_dec_and_test - decrement refcount for object and test + * @rcuref: reference counter in the object. + * @release: pointer to the function that will clean up the object + * when the last reference to the object is released. + * This pointer is required. + * + * Decrement the refcount, and if 0, return 1. Else return 0. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference + * in a lock-free reader-side critical section. + */ +static inline int rcuref_dec_and_test(atomic_t *rcuref) +{ + return atomic_dec_and_test(rcuref); +} + +/* + * cmpxchg is needed on UP too, if deletions to the list/array can happen + * in interrupt context. + */ + +/** + * rcuref_inc_lf - Take reference to an object in a read-side + * critical section protected by RCU. + * @rcuref: reference counter in the object in question. + * + * Try and increment the refcount by 1. The increment might fail if + * the reference counter has been through a 1 to 0 transition and + * is no longer part of the lock-free list. + * Returns non-zero on successful increment and zero otherwise. + */ +static inline int rcuref_inc_lf(atomic_t *rcuref) +{ + int c, old; + c = atomic_read(rcuref); + while (c && (old = cmpxchg(&rcuref->counter, c, c + 1)) != c) + c = old; + return c; +} + +#else /* !__HAVE_ARCH_CMPXCHG */ + +extern spinlock_t __rcuref_hash[]; + +/* + * Use a hash table of locks to protect the reference count + * since cmpxchg is not available in this arch. + */ +#ifdef CONFIG_SMP +#define RCUREF_HASH_SIZE 4 +#define RCUREF_HASH(k) \ + (&__rcuref_hash[(((unsigned long)k)>>8) & (RCUREF_HASH_SIZE-1)]) +#else +#define RCUREF_HASH_SIZE 1 +#define RCUREF_HASH(k) &__rcuref_hash[0] +#endif /* CONFIG_SMP */ + +/** + * rcuref_inc - increment refcount for object. + * @rcuref: reference counter in the object in question. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference in a lock-free + * reader-side critical section. + */ +static inline void rcuref_inc(atomic_t *rcuref) +{ + unsigned long flags; + spin_lock_irqsave(RCUREF_HASH(rcuref), flags); + rcuref->counter += 1; + spin_unlock_irqrestore(RCUREF_HASH(rcuref), flags); +} + +/** + * rcuref_dec - decrement refcount for object. + * @rcuref: reference counter in the object in question. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference in a lock-free + * reader-side critical section. + */ +static inline void rcuref_dec(atomic_t *rcuref) +{ + unsigned long flags; + spin_lock_irqsave(RCUREF_HASH(rcuref), flags); + rcuref->counter -= 1; + spin_unlock_irqrestore(RCUREF_HASH(rcuref), flags); +} + +/** + * rcuref_dec_and_test - decrement refcount for object and test + * @rcuref: reference counter in the object. + * @release: pointer to the function that will clean up the object + * when the last reference to the object is released. + * This pointer is required. + * + * Decrement the refcount, and if 0, return 1. Else return 0. + * + * This should be used only for objects where we use RCU and + * use the rcuref_inc_lf() api to acquire a reference in a lock-free + * reader-side critical section. + */ +static inline int rcuref_dec_and_test(atomic_t *rcuref) +{ + unsigned long flags; + spin_lock_irqsave(RCUREF_HASH(rcuref), flags); + rcuref->counter--; + if (!rcuref->counter) { + spin_unlock_irqrestore(RCUREF_HASH(rcuref), flags); + return 1; + } else { + spin_unlock_irqrestore(RCUREF_HASH(rcuref), flags); + return 0; + } +} + +/** + * rcuref_inc_lf - Take reference to an object of a lock-free collection + * by traversing a lock-free list/array. + * @rcuref: reference counter in the object in question. + * + * Try and increment the refcount by 1. The increment might fail if + * the reference counter has been through a 1 to 0 transition and + * object is no longer part of the lock-free list. + * Returns non-zero on successful increment and zero otherwise. + */ +static inline int rcuref_inc_lf(atomic_t *rcuref) +{ + int ret; + unsigned long flags; + spin_lock_irqsave(RCUREF_HASH(rcuref), flags); + if (rcuref->counter) + ret = rcuref->counter++; + else + ret = 0; + spin_unlock_irqrestore(RCUREF_HASH(rcuref), flags); + return ret; +} + + +#endif /* !__HAVE_ARCH_CMPXCHG */ + +#endif /* __KERNEL__ */ +#endif /* _RCUREF_H_ */ diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index f436993bd590..bef3b6901b76 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -45,6 +45,7 @@ #include #include #include +#include #include /* Definition for rcupdate control block. */ @@ -72,6 +73,19 @@ DEFINE_PER_CPU(struct rcu_data, rcu_bh_data) = { 0L }; static DEFINE_PER_CPU(struct tasklet_struct, rcu_tasklet) = {NULL}; static int maxbatch = 10; +#ifndef __HAVE_ARCH_CMPXCHG +/* + * We use an array of spinlocks for the rcurefs -- similar to ones in sparc + * 32 bit atomic_t implementations, and a hash function similar to that + * for our refcounting needs. + * Can't help multiprocessors which donot have cmpxchg :( + */ + +spinlock_t __rcuref_hash[RCUREF_HASH_SIZE] = { + [0 ... (RCUREF_HASH_SIZE-1)] = SPIN_LOCK_UNLOCKED +}; +#endif + /** * call_rcu - Queue an RCU callback for invocation after a grace period. * @head: structure to be used for queueing the RCU updates. -- cgit v1.2.3 From badf16621c1f9d1ac753be056fce11b43d6e0be5 Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Fri, 9 Sep 2005 13:04:10 -0700 Subject: [PATCH] files: break up files struct In order for the RCU to work, the file table array, sets and their sizes must be updated atomically. Instead of ensuring this through too many memory barriers, we put the arrays and their sizes in a separate structure. This patch takes the first step of putting the file table elements in a separate structure fdtable that is embedded withing files_struct. It also changes all the users to refer to the file table using files_fdtable() macro. Subsequent applciation of RCU becomes easier after this. Signed-off-by: Dipankar Sarma Signed-Off-By: David Howells Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/alpha/kernel/osf_sys.c | 4 ++- arch/ia64/kernel/perfmon.c | 7 ++-- arch/sparc64/solaris/ioctl.c | 8 +++-- drivers/char/tty_io.c | 4 ++- fs/exec.c | 8 +++-- fs/fcntl.c | 47 +++++++++++++++---------- fs/file.c | 42 ++++++++++++++--------- fs/locks.c | 8 +++-- fs/open.c | 41 +++++++++++++--------- fs/proc/array.c | 5 ++- fs/proc/base.c | 4 ++- fs/select.c | 12 +++++-- include/linux/file.h | 23 ++++++++----- include/linux/init_task.h | 13 ++++--- kernel/exit.c | 21 +++++++----- kernel/fork.c | 82 +++++++++++++++++++++++++++----------------- security/selinux/hooks.c | 6 ++-- 17 files changed, 211 insertions(+), 124 deletions(-) (limited to 'include/linux') diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c index 167fd89f8707..2b034182a0ca 100644 --- a/arch/alpha/kernel/osf_sys.c +++ b/arch/alpha/kernel/osf_sys.c @@ -974,6 +974,7 @@ osf_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, size_t size; long timeout; int ret = -EINVAL; + struct fdtable *fdt; timeout = MAX_SCHEDULE_TIMEOUT; if (tvp) { @@ -995,7 +996,8 @@ osf_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, } } - if (n < 0 || n > current->files->max_fdset) + fdt = files_fdtable(current->files); + if (n < 0 || n > fdt->max_fdset) goto out_nofds; /* diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c index f1201ac8a116..4ad97b3b39dc 100644 --- a/arch/ia64/kernel/perfmon.c +++ b/arch/ia64/kernel/perfmon.