From 9512a16b0e1217bbef73d276a67c28b5fbb46512 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 16 May 2017 15:57:42 -0400 Subject: nfsd: Revert "nfsd: check for oversized NFSv2/v3 arguments" This reverts commit 51f567777799 "nfsd: check for oversized NFSv2/v3 arguments", which breaks support for NFSv3 ACLs. That patch was actually an earlier draft of a fix for the problem that was eventually fixed by e6838a29ecb "nfsd: check for oversized NFSv2/v3 arguments". But somehow I accidentally left this earlier draft in the branch that was part of my 2.12 pull request. Reported-by: Eryu Guan Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/svc.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h index 94631026f79c..11cef5a7bc87 100644 --- a/include/linux/sunrpc/svc.h +++ b/include/linux/sunrpc/svc.h @@ -336,7 +336,8 @@ xdr_argsize_check(struct svc_rqst *rqstp, __be32 *p) { char *cp = (char *)p; struct kvec *vec = &rqstp->rq_arg.head[0]; - return cp == (char *)vec->iov_base + vec->iov_len; + return cp >= (char*)vec->iov_base + && cp <= (char*)vec->iov_base + vec->iov_len; } static inline int -- cgit v1.2.3 From b4d2ea2af95cb77e2f320e24da526280d4aa2f6b Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Mon, 8 May 2017 10:48:21 +0200 Subject: Revert "pinctrl: generic: Add bi-directional and output-enable" This reverts commit 8c58f1a7a4b6d1d723bf25fef9d842d5a11200d0. It turns out that applying these generic properties was premature: the properties used in the driver using this are of unclear electrical nature and the subject need to be discussed. Signed-off-by: Linus Walleij --- Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt | 2 -- drivers/pinctrl/pinconf-generic.c | 3 --- include/linux/pinctrl/pinconf-generic.h | 3 --- 3 files changed, 8 deletions(-) (limited to 'include') diff --git a/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt b/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt index 71a3c134af1b..f01d154090da 100644 --- a/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt +++ b/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt @@ -247,7 +247,6 @@ bias-bus-hold - latch weakly bias-pull-up - pull up the pin bias-pull-down - pull down the pin bias-pull-pin-default - use pin-default pull state -bi-directional - pin supports simultaneous input/output operations drive-push-pull - drive actively high and low drive-open-drain - drive with open drain drive-open-source - drive with open source @@ -260,7 +259,6 @@ input-debounce - debounce mode with debound time X power-source - select between different power supplies low-power-enable - enable low power mode low-power-disable - disable low power mode -output-enable - enable output on pin regardless of output value output-low - set the pin to output mode with low level output-high - set the pin to output mode with high level slew-rate - set the slew rate diff --git a/drivers/pinctrl/pinconf-generic.c b/drivers/pinctrl/pinconf-generic.c index 0d6b7f4b82af..720a19fd38d2 100644 --- a/drivers/pinctrl/pinconf-generic.c +++ b/drivers/pinctrl/pinconf-generic.c @@ -35,7 +35,6 @@ static const struct pin_config_item conf_items[] = { PCONFDUMP(PIN_CONFIG_BIAS_PULL_PIN_DEFAULT, "input bias pull to pin specific