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include @@ -2217,15 +2218,17 @@ static void pfm_free_fd(int fd, struct file *file) { struct files_struct *files = current->files; + struct fdtable *fdt = files_fdtable(files); /* * there ie no fd_uninstall(), so we do it here */ spin_lock(&files->file_lock); - files->fd[fd] = NULL; + rcu_assign_pointer(fdt->fd[fd], NULL); spin_unlock(&files->file_lock); - if (file) put_filp(file); + if (file) + put_filp(file); put_unused_fd(fd); } diff --git a/arch/sparc64/solaris/ioctl.c b/arch/sparc64/solaris/ioctl.c index cac0a1cf0050..374766455f5e 100644 --- a/arch/sparc64/solaris/ioctl.c +++ b/arch/sparc64/solaris/ioctl.c @@ -293,11 +293,13 @@ static struct module_info { static inline int solaris_sockmod(unsigned int fd, unsigned int cmd, u32 arg) { struct inode *ino; + struct fdtable *fdt; /* I wonder which of these tests are superfluous... --patrik */ spin_lock(¤t->files->file_lock); - if (! current->files->fd[fd] || - ! current->files->fd[fd]->f_dentry || - ! (ino = current->files->fd[fd]->f_dentry->d_inode) || + fdt = files_fdtable(current->files); + if (! fdt->fd[fd] || + ! fdt->fd[fd]->f_dentry || + ! (ino = fdt->fd[fd]->f_dentry->d_inode) || ! S_ISSOCK(ino->i_mode)) { spin_unlock(¤t->files->file_lock); return TBADF; diff --git a/drivers/char/tty_io.c b/drivers/char/tty_io.c index 6a56ae4f7725..0bfc7af68917 100644 --- a/drivers/char/tty_io.c +++ b/drivers/char/tty_io.c @@ -2454,6 +2454,7 @@ static void __do_SAK(void *arg) int i; struct file *filp; struct tty_ldisc *disc; + struct fdtable *fdt; if (!tty) return; @@ -2480,7 +2481,8 @@ static void __do_SAK(void *arg) task_lock(p); if (p->files) { spin_lock(&p->files->file_lock); - for (i=0; i < p->files->max_fds; i++) { + fdt = files_fdtable(p->files); + for (i=0; i < fdt->max_fds; i++) { filp = fcheck_files(p->files, i); if (!filp) continue; diff --git a/fs/exec.c b/fs/exec.c index 222ab1c572d8..14dd03907ccb 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -798,6 +798,7 @@ no_thread_group: static inline void flush_old_files(struct files_struct * files) { long j = -1; + struct fdtable *fdt; spin_lock(&files->file_lock); for (;;) { @@ -805,12 +806,13 @@ static inline void flush_old_files(struct files_struct * files) j++; i = j * __NFDBITS; - if (i >= files->max_fds || i >= files->max_fdset) + fdt = files_fdtable(files); + if (i >= fdt->max_fds || i >= fdt->max_fdset) break; - set = files->close_on_exec->fds_bits[j]; + set = fdt->close_on_exec->fds_bits[j]; if (!set) continue; - files->close_on_exec->fds_bits[j] = 0; + fdt->close_on_exec->fds_bits[j] = 0; spin_unlock(&files->file_lock); for ( ; set ; i++,set >>= 1) { if (set & 1) { diff --git a/fs/fcntl.c b/fs/fcntl.c index 6fbc9d8fcc36..bfecc6238083 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -24,20 +24,24 @@ void fastcall set_close_on_exec(unsigned int fd, int flag) { struct files_struct *files = current->files; + struct fdtable *fdt; spin_lock(&files->file_lock); + fdt = files_fdtable(files); if (flag) - FD_SET(fd, files->close_on_exec); + FD_SET(fd, fdt->close_on_exec); else - FD_CLR(fd, files->close_on_exec); + FD_CLR(fd, fdt->close_on_exec); spin_unlock(&files->file_lock); } static inline int get_close_on_exec(unsigned int fd) { struct files_struct *files = current->files; + struct fdtable *fdt; int res; spin_lock(&files->file_lock); - res = FD_ISSET(fd, files->close_on_exec); + fdt = files_fdtable(files); + res = FD_ISSET(fd, fdt->close_on_exec); spin_unlock(&files->file_lock); return res; } @@ -54,24 +58,26 @@ static int locate_fd(struct files_struct *files, unsigned int newfd; unsigned int start; int error; + struct fdtable *fdt; error = -EINVAL; if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; + fdt = files_fdtable(files); repeat: /* * Someone might have closed fd's in the range - * orig_start..files->next_fd + * orig_start..fdt->next_fd */ start = orig_start; - if (start < files->next_fd) - start = files->next_fd; + if (start < fdt->next_fd) + start = fdt->next_fd; newfd = start; - if (start < files->max_fdset) { - newfd = find_next_zero_bit(files->open_fds->fds_bits, - files->max_fdset, start); + if (start < fdt->max_fdset) { + newfd = find_next_zero_bit(fdt->open_fds->fds_bits, + fdt->max_fdset, start); } error = -EMFILE; @@ -89,8 +95,8 @@ repeat: if (error) goto repeat; - if (start <= files->next_fd) - files->next_fd = newfd + 1; + if (start <= fdt->next_fd) + fdt->next_fd = newfd + 1; error = newfd; @@ -101,13 +107,16 @@ out: static int dupfd(struct file *file, unsigned int start) { struct files_struct * files = current->files; + struct fdtable *fdt; int fd; spin_lock(&files->file_lock); fd = locate_fd(files, file, start); if (fd >= 0) { - FD_SET(fd, files->open_fds); - FD_CLR(fd, files->close_on_exec); + /* locate_fd() may have expanded fdtable, load the ptr */ + fdt = files_fdtable(files); + FD_SET(fd, fdt->open_fds); + FD_CLR(fd, fdt->close_on_exec); spin_unlock(&files->file_lock); fd_install(fd, file); } else { @@ -123,6 +132,7 @@ asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) int err = -EBADF; struct file * file, *tofree; struct files_struct * files = current->files; + struct fdtable *fdt; spin_lock(&files->file_lock); if (!(file = fcheck(oldfd))) @@ -148,13 +158,14 @@ asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) /* Yes. It's a race. In user space. Nothing sane to do */ err = -EBUSY; - tofree = files->fd[newfd]; - if (!tofree && FD_ISSET(newfd, files->open_fds)) + fdt = files_fdtable(files); + tofree = fdt->fd[newfd]; + if (!tofree && FD_ISSET(newfd, fdt->open_fds)) goto out_fput; - files->fd[newfd] = file; - FD_SET(newfd, files->open_fds); - FD_CLR(newfd, files->close_on_exec); + fdt->fd[newfd] = file; + FD_SET(newfd, fdt->open_fds); + FD_CLR(newfd, fdt->close_on_exec); spin_unlock(&files->file_lock); if (tofree) diff --git a/fs/file.c b/fs/file.c index 92b5f25985d2..f5926ce73f37 100644 --- a/fs/file.c +++ b/fs/file.c @@ -59,13 +59,15 @@ static int expand_fd_array(struct files_struct *files, int nr) { struct file **new_fds; int error, nfds; + struct fdtable *fdt; error = -EMFILE; - if (files->max_fds >= NR_OPEN || nr >= NR_OPEN) + fdt = files_fdtable(files); + if (fdt->max_fds >= NR_OPEN || nr >= NR_OPEN) goto out; - nfds = files->max_fds; + nfds = fdt->max_fds; spin_unlock(&files->file_lock); /* @@ -95,13 +97,14 @@ static int expand_fd_array(struct files_struct *files, int nr) goto out; /* Copy the existing array and install the new pointer */ + fdt = files_fdtable(files); - if (nfds > files->max_fds) { + if (nfds > fdt->max_fds) { struct file **old_fds; int i; - old_fds = xchg(&files->fd, new_fds); - i = xchg(&files->max_fds, nfds); + old_fds = xchg(&fdt->fd, new_fds); + i = xchg(&fdt->max_fds, nfds); /* Don't copy/clear the array if we are creating a new fd array for fork() */ @@ -164,12 +167,14 @@ static int expand_fdset(struct files_struct *files, int nr) { fd_set *new_openset = NULL, *new_execset = NULL; int error, nfds = 0; + struct fdtable *fdt; error = -EMFILE; - if (files->max_fdset >= NR_OPEN || nr >= NR_OPEN) + fdt = files_fdtable(files); + if (fdt->max_fdset >= NR_OPEN || nr >= NR_OPEN) goto out; - nfds = files->max_fdset; + nfds = fdt->max_fdset; spin_unlock(&files->file_lock); /* Expand to the max in easy steps */ @@ -193,24 +198,25 @@ static int expand_fdset(struct files_struct *files, int nr) error = 0; /* Copy the existing tables and install the new pointers */ - if (nfds > files->max_fdset) { - int i = files->max_fdset / (sizeof(unsigned long) * 8); - int count = (nfds - files->max_fdset) / 8; + fdt = files_fdtable(files); + if (nfds > fdt->max_fdset) { + int i = fdt->max_fdset / (sizeof(unsigned long) * 8); + int count = (nfds - fdt->max_fdset) / 8; /* * Don't copy the entire array if the current fdset is * not yet initialised. */ if (i) { - memcpy (new_openset, files->open_fds, files->max_fdset/8); - memcpy (new_execset, files->close_on_exec, files->max_fdset/8); + memcpy (new_openset, fdt->open_fds, fdt->max_fdset/8); + memcpy (new_execset, fdt->close_on_exec, fdt->max_fdset/8); memset (&new_openset->fds_bits[i], 0, count); memset (&new_execset->fds_bits[i], 0, count); } - nfds = xchg(&files->max_fdset, nfds); - new_openset = xchg(&files->open_fds, new_openset); - new_execset = xchg(&files->close_on_exec, new_execset); + nfds = xchg(&fdt->max_fdset, nfds); + new_openset = xchg(&fdt->open_fds, new_openset); + new_execset = xchg(&fdt->close_on_exec, new_execset); spin_unlock(&files->file_lock); free_fdset (new_openset, nfds); free_fdset (new_execset, nfds); @@ -237,13 +243,15 @@ out: int expand_files(struct files_struct *files, int nr) { int err, expand = 0; + struct fdtable *fdt; - if (nr >= files->max_fdset) { + fdt = files_fdtable(files); + if (nr >= fdt->max_fdset) { expand = 1; if ((err = expand_fdset(files, nr))) goto out; } - if (nr >= files->max_fds) { + if (nr >= fdt->max_fds) { expand = 1; if ((err = expand_fd_array(files, nr))) goto out; diff --git a/fs/locks.c b/fs/locks.c index 11956b6179ff..c2c09b4798d6 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -2198,21 +2198,23 @@ void steal_locks(fl_owner_t from) { struct files_struct *files = current->files; int i, j; + struct fdtable *fdt; if (from == files) return; lock_kernel(); j = 0; + fdt = files_fdtable(files); for (;;) { unsigned long set; i = j * __NFDBITS; - if (i >= files->max_fdset || i >= files->max_fds) + if (i >= fdt->max_fdset || i >= fdt->max_fds) break; - set = files->open_fds->fds_bits[j++]; + set = fdt->open_fds->fds_bits[j++]; while (set) { if (set & 1) { - struct file *file = files->fd[i]; + struct file *file = fdt->fd[i]; if (file) __steal_locks(file, from); } diff --git a/fs/open.c b/fs/open.c index 4ee2dcc31c28..b6542516a0ca 100644 --- a/fs/open.c +++ b/fs/open.c @@ -842,14 +842,16 @@ int get_unused_fd(void) { struct files_struct * files = current->files; int fd, error; + struct fdtable *fdt; error = -EMFILE; spin_lock(&files->file_lock); repeat: - fd = find_next_zero_bit(files->open_fds->fds_bits, - files->max_fdset, - files->next_fd); + fdt = files_fdtable(files); + fd = find_next_zero_bit(fdt->open_fds->fds_bits, + fdt->max_fdset, + fdt->next_fd); /* * N.B. For clone tasks sharing a files structure, this test @@ -872,14 +874,14 @@ repeat: goto repeat; } - FD_SET(fd, files->open_fds); - FD_CLR(fd, files->close_on_exec); - files->next_fd = fd + 1; + FD_SET(fd, fdt->open_fds); + FD_CLR(fd, fdt->close_on_exec); + fdt->next_fd = fd + 1; #if 1 /* Sanity check */ - if (files->fd[fd] != NULL) { + if (fdt->fd[fd] != NULL) { printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", fd); - files->fd[fd] = NULL; + fdt->fd[fd] = NULL; } #endif error = fd; @@ -893,9 +895,10 @@ EXPORT_SYMBOL(get_unused_fd); static inline void __put_unused_fd(struct files_struct *files, unsigned int fd) { - __FD_CLR(fd, files->open_fds); - if (fd < files->next_fd) - files->next_fd = fd; + struct fdtable *fdt = files_fdtable(files); + __FD_CLR(fd, fdt->open_fds); + if (fd < fdt->next_fd) + fdt->next_fd = fd; } void fastcall put_unused_fd(unsigned int fd) @@ -924,10 +927,12 @@ EXPORT_SYMBOL(put_unused_fd); void fastcall fd_install(unsigned int fd, struct file * file) { struct files_struct *files = current->files; + struct fdtable *fdt; spin_lock(&files->file_lock); - if (unlikely(files->fd[fd] != NULL)) + fdt = files_fdtable(files); + if (unlikely(fdt->fd[fd] != NULL)) BUG(); - files->fd[fd] = file; + fdt->fd[fd] = file; spin_unlock(&files->file_lock); } @@ -1010,15 +1015,17 @@ asmlinkage long sys_close(unsigned int fd) { struct file * filp; struct files_struct *files = current->files; + struct fdtable *fdt; spin_lock(&files->file_lock); - if (fd >= files->max_fds) + fdt = files_fdtable(files); + if (fd >= fdt->max_fds) goto out_unlock; - filp = files->fd[fd]; + filp = fdt->fd[fd]; if (!filp) goto out_unlock; - files->fd[fd] = NULL; - FD_CLR(fd, files->close_on_exec); + fdt->fd[fd] = NULL; + FD_CLR(fd, fdt->close_on_exec); __put_unused_fd(files, fd); spin_unlock(&files->file_lock); return filp_close(filp, files); diff --git a/fs/proc/array.c b/fs/proc/array.c index 37668fe998ad..d88d518d30f6 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -159,6 +159,7 @@ static inline char * task_state(struct task_struct *p, char *buffer) { struct group_info *group_info; int g; + struct fdtable *fdt = NULL; read_lock(&tasklist_lock); buffer += sprintf(buffer, @@ -179,10 +180,12 @@ static inline char * task_state(struct task_struct *p, char *buffer) p->gid, p->egid, p->sgid, p->fsgid); read_unlock(&tasklist_lock); task_lock(p); + if (p->files) + fdt = files_fdtable(p->files); buffer += sprintf(buffer, "FDSize:\t%d\n" "Groups:\t", - p->files ? p->files->max_fds : 0); + fdt ? fdt->max_fds : 0); group_info = p->group_info; get_group_info(group_info); diff --git a/fs/proc/base.c b/fs/proc/base.c index 84751f3f52d5..d0087a0b024b 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -1039,6 +1039,7 @@ static int proc_readfd(struct file * filp, void * dirent, filldir_t filldir) int retval; char buf[NUMBUF]; struct files_struct * files; + struct fdtable *fdt; retval = -ENOENT; if (!pid_alive(p)) @@ -1062,8 +1063,9 @@ static int proc_readfd(struct file * filp, void * dirent, filldir_t filldir) if (!files) goto out; spin_lock(&files->file_lock); + fdt = files_fdtable(files); for (fd = filp->f_pos-2; - fd < files->max_fds; + fd < fdt->max_fds; fd++, filp->f_pos++) { unsigned int i,j; diff --git a/fs/select.c b/fs/select.c index b80e7eb0ac0d..2e56325c73c4 100644 --- a/fs/select.c +++ b/fs/select.c @@ -132,11 +132,13 @@ static int max_select_fd(unsigned long n, fd_set_bits *fds) unsigned long *open_fds; unsigned long set; int max; + struct fdtable *fdt; /* handle last in-complete long-word first */ set = ~(~0UL << (n & (__NFDBITS-1))); n /= __NFDBITS; - open_fds = current->files->open_fds->fds_bits+n; + fdt = files_fdtable(current->files); + open_fds = fdt->open_fds->fds_bits+n; max = 0; if (set) { set &= BITS(fds, n); @@ -299,6 +301,7 @@ sys_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, s char *bits; long timeout; int ret, size, max_fdset; + struct fdtable *fdt; timeout = MAX_SCHEDULE_TIMEOUT; if (tvp) { @@ -326,7 +329,8 @@ sys_select(int n, fd_set __user *inp, fd_set __user *outp, fd_set __user *exp, s goto out_nofds; /* max_fdset can increase, so grab it once to avoid race */ - max_fdset = current->files->max_fdset; + fdt = files_fdtable(current->files); + max_fdset = fdt->max_fdset; if (n > max_fdset) n = max_fdset; @@ -464,9 +468,11 @@ asmlinkage long sys_poll(struct pollfd __user * ufds, unsigned int nfds, long ti unsigned int i; struct poll_list *head; struct poll_list *walk; + struct fdtable *fdt; /* Do a sanity check on nfds ... */ - if (nfds > current->files->max_fdset && nfds > OPEN_MAX) + fdt = files_fdtable(current->files); + if (nfds > fdt->max_fdset && nfds > OPEN_MAX) return -EINVAL; if (timeout) { diff --git a/include/linux/file.h b/include/linux/file.h index 5206beb9a80e..db372230848e 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -16,23 +16,29 @@ */ #define NR_OPEN_DEFAULT BITS_PER_LONG +struct fdtable { + unsigned int max_fds; + int max_fdset; + int next_fd; + struct file ** fd; /* current fd array */ + fd_set *close_on_exec; + fd_set *open_fds; +}; + /* * Open file table structure */ struct files_struct { atomic_t count; spinlock_t file_lock; /* Protects all the below members. Nests inside tsk->alloc_lock */ - int max_fds; - int max_fdset; - int next_fd; - struct file ** fd; /* current fd array */ - fd_set *close_on_exec; - fd_set *open_fds; + struct fdtable fdtab; fd_set close_on_exec_init; fd_set open_fds_init; struct file * fd_array[NR_OPEN_DEFAULT]; }; +#define files_fdtable(files) (&(files)->fdtab) + extern void FASTCALL(__fput(struct file *)); extern void FASTCALL(fput(struct file *)); @@ -63,9 +69,10 @@ extern int expand_files(struct files_struct *, int nr); static inline struct file * fcheck_files(struct files_struct *files, unsigned int fd) { struct file * file = NULL; + struct fdtable *fdt = files_fdtable(files); - if (fd < files->max_fds) - file = files->fd[fd]; + if (fd < fdt->max_fds) + file = fdt->fd[fd]; return file; } diff --git a/include/linux/init_task.h b/include/linux/init_task.h index c727c195a91a..94aefa54a1b5 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -3,16 +3,21 @@ #include -#define INIT_FILES \ -{ \ - .count = ATOMIC_INIT(1), \ - .file_lock = SPIN_LOCK_UNLOCKED, \ +#define INIT_FDTABLE \ +{ \ .max_fds = NR_OPEN_DEFAULT, \ .max_fdset = __FD_SETSIZE, \ .next_fd = 0, \ .fd = &init_files.fd_array[0], \ .close_on_exec = &init_files.close_on_exec_init, \ .open_fds = &init_files.open_fds_init, \ +} + +#define INIT_FILES \ +{ \ + .count = ATOMIC_INIT(1), \ + .file_lock = SPIN_LOCK_UNLOCKED, \ + .fdtab = INIT_FDTABLE, \ .close_on_exec_init = { { 0, } }, \ .open_fds_init = { { 0, } }, \ .fd_array = { NULL, } \ diff --git a/kernel/exit.c b/kernel/exit.c index 5b0fb9f09f21..83beb1e93b18 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -368,17 +368,19 @@ EXPORT_SYMBOL(daemonize); static inline void close_files(struct files_struct * files) { int i, j; + struct fdtable *fdt; j = 0; + fdt = files_fdtable(files); for (;;) { unsigned long set; i = j * __NFDBITS; - if (i >= files->max_fdset || i >= files->max_fds) + if (i >= fdt->max_fdset || i >= fdt->max_fds) break; - set = files->open_fds->fds_bits[j++]; + set = fdt->open_fds->fds_bits[j++]; while (set) { if (set & 1) { - struct file * file = xchg(&files->fd[i], NULL); + struct file * file = xchg(&fdt->fd[i], NULL); if (file) filp_close(file, files); } @@ -403,16 +405,19 @@ struct files_struct *get_files_struct(struct task_struct *task) void fastcall put_files_struct(struct files_struct *files) { + struct fdtable *fdt; + if (atomic_dec_and_test(&files->count)) { close_files(files); /* * Free the fd and fdset arrays if we expanded them. */ - if (files->fd != &files->fd_array[0]) - free_fd_array(files->fd, files->max_fds); - if (files->max_fdset > __FD_SETSIZE) { - free_fdset(files->open_fds, files->max_fdset); - free_fdset(files->close_on_exec, files->max_fdset); + fdt = files_fdtable(files); + if (fdt->fd != &files->fd_array[0]) + free_fd_array(fdt->fd, fdt->max_fds); + if (fdt->max_fdset > __FD_SETSIZE) { + free_fdset(fdt->open_fds, fdt->max_fdset); + free_fdset(fdt->close_on_exec, fdt->max_fdset); } kmem_cache_free(files_cachep, files); } diff --git a/kernel/fork.c b/kernel/fork.c index b25802065031..ecc694debb50 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -568,21 +568,47 @@ static inline int copy_fs(unsigned long clone_flags, struct task_struct * tsk) static int count_open_files(struct files_struct *files, int size) { int i; + struct fdtable *fdt; /* Find the last open fd */ + fdt = files_fdtable(files); for (i = size/(8*sizeof(long)); i > 0; ) { - if (files->open_fds->fds_bits[--i]) + if (fdt->open_fds->fds_bits[--i]) break; } i = (i+1) * 8 * sizeof(long); return i; } +static struct files_struct *alloc_files(void) +{ + struct files_struct *newf; + struct fdtable *fdt; + + newf = kmem_cache_alloc(files_cachep, SLAB_KERNEL); + if (!newf) + goto out; + + atomic_set(&newf->count, 1); + + spin_lock_init(&newf->file_lock); + fdt = files_fdtable(newf); + fdt->next_fd = 0; + fdt->max_fds = NR_OPEN_DEFAULT; + fdt->max_fdset = __FD_SETSIZE; + fdt->close_on_exec = &newf->close_on_exec_init; + fdt->open_fds = &newf->open_fds_init; + fdt->fd = &newf->fd_array[0]; +out: + return newf; +} + static int copy_files(unsigned long clone_flags, struct task_struct * tsk) { struct files_struct *oldf, *newf; struct file **old_fds, **new_fds; int open_files, size, i, error = 0, expand; + struct fdtable *old_fdt, *new_fdt; /* * A background process may not have any files ... @@ -603,35 +629,27 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) */ tsk->files = NULL; error = -ENOMEM; - newf = kmem_cache_alloc(files_cachep, SLAB_KERNEL); - if (!newf) + newf = alloc_files(); + if (!newf) goto out; - atomic_set(&newf->count, 1); - - spin_lock_init(&newf->file_lock); - newf->next_fd = 0; - newf->max_fds = NR_OPEN_DEFAULT; - newf->max_fdset = __FD_SETSIZE; - newf->close_on_exec = &newf->close_on_exec_init; - newf->open_fds = &newf->open_fds_init; - newf->fd = &newf->fd_array[0]; - spin_lock(&oldf->file_lock); - - open_files = count_open_files(oldf, oldf->max_fdset); + old_fdt = files_fdtable(oldf); + new_fdt = files_fdtable(newf); + size = old_fdt->max_fdset; + open_files = count_open_files(oldf, old_fdt->max_fdset); expand = 0; /* * Check whether we need to allocate a larger fd array or fd set. * Note: we're not a clone task, so the open count won't change. */ - if (open_files > newf->max_fdset) { - newf->max_fdset = 0; + if (open_files > new_fdt->max_fdset) { + new_fdt->max_fdset = 0; expand = 1; } - if (open_files > newf->max_fds) { - newf->max_fds = 0; + if (open_files > new_fdt->max_fds) { + new_fdt->max_fds = 0; expand = 1; } @@ -646,11 +664,11 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) spin_lock(&oldf->file_lock); } - old_fds = oldf->fd; - new_fds = newf->fd; + old_fds = old_fdt->fd; + new_fds = new_fdt->fd; - memcpy(newf->open_fds->fds_bits, oldf->open_fds->fds_bits, open_files/8); - memcpy(newf->close_on_exec->fds_bits, oldf->close_on_exec->fds_bits, open_files/8); + memcpy(new_fdt->open_fds->fds_bits, old_fdt->open_fds->fds_bits, open_files/8); + memcpy(new_fdt->close_on_exec->fds_bits, old_fdt->close_on_exec->fds_bits, open_files/8); for (i = open_files; i != 0; i--) { struct file *f = *old_fds++; @@ -663,24 +681,24 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) * is partway through open(). So make sure that this * fd is available to the new process. */ - FD_CLR(open_files - i, newf->open_fds); + FD_CLR(open_files - i, new_fdt->open_fds); } *new_fds++ = f; } spin_unlock(&oldf->file_lock); /* compute the remainder to be cleared */ - size = (newf->max_fds - open_files) * sizeof(struct file *); + size = (new_fdt->max_fds - open_files) * sizeof(struct file *); /* This is long word aligned thus could use a optimized version */ memset(new_fds, 0, size); - if (newf->max_fdset > open_files) { - int left = (newf->max_fdset-open_files)/8; + if (new_fdt->max_fdset > open_files) { + int left = (new_fdt->max_fdset-open_files)/8; int