state", NULL, false), PCONFDUMP(PIN_CONFIG_BIAS_PULL_UP, "input bias pull up", NULL, false), - PCONFDUMP(PIN_CONFIG_BIDIRECTIONAL, "bi-directional pin operations", NULL, false), PCONFDUMP(PIN_CONFIG_DRIVE_OPEN_DRAIN, "output drive open drain", NULL, false), PCONFDUMP(PIN_CONFIG_DRIVE_OPEN_SOURCE, "output drive open source", NULL, false), PCONFDUMP(PIN_CONFIG_DRIVE_PUSH_PULL, "output drive push pull", NULL, false), @@ -161,7 +160,6 @@ static const struct pinconf_generic_params dt_params[] = { { "bias-pull-up", PIN_CONFIG_BIAS_PULL_UP, 1 }, { "bias-pull-pin-default", PIN_CONFIG_BIAS_PULL_PIN_DEFAULT, 1 }, { "bias-pull-down", PIN_CONFIG_BIAS_PULL_DOWN, 1 }, - { "bi-directional", PIN_CONFIG_BIDIRECTIONAL, 1 }, { "drive-open-drain", PIN_CONFIG_DRIVE_OPEN_DRAIN, 0 }, { "drive-open-source", PIN_CONFIG_DRIVE_OPEN_SOURCE, 0 }, { "drive-push-pull", PIN_CONFIG_DRIVE_PUSH_PULL, 0 }, @@ -174,7 +172,6 @@ static const struct pinconf_generic_params dt_params[] = { { "input-schmitt-enable", PIN_CONFIG_INPUT_SCHMITT_ENABLE, 1 }, { "low-power-disable", PIN_CONFIG_LOW_POWER_MODE, 0 }, { "low-power-enable", PIN_CONFIG_LOW_POWER_MODE, 1 }, - { "output-enable", PIN_CONFIG_OUTPUT, 1, }, { "output-high", PIN_CONFIG_OUTPUT, 1, }, { "output-low", PIN_CONFIG_OUTPUT, 0, }, { "power-source", PIN_CONFIG_POWER_SOURCE, 0 }, diff --git a/include/linux/pinctrl/pinconf-generic.h b/include/linux/pinctrl/pinconf-generic.h index 279e3c5326e3..7620eb127cff 100644 --- a/include/linux/pinctrl/pinconf-generic.h +++ b/include/linux/pinctrl/pinconf-generic.h @@ -42,8 +42,6 @@ * @PIN_CONFIG_BIAS_PULL_UP: the pin will be pulled up (usually with high * impedance to VDD). If the argument is != 0 pull-up is enabled, * if it is 0, pull-up is total, i.e. the pin is connected to VDD. - * @PIN_CONFIG_BIDIRECTIONAL: the pin will be configured to allow simultaneous - * input and output operations. * @PIN_CONFIG_DRIVE_OPEN_DRAIN: the pin will be driven with open drain (open * collector) which means it is usually wired with other output ports * which are then pulled up with an external resistor. Setting this @@ -98,7 +96,6 @@ enum pin_config_param { PIN_CONFIG_BIAS_PULL_DOWN, PIN_CONFIG_BIAS_PULL_PIN_DEFAULT, PIN_CONFIG_BIAS_PULL_UP, - PIN_CONFIG_BIDIRECTIONAL, PIN_CONFIG_DRIVE_OPEN_DRAIN, PIN_CONFIG_DRIVE_OPEN_SOURCE, PIN_CONFIG_DRIVE_PUSH_PULL, -- cgit v1.2.3 From 020e0b1c8f19f1fc3bce23beeccd80c574ca0e49 Mon Sep 17 00:00:00 2001 From: Anatolij Gustschin Date: Thu, 11 May 2017 20:24:31 +0200 Subject: gpiolib: Add stubs for gpiod lookup table interface Add stubs for gpiod_add_lookup_table() and gpiod_remove_lookup_table() for the !GPIOLIB case to prevent build errors. Signed-off-by: Anatolij Gustschin Reviewed-by: Andy Shevchenko Signed-off-by: Linus Walleij --- include/linux/gpio/machine.h | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'include') diff --git a/include/linux/gpio/machine.h b/include/linux/gpio/machine.h index c0d712d22b07..f738d50cc17d 100644 --- a/include/linux/gpio/machine.h +++ b/include/linux/gpio/machine.h @@ -56,7 +56,14 @@ struct gpiod_lookup_table { .flags = _flags, \ } +#ifdef CONFIG_GPIOLIB void gpiod_add_lookup_table(struct gpiod_lookup_table *table); void gpiod_remove_lookup_table(struct