start = open_files / (8 * sizeof(unsigned long)); - memset(&newf->open_fds->fds_bits[start], 0, left); - memset(&newf->close_on_exec->fds_bits[start], 0, left); + memset(&new_fdt->open_fds->fds_bits[start], 0, left); + memset(&new_fdt->close_on_exec->fds_bits[start], 0, left); } tsk->files = newf; @@ -689,9 +707,9 @@ out: return error; out_release: - free_fdset (newf->close_on_exec, newf->max_fdset); - free_fdset (newf->open_fds, newf->max_fdset); - free_fd_array(newf->fd, newf->max_fds); + free_fdset (new_fdt->close_on_exec, new_fdt->max_fdset); + free_fdset (new_fdt->open_fds, new_fdt->max_fdset); + free_fd_array(new_fdt->fd, new_fdt->max_fds); kmem_cache_free(files_cachep, newf); goto out; } diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 3f0b533be92c..acb5a495a902 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -1594,6 +1594,7 @@ static inline void flush_unauthorized_files(struct files_struct * files) struct avc_audit_data ad; struct file *file, *devnull = NULL; struct tty_struct *tty = current->signal->tty; + struct fdtable *fdt; long j = -1; if (tty) { @@ -1627,9 +1628,10 @@ static inline void flush_unauthorized_files(struct files_struct * files) j++; i = j * __NFDBITS; - if (i >= files->max_fds || i >= files->max_fdset) + fdt = files_fdtable(files); + if (i >= fdt->max_fds || i >= fdt->max_fdset) break; - set = files->open_fds->fds_bits[j]; + set = fdt->open_fds->fds_bits[j]; if (!set) continue; spin_unlock(&files->file_lock); -- cgit v1.2.3 From ab2af1f5005069321c5d130f09cce577b03f43ef Mon Sep 17 00:00:00 2001 From: Dipankar Sarma Date: Fri, 9 Sep 2005 13:04:13 -0700 Subject: [PATCH] files: files struct with RCU Patch to eliminate struct files_struct.file_lock spinlock on the reader side and use rcu refcounting rcuref_xxx api for the f_count refcounter. The updates to the fdtable are done by allocating a new fdtable structure and setting files->fdt to point to the new structure. The fdtable structure is protected by RCU thereby allowing lock-free lookup. For fd arrays/sets that are vmalloced, we use keventd to free them since RCU callbacks can't sleep. A global list of fdtable to be freed is not scalable, so we use a per-cpu list. If keventd is already handling the current cpu's work, we use a timer to defer queueing of that work. Since the last publication, this patch has been re-written to avoid using explicit memory barriers and use rcu_assign_pointer(), rcu_dereference() premitives instead. This required that the fd information is kept in a separate structure (fdtable) and updated atomically. Signed-off-by: Dipankar Sarma Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/aio.c | 3 +- fs/fcntl.c | 13 +- fs/file.c | 389 ++++++++++++++++++++++++++++++---------------- fs/file_table.c | 40 +++-- fs/open.c | 8 +- include/linux/file.h | 11 +- include/linux/fs.h | 4 +- include/linux/init_task.h | 5 + kernel/exit.c | 15 +- kernel/fork.c | 23 ++- 10 files changed, 345 insertions(+), 166 deletions(-) (limited to 'include/linux') diff --git a/fs/aio.c b/fs/aio.c index 201c1847fa07..38f62680fd63 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -499,7 +500,7 @@ static int __aio_put_req(struct kioctx *ctx, struct kiocb *req) /* Must be done under the lock to serialise against cancellation. * Call this aio_fput as it duplicates fput via the fput_work. */ - if (unlikely(atomic_dec_and_test(&req->ki_filp->f_count))) { + if (unlikely(rcuref_dec_and_test(&req->ki_filp->f_count))) { get_ioctx(ctx); spin_lock(&fput_lock); list_add(&req->ki_list, &fput_head); diff --git a/fs/fcntl.c b/fs/fcntl.c index bfecc6238083..d2f3ed8acd93 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include @@ -64,8 +65,8 @@ static int locate_fd(struct files_struct *files, if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; - fdt = files_fdtable(files); repeat: + fdt = files_fdtable(files); /* * Someone might have closed fd's in the range * orig_start..fdt->next_fd @@ -95,9 +96,15 @@ repeat: if (error) goto repeat; + /* + * We reacquired files_lock, so we are safe as long as + * we reacquire the fdtable pointer and use it while holding + * the lock, no one can free it during that time. + */ + fdt = files_fdtable(files); if (start <= fdt->next_fd) fdt->next_fd = newfd + 1; - + error = newfd; out: @@ -163,7 +170,7 @@ asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd) if (!tofree && FD_ISSET(newfd, fdt->open_fds)) goto out_fput; - fdt->fd[newfd] = file; + rcu_assign_pointer(fdt->fd[newfd], file); FD_SET(newfd, fdt->open_fds); FD_CLR(newfd, fdt->close_on_exec); spin_unlock(&files->file_lock); diff --git a/fs/file.c b/fs/file.c index f5926ce73f37..2127a7b9dc3a 100644 --- a/fs/file.c +++ b/fs/file.c @@ -13,6 +13,25 @@ #include #include #include +#include +#include +#include +#include + +struct fdtable_defer { + spinlock_t lock; + struct work_struct wq; + struct timer_list timer; + struct fdtable *next; +}; + +/* + * We use this list to defer free fdtables that have vmalloced + * sets/arrays. By keeping a per-cpu list, we avoid having to embed + * the work_struct in fdtable itself which avoids a 64 byte (i386) increase in + * this per-task structure. + */ +static DEFINE_PER_CPU(struct fdtable_defer, fdtable_defer_list); /* @@ -48,85 +67,143 @@ void free_fd_array(struct file **array, int num) vfree(array); } -/* - * Expand the fd array in the files_struct. Called with the files - * spinlock held for write. - */ - -static int expand_fd_array(struct files_struct *files, int nr) - __releases(files->file_lock) - __acquires(files->file_lock) +static void __free_fdtable(struct fdtable *fdt) { - struct file **new_fds; - int error, nfds; - struct fdtable *fdt; + int fdset_size, fdarray_size; - - error = -EMFILE; - fdt = files_fdtable(files); - if (fdt->max_fds >= NR_OPEN || nr >= NR_OPEN) - goto out; + fdset_size = fdt->max_fdset / 8; + fdarray_size = fdt->max_fds * sizeof(struct file *); + free_fdset(fdt->open_fds, fdset_size); + free_fdset(fdt->close_on_exec, fdset_size); + free_fd_array(fdt->fd, fdarray_size); + kfree(fdt); +} - nfds = fdt->max_fds; - spin_unlock(&files->file_lock); +static void fdtable_timer(unsigned long data) +{ + struct fdtable_defer *fddef = (struct fdtable_defer *)data; - /* - * Expand to the max in easy steps, and keep expanding it until - * we have enough for the requested fd array size. + spin_lock(&fddef->lock); + /* + * If someone already emptied the queue return. */ + if (!fddef->next) + goto out; + if (!schedule_work(&fddef->wq)) + mod_timer(&fddef->timer, 5); +out: + spin_unlock(&fddef->lock); +} - do { -#if NR_OPEN_DEFAULT < 256 - if (nfds < 256) - nfds = 256; - else -#endif - if (nfds < (PAGE_SIZE / sizeof(struct file *))) - nfds = PAGE_SIZE / sizeof(struct file *); - else { - nfds = nfds * 2; - if (nfds > NR_OPEN) - nfds = NR_OPEN; - } - } while (nfds <= nr); +static void free_fdtable_work(struct fdtable_defer *f) +{ + struct fdtable *fdt; - error = -ENOMEM; - new_fds = alloc_fd_array(nfds); - spin_lock(&files->file_lock); - if (!new_fds) - goto out; + spin_lock_bh(&f->lock); + fdt = f->next; + f->next = NULL; + spin_unlock_bh(&f->lock); + while(fdt) { + struct fdtable *next = fdt->next; + __free_fdtable(fdt); + fdt = next; + } +} - /* Copy the existing array and install the new pointer */ - fdt = files_fdtable(files); +static void free_fdtable_rcu(struct rcu_head *rcu) +{ + struct fdtable *fdt = container_of(rcu, struct