gpiod_lookup_table *table); +#else +static inline +void gpiod_add_lookup_table(struct gpiod_lookup_table *table) {} +static inline +void gpiod_remove_lookup_table(struct gpiod_lookup_table *table) {} +#endif #endif /* __LINUX_GPIO_MACHINE_H */ -- cgit v1.2.3 From c61872c9833d17d3807fb999096917c1f9eaada0 Mon Sep 17 00:00:00 2001 From: Mika Westerberg Date: Wed, 17 May 2017 13:25:12 +0300 Subject: firmware: dmi: Add DMI_PRODUCT_FAMILY identification string Sometimes it is more convenient to be able to match a whole family of products, like in case of bunch of Chromebooks based on Intel_Strago to apply a driver quirk instead of quirking each machine one-by-one. This adds support for DMI_PRODUCT_FAMILY identification string and also exports it to the userspace through sysfs attribute just like the existing ones. Suggested-by: Dmitry Torokhov Signed-off-by: Mika Westerberg Reviewed-by: Andy Shevchenko Signed-off-by: Linus Walleij --- drivers/firmware/dmi-id.c | 2 ++ drivers/firmware/dmi_scan.c | 1 + include/linux/mod_devicetable.h | 1 + 3 files changed, 4 insertions(+) (limited to 'include') diff --git a/drivers/firmware/dmi-id.c b/drivers/firmware/dmi-id.c index 44c01390d035..dc269cb288c2 100644 --- a/drivers/firmware/dmi-id.c +++ b/drivers/firmware/dmi-id.c @@ -47,6 +47,7 @@ DEFINE_DMI_ATTR_WITH_SHOW(product_name, 0444, DMI_PRODUCT_NAME); DEFINE_DMI_ATTR_WITH_SHOW(product_version, 0444, DMI_PRODUCT_VERSION); DEFINE_DMI_ATTR_WITH_SHOW(product_serial, 0400, DMI_PRODUCT_SERIAL); DEFINE_DMI_ATTR_WITH_SHOW(product_uuid, 0400, DMI_PRODUCT_UUID); +DEFINE_DMI_ATTR_WITH_SHOW(product_family, 0400, DMI_PRODUCT_FAMILY); DEFINE_DMI_ATTR_WITH_SHOW(board_vendor, 0444, DMI_BOARD_VENDOR); DEFINE_DMI_ATTR_WITH_SHOW(board_name, 0444, DMI_BOARD_NAME); DEFINE_DMI_ATTR_WITH_SHOW(board_version, 0444, DMI_BOARD_VERSION); @@ -191,6 +192,7 @@ static void __init dmi_id_init_attr_table(void) ADD_DMI_ATTR(product_version, DMI_PRODUCT_VERSION); ADD_DMI_ATTR(product_serial, DMI_PRODUCT_SERIAL); ADD_DMI_ATTR(product_uuid, DMI_PRODUCT_UUID); + ADD_DMI_ATTR(product_family, DMI_PRODUCT_FAMILY); ADD_DMI_ATTR(board_vendor, DMI_BOARD_VENDOR); ADD_DMI_ATTR(board_name, DMI_BOARD_NAME); ADD_DMI_ATTR(board_version, DMI_BOARD_VERSION); diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c index 54be60ead08f..93f7acdaac7a 100644 --- a/drivers/firmware/dmi_scan.c +++ b/drivers/firmware/dmi_scan.c @@ -430,6 +430,7 @@ static void __init dmi_decode(const struct dmi_header *dm, void *dummy) dmi_save_ident(dm, DMI_PRODUCT_VERSION, 6); dmi_save_ident(dm, DMI_PRODUCT_SERIAL, 7); dmi_save_uuid(dm, DMI_PRODUCT_UUID, 8); + dmi_save_ident(dm, DMI_PRODUCT_FAMILY, 26); break; case 2: /* Base Board Information */ dmi_save_ident(dm, DMI_BOARD_VENDOR, 4); diff --git a/include/linux/mod_devicetable.h b/include/linux/mod_devicetable.h index 566fda587fcf..3f74ef2281e8 100644 --- a/include/linux/mod_devicetable.h +++ b/include/linux/mod_devicetable.h @@ -467,6 +467,7 @@ enum dmi_field { DMI_PRODUCT_VERSION, DMI_PRODUCT_SERIAL, DMI_PRODUCT_UUID, + DMI_PRODUCT_FAMILY, DMI_BOARD_VENDOR, DMI_BOARD_NAME, DMI_BOARD_VERSION, -- cgit v1.2.3 From 118b90f3f18e733c99f0e8b98ea31a815ffc4d14 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Thu, 18 May 2017 14:10:22 +0300 Subject: drm/dp: add helper for reading DP sink/branch device desc from DPCD Reviewed-by: Daniel Vetter Signed-off-by: Jani Nikula Link: http://patchwork.freedesktop.org/patch/msgid/acba54da7d80eafea9e59a893e27e3c31028c0ba.1495105635.git.jani.nikula@intel.com --- drivers/gpu/drm/drm_dp_helper.c | 35 +++++++++++++++++++++++++++++++++++ include/drm/drm_dp_helper.h | 19 +++++++++++++++++++ 2 files changed, 54 insertions(+) (limited to 'include') diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 3e5f52110ea1..52e0ca9a5bb1 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -1208,3 +1208,38 @@ int drm_dp_stop_crc(struct drm_dp_aux *aux) return 0; } EXPORT_SYMBOL(drm_dp_stop_crc); + +/** + * drm_dp_read_desc - read sink/branch descriptor from DPCD + * @aux: DisplayPort AUX channel + * @desc: Device decriptor to fill from DPCD + * @is_branch: true for branch devices, false for sink devices + * + * Read DPCD 0x400 (sink) or 0x500 (branch) into @desc. Also debug log the + * identification. + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_read_desc(struct drm_dp_aux *aux, struct drm_dp_desc *desc, + bool is_branch) +{ + struct drm_dp_dpcd_ident *ident = &desc->ident; + unsigned int offset = is_branch ? DP_BRANCH_OUI : DP_SINK_OUI; + int ret, dev_id_len; + + ret = drm_dp_dpcd_read(aux, offset, ident, sizeof(*ident)); + if (ret < 0) + return ret; + + dev_id_len = strnlen(ident->device_id, sizeof(ident->device_id)); + + DRM_DEBUG_KMS("DP %s: OUI %*phD dev-ID %*pE HW-rev %d.%d SW-rev %d.%d\n", + is_branch ? "branch" : "sink", + (int)sizeof(ident->oui), ident->oui, + dev_id_len, ident->device_id, + ident->hw_rev >> 4, ident->hw_rev & 0xf, + ident->sw_major_rev, ident->sw_minor_rev); + + return 0; +} +EXPORT_SYMBOL(drm_dp_read_desc); diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index c0bd0d7651a9..84502da177a1 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -913,4 +913,23 @@ void drm_dp_aux_unregister(struct drm_dp_aux *aux); int drm_dp_start_crc(struct drm_dp_aux *aux, struct drm_crtc *crtc); int drm_dp_stop_crc(struct drm_dp_aux *aux); +struct drm_dp_dpcd_ident { + u8 oui[3]; + u8 device_id[6]; + u8 hw_rev; + u8 sw_major_rev; + u8 sw_minor_rev; +} __packed; + +/** + * struct drm_dp_desc - DP branch/sink device descriptor + * @ident: DP device identification from DPCD 0x400 (sink) or 0x500 (branch). + */ +struct drm_dp_desc { + struct drm_dp_dpcd_ident ident; +}; + +int drm_dp_read_desc(struct drm_dp_aux *aux, struct drm_dp_desc *desc, + bool is_branch); + #endif /* _DRM_DP_HELPER_H_ */ -- cgit v1.2.3 From 76fa998acd86b6b40d0217e12af39c2406bdcd2b Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Thu, 18 May 2017 14:10:24 +0300 Subject: drm/dp: start a DPCD based DP sink/branch device quirk database MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Face the fact, there are Display Port sink and branch devices out there in the wild that don't follow the Display Port specifications, or they have bugs, or just otherwise require special treatment. Start a common quirk database the drivers can query based on the DP device identification. At least for now, we leave the workarounds for the drivers to implement as they see fit. For