fdtable, rcu); + int fdset_size, fdarray_size; + struct fdtable_defer *fddef; - if (nfds > fdt->max_fds) { - struct file **old_fds; - int i; - - old_fds = xchg(&fdt->fd, new_fds); - i = xchg(&fdt->max_fds, nfds); - - /* Don't copy/clear the array if we are creating a new - fd array for fork() */ - if (i) { - memcpy(new_fds, old_fds, i * sizeof(struct file *)); - /* clear the remainder of the array */ - memset(&new_fds[i], 0, - (nfds-i) * sizeof(struct file *)); - - spin_unlock(&files->file_lock); - free_fd_array(old_fds, i); - spin_lock(&files->file_lock); - } + BUG_ON(!fdt); + fdset_size = fdt->max_fdset / 8; + fdarray_size = fdt->max_fds * sizeof(struct file *); + + if (fdt->free_files) { + /* + * The this fdtable was embedded in the files structure + * and the files structure itself was getting destroyed. + * It is now safe to free the files structure. + */ + kmem_cache_free(files_cachep, fdt->free_files); + return; + } + if (fdt->max_fdset <= __FD_SETSIZE && fdt->max_fds <= NR_OPEN_DEFAULT) { + /* + * The fdtable was embedded + */ + return; + } + if (fdset_size <= PAGE_SIZE && fdarray_size <= PAGE_SIZE) { + kfree(fdt->open_fds); + kfree(fdt->close_on_exec); + kfree(fdt->fd); + kfree(fdt); } else { - /* Somebody expanded the array while we slept ... */ - spin_unlock(&files->file_lock); - free_fd_array(new_fds, nfds); - spin_lock(&files->file_lock); + fddef = &get_cpu_var(fdtable_defer_list); + spin_lock(&fddef->lock); + fdt->next = fddef->next; + fddef->next = fdt; + /* + * vmallocs are handled from the workqueue context. + * If the per-cpu workqueue is running, then we + * defer work scheduling through a timer. + */ + if (!schedule_work(&fddef->wq)) + mod_timer(&fddef->timer, 5); + spin_unlock(&fddef->lock); + put_cpu_var(fdtable_defer_list); } - error = 0; -out: - return error; +} + +void free_fdtable(struct fdtable *fdt) +{ + if (fdt->free_files || fdt->max_fdset > __FD_SETSIZE || + fdt->max_fds > NR_OPEN_DEFAULT) + call_rcu(&fdt->rcu, free_fdtable_rcu); +} + +/* + * Expand the fdset in the files_struct. Called with the files spinlock + * held for write. + */ +static void copy_fdtable(struct fdtable *nfdt, struct fdtable *fdt) +{ + int i; + int count; + + BUG_ON(nfdt->max_fdset < fdt->max_fdset); + BUG_ON(nfdt->max_fds < fdt->max_fds); + /* Copy the existing tables and install the new pointers */ + + i = fdt->max_fdset / (sizeof(unsigned long) * 8); + count = (nfdt->max_fdset - fdt->max_fdset) / 8; + + /* + * Don't copy the entire array if the current fdset is + * not yet initialised. + */ + if (i) { + memcpy (nfdt->open_fds, fdt->open_fds, + fdt->max_fdset/8); + memcpy (nfdt->close_on_exec, fdt->close_on_exec, + fdt->max_fdset/8); + memset (&nfdt->open_fds->fds_bits[i], 0, count); + memset (&nfdt->close_on_exec->fds_bits[i], 0, count); + } + + /* Don't copy/clear the array if we are creating a new + fd array for fork() */ + if (fdt->max_fds) { + memcpy(nfdt->fd, fdt->fd, + fdt->max_fds * sizeof(struct file *)); + /* clear the remainder of the array */ + memset(&nfdt->fd[fdt->max_fds], 0, + (nfdt->max_fds - fdt->max_fds) * + sizeof(struct file *)); + } + nfdt->next_fd = fdt->next_fd; } /* @@ -157,28 +234,21 @@ void free_fdset(fd_set *array, int num) vfree(array); } -/* - * Expand the fdset in the files_struct. Called with the files spinlock - * held for write. - */ -static int expand_fdset(struct files_struct *files, int nr) - __releases(file->file_lock) - __acquires(file->file_lock) +static struct fdtable *alloc_fdtable(int nr) { - fd_set *new_openset = NULL, *new_execset = NULL; - int error, nfds = 0; - struct fdtable *fdt; - - error = -EMFILE; - fdt = files_fdtable(files); - if (fdt->max_fdset >= NR_OPEN || nr >= NR_OPEN) - goto out; + struct fdtable *fdt = NULL; + int nfds = 0; + fd_set *new_openset = NULL, *new_execset = NULL; + struct file **new_fds; - nfds = fdt->max_fdset; - spin_unlock(&files->file_lock); + fdt = kmalloc(sizeof(*fdt), GFP_KERNEL); + if (!fdt) + goto out; + memset(fdt, 0, sizeof(*fdt)); - /* Expand to the max in easy steps */ - do { + nfds = __FD_SETSIZE; + /* Expand to the max in easy steps */ + do { if (nfds < (PAGE_SIZE * 8)) nfds = PAGE_SIZE * 8; else { @@ -188,50 +258,88 @@ static int expand_fdset(struct files_struct *files, int nr) } } while (nfds <= nr); - error = -ENOMEM; - new_openset = alloc_fdset(nfds); - new_execset = alloc_fdset(nfds); - spin_lock(&files->file_lock); - if (!new_openset || !new_execset) + new_openset = alloc_fdset(nfds); + new_execset = alloc_fdset(nfds); + if (!new_openset || !new_execset) + goto out; + fdt->open_fds = new_openset; + fdt->close_on_exec = new_execset; + fdt->max_fdset = nfds; + + nfds = NR_OPEN_DEFAULT; + /* + * Expand to the max in easy steps, and keep expanding it until + * we have enough for the requested fd array size. + */ + do { +#if NR_OPEN_DEFAULT < 256 + if (nfds < 256) + nfds = 256; + else +#endif + if (nfds < (PAGE_SIZE / sizeof(struct file *))) + nfds = PAGE_SIZE / sizeof(struct file *); + else { + nfds = nfds * 2; + if (nfds > NR_OPEN) + nfds = NR_OPEN; + } + } while (nfds <= nr); + new_fds = alloc_fd_array(nfds); + if (!new_fds) goto out; + fdt->fd = new_fds; + fdt->max_fds = nfds; + fdt->free_files = NULL; + return fdt; +out: + if (new_openset) + free_fdset(new_openset, nfds); + if (new_execset) + free_fdset(new_execset, nfds); + kfree(fdt); + return NULL; +} - error = 0; - - /* Copy the existing tables and install the new pointers */ +/* + * Expands the file descriptor table - it will allocate a new fdtable and + * both fd array and fdset. It is expected to be called with the + * files_lock held. + */ +static int expand_fdtable(struct files_struct *files, int nr) + __releases(files->file_lock) + __acquires(files->file_lock) +{ + int error = 0; + struct fdtable *fdt; + struct fdtable *nfdt = NULL; + + spin_unlock(&files->file_lock); + nfdt = alloc_fdtable(nr); + if (!nfdt) { + error = -ENOMEM; + spin_lock(&files->file_lock); + goto out; + } + + spin_lock(&files->file_lock); fdt = files_fdtable(files); - if (nfds > fdt->max_fdset) { - int i = fdt->max_fdset / (sizeof(unsigned long) * 8); - int count = (nfds - fdt->max_fdset) / 8; - - /* - * Don't copy the entire array if the current fdset is - * not yet initialised. - */ - if (i) { - memcpy (new_openset, fdt->open_fds, fdt->max_fdset/8); - memcpy (new_execset, fdt->close_on_exec, fdt->max_fdset/8); - memset (&new_openset->fds_bits[i], 0, count); - memset (&new_execset->fds_bits[i], 0, count); - } - - nfds = xchg(&fdt->max_fdset, nfds); - new_openset = xchg(&fdt->open_fds, new_openset); - new_execset = xchg(&fdt->close_on_exec, new_execset); + /* + * Check again since another task may have expanded the + * fd table while we dropped the lock + */ + if (nr >= fdt->max_fds || nr >= fdt->max_fdset) { + copy_fdtable(nfdt, fdt); + } else { + /* Somebody expanded while we dropped file_lock */ spin_unlock(&files->file_lock); - free_fdset (new_openset, nfds); - free_fdset (new_execset, nfds); + __free_fdtable(nfdt); spin_lock(&files->file_lock); - return 0; - } - /* Somebody expanded the array while we slept ... */ - + goto out; + } + rcu_assign_pointer(files->fdt, nfdt); + free_fdtable(fdt); out: - spin_unlock(&files->file_lock); - if (new_openset) - free_fdset(new_openset, nfds); - if (new_execset) - free_fdset(new_execset, nfds); - spin_lock(&files->file_lock); return error; } @@ -246,17 +354,36 @@ int expand_files(struct files_struct *files, int nr) struct fdtable *fdt; fdt = files_fdtable(files); - if (nr >= fdt->max_fdset) { - expand = 1; - if ((err = expand_fdset(files, nr))) + if (nr >= fdt->max_fdset || nr >= fdt->max_fds) { + if (fdt->max_fdset >= NR_OPEN || + fdt->max_fds >= NR_OPEN || nr >= NR_OPEN) { + err = -EMFILE; goto out; - } - if (nr >= fdt->max_fds) { + } expand = 1; - if ((err = expand_fd_array(files, nr))) + if ((err = expand_fdtable(files, nr))) goto out; } err = expand; out: return err; } + +static void __devinit fdtable_defer_list_init(int cpu) +{ + struct fdtable_defer *fddef = &per_cpu(fdtable_defer_list, cpu); + spin_lock_init(&fddef->lock); + INIT_WORK(&fddef->wq, (void (*)(void *))free_fdtable_work, fddef); + init_timer(&fddef->timer); + fddef->timer.data = (unsigned long)fddef; + fddef->timer.function = fdtable_timer; + fddef->next = NULL; +} + +void __init files_defer_init(void) +{ + int i; + /* Really early - can't use for_each_cpu */ + for (i = 0; i < NR_CPUS; i++) + fdtable_defer_list_init(i); +} diff --git a/fs/file_table.c b/fs/file_table.c index 43e9e1737de2..86ec8ae985b4 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -53,11 +54,17 @@ void filp_dtor(void * objp, struct kmem_cache_s *cachep, unsigned long dflags) spin_unlock_irqrestore(&filp_count_lock, flags); } -static inline void file_free(struct file *f) +static inline void file_free_rcu(struct rcu_head *head) { + struct file *f = container_of(head, struct file, f_rcuhead); kmem_cache_free(filp_cachep, f); } +static inline void file_free(struct file *f) +{ + call_rcu(&f->f_rcuhead, file_free_rcu); +} + /* Find an unused file structure and return a pointer to it. * Returns NULL, if there are no more free file structures or * we run out of memory. @@ -110,7 +117,7 @@ EXPORT_SYMBOL(get_empty_filp); void fastcall fput(struct file *file) { - if (atomic_dec_and_test(&file->f_count)) + if (rcuref_dec_and_test(&file->f_count)) __fput(file); } @@ -156,11 +163,17 @@ struct file fastcall *fget(unsigned int fd) struct file *file; struct files_struct *files = current->files; - spin_lock(&files->file_lock); + rcu_read_lock(); file = fcheck_files(files, fd); - if (file) - get_file(file); - spin_unlock(&files->file_lock); + if (file) { + if (!rcuref_inc_lf(&file->f_count)) { + /* File object ref couldn't be taken */ + rcu_read_unlock(); + return NULL; + } + } + rcu_read_unlock(); + return file; } @@ -182,21 +195,25 @@ struct file fastcall *fget_light(unsigned int fd, int *fput_needed) if (likely((atomic_read(&files->count) == 1))) { file = fcheck_files(files, fd); } else { - spin_lock(&files->file_lock); + rcu_read_lock(); file = fcheck_files(files, fd); if (file) { - get_file(file); - *fput_needed = 1; + if (rcuref_inc_lf(&file->f_count)) + *fput_needed = 1; + else + /* Didn't get the reference, someone's freed */ + file = NULL; } - spin_unlock(&files->file_lock); + rcu_read_unlock(); } + return file; } void put_filp(struct file *file) { - if (atomic_dec_and_test(&file->f_count)) { + if (rcuref_dec_and_test(&file->f_count)) { security_file_free(file); file_kill(file); file_free(file); @@ -257,4 +274,5 @@ void __init files_init(unsigned long mempages) files_stat.max_files = n; if (files_stat.max_files < NR_FILE) files_stat.max_files = NR_FILE; + files_defer_init(); } diff --git a/fs/open.c b/fs/open.c index b6542516a0ca..2fac58c51910 100644 --- a/fs/open.c +++ b/fs/open.c @@ -24,6 +24,7 @@ #include #include #include +#include #include @@ -930,9 +931,8 @@ void fastcall fd_install(unsigned int fd, struct file * file) struct fdtable *fdt; spin_lock(&files->file_lock); fdt = files_fdtable(files); - if (unlikely(fdt->fd[fd] != NULL)) - BUG(); - fdt->fd[fd] = file; + BUG_ON(fdt->fd[fd] != NULL); + rcu_assign_pointer(fdt->fd[fd], file); spin_unlock(&files->file_lock); } @@ -1024,7 +1024,7 @@ asmlinkage long sys_close(unsigned int fd) filp = fdt->fd[fd]; if (!filp) goto out_unlock; - fdt->fd[fd] = NULL; + rcu_assign_pointer(fdt->fd[fd], NULL); FD_CLR(fd, fdt->close_on_exec); __put_unused_fd(files, fd); spin_unlock(&files->file_lock); diff --git a/include/linux/file.h b/include/linux/file.h index db372230848e..f5bbd4c508b3 100644 --- a/include/linux/file.h +++ b/include/linux/file.h @@ -9,6 +9,7 @@ #include #include #include +#include /* * The default fd array needs to be at least BITS_PER_LONG, @@ -23,6 +24,9 @@ struct fdtable { struct file ** fd; /* current fd array */ fd_set *close_on_exec; fd_set *open_fds; + struct rcu_head rcu; + struct files_struct *free_files; + struct fdtable *next; }; /* @@ -31,13 +35,14 @@ struct fdtable { struct files_struct { atomic_t count; spinlock_t file_lock; /* Protects all the below members. Nests inside tsk->alloc_lock */ + struct fdtable *fdt; struct fdtable fdtab; fd_set close_on_exec_init; fd_set open_fds_init; struct file * fd_array[NR_OPEN_DEFAULT]; }; -#define files_fdtable(files) (&(files)->fdtab) +#define files_fdtable(files) (rcu_dereference((files)->fdt)) extern void FASTCALL(__fput(struct file *)); extern void FASTCALL(fput(struct file *)); @@ -65,6 +70,8 @@ extern fd_set *alloc_fdset(int); extern void free_fdset(fd_set *, int); extern int expand_files(struct files_struct *, int nr); +extern void free_fdtable(struct fdtable *fdt); +extern void __init files_defer_init(void); static inline struct file * fcheck_files(struct files_struct *files, unsigned int fd) { @@ -72,7 +79,7 @@ static inline struct file * fcheck_files(struct files_struct *files, unsigned in struct fdtable *fdt = files_fdtable(files); if (fd < fdt->max_fds) - file = fdt->fd[fd]; + file = rcu_dereference(fdt->fd[fd]); return file; } diff --git a/include/linux/fs.h b/include/linux/fs.h index fd93ab7da905..7f61227827d7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -9,6 +9,7 @@ #include #include #include +#include /* * It's silly to have NR_OPEN bigger than NR_FILE, but you can change @@ -597,12 +598,13 @@ struct file { spinlock_t f_ep_lock; #endif /* #ifdef CONFIG_EPOLL */ struct address_space *f_mapping; + struct rcu_head f_rcuhead; }; extern spinlock_t files_lock; #define file_list_lock() spin_lock(&files_lock); #define file_list_unlock() spin_unlock(&files_lock); -#define get_file(x) atomic_inc(&(x)->f_count) +#define get_file(x) rcuref_inc(&(x)->f_count) #define file_count(x) atomic_read(&(x)->f_count) #define MAX_NON_LFS ((1UL<<31) - 1) diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 94aefa54a1b5..68ab5f2ab9cd 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -2,6 +2,7 @@ #define _LINUX__INIT_TASK_H #include +#include #define INIT_FDTABLE \ { \ @@ -11,12 +12,16 @@ .fd = &init_files.fd_array[0], \ .close_on_exec = &init_files.close_on_exec_init, \ .open_fds = &init_files.open_fds_init, \ + .rcu = RCU_HEAD_INIT, \ + .free_files = NULL, \ + .next = NULL, \ } #define INIT_FILES \ { \ .count = ATOMIC_INIT(1), \ .file_lock = SPIN_LOCK_UNLOCKED, \ + .fdt = &init_files.fdtab, \ .fdtab = INIT_FDTABLE, \ .close_on_exec_init = { { 0, } }, \ .open_fds_init = { { 0, } }, \ diff --git a/kernel/exit.c b/kernel/exit.c index 83beb1e93b18..6d2089a1bce7 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -411,15 +411,16 @@ void fastcall put_files_struct(struct files_struct *files) close_files(files); /* * Free the fd and fdset arrays if we expanded them. + * If the fdtable was embedded, pass files for freeing + * at the end of the RCU grace period. Otherwise, + * you can free files immediately. */ fdt = files_fdtable(files); - if (fdt->fd != &files->fd_array[0]) - free_fd_array(fdt->fd, fdt->max_fds); - if (fdt->max_fdset > __FD_SETSIZE) { - free_fdset(fdt->open_fds, fdt->max_fdset); - free_fdset(fdt->close_on_exec, fdt->max_fdset); - } - kmem_cache_free(files_cachep, files); + if (fdt == &files->fdtab) + fdt->free_files = files; + else + kmem_cache_free(files_cachep, files); + free_fdtable(fdt); } } diff --git a/kernel/fork.c b/kernel/fork.c index ecc694debb50..8149f3602881 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -35,6 +35,7 @@ #include #include #include +#include #include #include #include @@ -565,13 +566,12 @@ static inline int copy_fs(unsigned long clone_flags, struct task_struct * tsk) return 0; } -static int count_open_files(struct files_struct *files, int size) +static int count_open_files(struct fdtable *fdt) { + int size = fdt->max_fdset; int i; - struct fdtable *fdt; /* Find the last open fd */ - fdt = files_fdtable(files); for (i = size/(8*sizeof(long)); i > 0; ) { if (fdt->open_fds->fds_bits[--i]) break; @@ -592,13 +592,17 @@ static struct files_struct *alloc_files(void) atomic_set(&newf->count, 1); spin_lock_init(&newf->file_lock); - fdt = files_fdtable(newf); + fdt = &newf->fdtab; fdt->next_fd = 0; fdt->max_fds = NR_OPEN_DEFAULT; fdt->max_fdset = __FD_SETSIZE; fdt->close_on_exec = &newf->close_on_exec_init; fdt->open_fds = &newf->open_fds_init; fdt->fd = &newf->fd_array[0]; + INIT_RCU_HEAD(&fdt->rcu); + fdt->free_files = NULL; + fdt->next = NULL; + rcu_assign_pointer(newf->fdt, fdt); out: return newf; } @@ -637,7 +641,7 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) old_fdt = files_fdtable(oldf); new_fdt = files_fdtable(newf); size = old_fdt->max_fdset; - open_files = count_open_files(oldf, old_fdt->max_fdset); + open_files = count_open_files(old_fdt); expand = 0; /* @@ -661,7 +665,14 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) spin_unlock(&newf->file_lock); if (error < 0) goto out_release; + new_fdt = files_fdtable(newf); + /* + * Reacquire the oldf lock and a pointer to its fd table + * who knows it may have a new bigger fd table. We need + * the latest pointer. + */ spin_lock(&oldf->file_lock); + old_fdt = files_fdtable(oldf); } old_fds = old_fdt->fd; @@ -683,7 +694,7 @@ static int copy_files(unsigned long clone_flags, struct task_struct * tsk) */ FD_CLR(open_files - i, new_fdt->open_fds); } - *new_fds++ = f; + rcu_assign_pointer(*new_fds++, f); } spin_unlock(&oldf->file_lock); -- cgit v1.2.3 From 7726e9e10fc6e026ed2dc00e48f4a3ffc1254ad2 Mon Sep 17 00:00:00 2001 From: "Antonino A. Daplas" Date: Fri, 9 Sep 2005 13:04:29 -0700 Subject: [PATCH] fbdev: Add fbset -a support Add capability to fbdev to listen to the FB_ACTIVATE_ALL flag. If set, it notifies fbcon that all consoles must be set to the current var. Signed-off-by: Antonino Daplas Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/video/console/fbcon.c | 48 +++++++++++++++++++++++++++++++++++++++++++ drivers/video/fbmem.c | 6 ++++-- include/linux/fb.h | 3 +++ 3 files changed, 55 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/drivers/video/console/fbcon.c b/drivers/video/console/fbcon.c index 35c88bd7ba5e..751890a5f5f3 100644 --- a/drivers/video/console/fbcon.c +++ b/drivers/video/console/fbcon.c @@ -2593,6 +2593,51 @@ static void fbcon_modechanged(struct fb_info *info) } } +static void fbcon_set_all_vcs(struct fb_info *info) +{ + struct fbcon_ops *ops = info->fbcon_par; + struct vc_data *vc; + struct display *p; + int i, rows, cols; + + if (!ops || ops->currcon < 0) + return; + + for (i = 0; i < MAX_NR_CONSOLES; i++) { + vc = vc_cons[i].d; + if (!vc || vc->vc_mode != KD_TEXT || + registered_fb[con2fb_map[i]] != info) + continue; + + p = &fb_display[vc->vc_num]; + + info->var.xoffset = info->var.yoffset = p->yscroll = 0; + var_to_display(p, &info->var, info); + cols = info->var.xres / vc->vc_font.width; + rows = info->var.yres / vc->vc_font.height; + vc_resize(vc, cols, rows); + + if (CON_IS_VISIBLE(vc)) { + updatescrollmode(p, info, vc); + scrollback_max = 0; + scrollback_current = 0; + update_var(vc->vc_num, info); + fbcon_set_palette(vc, color_table); + update_screen(vc); + if (softback_buf) { + int l = fbcon_softback_size / vc->vc_size_row; + if (l > 5) + softback_end = softback_buf + l * vc->vc_size_row; + else { + /* Smaller scrollback makes no sense, and 0 + would screw the operation totally */ + softback_top = 0; + } + } + } + } +} + static int fbcon_mode_deleted(struct fb_info *info, struct fb_videomode *mode) { @@ -2708,6 +2753,9 @@ static int fbcon_event_notify(struct notifier_block *self, case FB_EVENT_MODE_CHANGE: fbcon_modechanged(info); break; + case FB_EVENT_MODE_CHANGE_ALL: + fbcon_set_all_vcs(info); + break; case FB_EVENT_MODE_DELETE: mode = event->data; ret = fbcon_mode_deleted(info, mode); diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c index 4ff853fbe0be..a8eee79e117d 100644 --- a/drivers/video/fbmem.c +++ b/drivers/video/fbmem.c @@ -684,11 +684,13 @@ fb_set_var(struct fb_info *info, struct fb_var_screeninfo *var) if (!err && (flags & FBINFO_MISC_USEREVENT)) { struct fb_event event; + int evnt = (var->activate & FB_ACTIVATE_ALL) ? + FB_EVENT_MODE_CHANGE_ALL : + FB_EVENT_MODE_CHANGE; info->flags &= ~FBINFO_MISC_USEREVENT; event.info = info; - notifier_call_chain(&fb_notifier_list, - FB_EVENT_MODE_CHANGE, + notifier_call_chain(&fb_notifier_list, evnt, &event); } } diff --git a/include/linux/fb.h b/include/linux/fb.h index bc24beeed971..70da819df0f5 100644 --- a/include/linux/fb.h +++ b/include/linux/fb.h @@ -495,6 +495,9 @@ struct fb_cursor_user { #define FB_EVENT_BLANK 0x08 /* Private modelist is to be replaced */ #define FB_EVENT_NEW_MODELIST 0x09 +/* The resolution of the passed in fb_info about to change and + all vc's should be changed */ +#define FB_EVENT_MODE_CHANGE_ALL 0x0A struct fb_event { struct fb_info *info; -- cgit v1.2.3 From d2d58384fc5d4c0fe2d8e34bc2d15a90a9bb372a Mon Sep 17 00:00:00 2001 From: "Antonino A. Daplas" Date: Fri, 9 Sep 2005 13:04:31 -0700 Subject: [PATCH] vesafb: Add blanking support Add rudimentary support by manipulating the VGA registers. However, not all vesa modes are VGA compatible, so VGA compatiblity is checked first. Only 2 levels are supported, powerup and powerdown. Signed-off-by: Antonino Daplas Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/boot/video.S | 5 +++++ drivers/video/vesafb.c | 38 ++++++++++++++++++++++++++++++++++++++ include/linux/tty.h | 3 ++- 3 files changed, 45 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/arch/i386/boot/video.S b/arch/i386/boot/video.S index 0587477c99f2..02bf625a681b 100644 --- a/arch/i386/boot/video.S +++ b/arch/i386/boot/video.S @@ -97,6 +97,7 @@ #define PARAM_VESAPM_OFF 0x30 #define PARAM_LFB_PAGES 0x32 #define PARAM_VESA_ATTRIB 0x34 +#define PARAM_CAPABILITIES 0x36 /* Define DO_STORE according to CONFIG_VIDEO_RETAIN */ #ifdef CONFIG_VIDEO_RETAIN @@ -233,6 +234,10 @@ mopar_gr: movw 18(%di), %ax movl %eax, %fs:(PARAM_LFB_SIZE) +# store mode capabilities + movl 10(%di), %eax + movl %eax, %fs:(PARAM_CAPABILITIES) + # switching the DAC to 8-bit is for <= 8 bpp only movw %fs:(PARAM_LFB_DEPTH), %ax cmpw $8, %ax diff --git a/drivers/video/vesafb.c b/drivers/video/vesafb.c index a272592b0373..1ca80264c7b0 100644 --- a/drivers/video/vesafb.c +++ b/drivers/video/vesafb.c @@ -19,6 +19,7 @@ #include #include #include +#include