starters, add a branch device that can't handle full 24-bit main link Mdiv and Ndiv main link attributes properly. Naturally, the workaround of reducing main link attributes for all devices ended up in regressions for other devices. So here we are. v2: Rebase on DRM DP desc read helpers v3: Fix the OUI memcmp blunder (Clint) Cc: Ville Syrjälä Cc: Dhinakaran Pandiyan Cc: Clint Taylor Cc: Adam Jackson Cc: Harry Wentland Tested-by: Clinton Taylor Reviewed-by: Clinton Taylor Reviewed-by: Daniel Vetter # v2 Signed-off-by: Jani Nikula Link: http://patchwork.freedesktop.org/patch/msgid/91ec198dd95258dbf3bee2f6be739e0da73b4fdd.1495105635.git.jani.nikula@intel.com --- drivers/gpu/drm/drm_dp_helper.c | 52 +++++++++++++++++++++++++++++++++++++++-- include/drm/drm_dp_helper.h | 32 +++++++++++++++++++++++++ 2 files changed, 82 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 52e0ca9a5bb1..213fb837e1c4 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -1209,6 +1209,51 @@ int drm_dp_stop_crc(struct drm_dp_aux *aux) } EXPORT_SYMBOL(drm_dp_stop_crc); +struct dpcd_quirk { + u8 oui[3]; + bool is_branch; + u32 quirks; +}; + +#define OUI(first, second, third) { (first), (second), (third) } + +static const struct dpcd_quirk dpcd_quirk_list[] = { + /* Analogix 7737 needs reduced M and N at HBR2 link rates */ + { OUI(0x00, 0x22, 0xb9), true, BIT(DP_DPCD_QUIRK_LIMITED_M_N) }, +}; + +#undef OUI + +/* + * Get a bit mask of DPCD quirks for the sink/branch device identified by + * ident. The quirk data is shared but it's up to the drivers to act on the + * data. + * + * For now, only the OUI (first three bytes) is used, but this may be extended + * to device identification string and hardware/firmware revisions later. + */ +static u32 +drm_dp_get_quirks(const struct drm_dp_dpcd_ident *ident, bool is_branch) +{ + const struct dpcd_quirk *quirk; + u32 quirks = 0; + int i; + + for (i = 0; i < ARRAY_SIZE(dpcd_quirk_list); i++) { + quirk = &dpcd_quirk_list[i]; + + if (quirk->is_branch != is_branch) + continue; + + if (memcmp(quirk->oui, ident->oui, sizeof(ident->oui)) != 0) + continue; + + quirks |= quirk->quirks; + } + + return quirks; +} + /** * drm_dp_read_desc - read sink/branch descriptor from DPCD * @aux: DisplayPort AUX channel @@ -1231,14 +1276,17 @@ int drm_dp_read_desc(struct drm_dp_aux *aux, struct drm_dp_desc *desc, if (ret < 0) return ret; + desc->quirks = drm_dp_get_quirks(ident, is_branch); + dev_id_len = strnlen(ident->device_id, sizeof(ident->device_id)); - DRM_DEBUG_KMS("DP %s: OUI %*phD dev-ID %*pE HW-rev %d.%d SW-rev %d.%d\n", + DRM_DEBUG_KMS("DP %s: OUI %*phD dev-ID %*pE HW-rev %d.%d SW-rev %d.%d quirks 0x%04x\n", is_branch ? "branch" : "sink", (int)sizeof(ident->oui), ident->oui, dev_id_len, ident->device_id, ident->hw_rev >> 4, ident->hw_rev & 0xf, - ident->sw_major_rev, ident->sw_minor_rev); + ident->sw_major_rev, ident->sw_minor_rev, + desc->quirks); return 0; } diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 84502da177a1..bb837310c07e 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -924,12 +924,44 @@ struct drm_dp_dpcd_ident { /** * struct drm_dp_desc - DP branch/sink device descriptor * @ident: DP device identification from DPCD 0x400 (sink) or 0x500 (branch). + * @quirks: Quirks; use drm_dp_has_quirk() to query for the quirks. */ struct drm_dp_desc { struct drm_dp_dpcd_ident ident; + u32 quirks; }; int drm_dp_read_desc(struct drm_dp_aux *aux, struct drm_dp_desc *desc, bool is_branch); +/** + * enum drm_dp_quirk - Display Port sink/branch device specific quirks + * + * Display Port sink and branch devices in the wild have a variety of bugs, try + * to collect them here. The quirks are shared, but it's up to the drivers to + * implement workarounds for them. + */ +enum drm_dp_quirk { + /** + * @DP_DPCD_QUIRK_LIMITED_M_N: + * + * The device requires main link attributes Mvid and Nvid to be limited + * to 16 bits. + */ + DP_DPCD_QUIRK_LIMITED_M_N, +}; + +/** + * drm_dp_has_quirk() - does the DP device have a specific quirk + * @desc: Device decriptor filled by drm_dp_read_desc() + * @quirk: Quirk to query for + * + * Return true if DP device identified by @desc has @quirk. + */ +static inline bool +drm_dp_has_quirk(const struct drm_dp_desc *desc, enum drm_dp_quirk quirk) +{ + return desc->quirks & BIT(quirk); +} + #endif /* _DRM_DP_HELPER_H_ */ -- cgit v1.2.3 From 25cdda95fda78d22d44157da15aa7ea34be3c804 Mon Sep 17 00:00:00 2001 From: Nicholas Bellinger Date: Wed, 24 May 2017 21:47:09 -0700 Subject: iscsi-target: Fix initial login PDU asynchronous socket close OOPs This patch fixes a OOPs originally introduced by: commit bb048357dad6d604520c91586334c9c230366a14 Author: Nicholas Bellinger Date: Thu Sep 5 14:54:04 2013 -0700 iscsi-target: Add sk->sk_state_change to cleanup after TCP failure which would trigger a NULL pointer dereference when a TCP connection was closed asynchronously via iscsi_target_sk_state_change(), but only when the initial PDU processing in iscsi_target_do_login() from iscsi_np process context was blocked waiting for backend I/O to complete. To address this issue, this patch makes the following changes. First, it introduces some common helper functions used for checking socket closing state, checking login_flags, and atomically checking socket closing state + setting login_flags. Second, it introduces a LOGIN_FLAGS_INITIAL_PDU bit to know when a TCP connection has dropped via iscsi_target_sk_state_change(), but the initial PDU processing within iscsi_target_do_login() in iscsi_np context is still running. For this case, it sets LOGIN_FLAGS_CLOSED, but doesn't invoke schedule_delayed_work(). The original NULL pointer dereference case reported by MNC is now handled by iscsi_target_do_login() doing a iscsi_target_sk_check_close() before transitioning to FFP to determine when the socket has already closed, or iscsi_target_start_negotiation() if the login needs to exchange more PDUs (eg: iscsi_target_do_login returned 0) but the socket has closed. For both of these cases, the cleanup up of remaining connection resources will occur in iscsi_target_start_negotiation() from iscsi_np process context once the failure is detected. Finally, to handle to case where iscsi_target_sk_state_change() is called after the initial PDU procesing is complete, it now invokes conn->login_work -> iscsi_target_do_login_rx() to perform cleanup once existing iscsi_target_sk_check_close() checks detect connection failure. For this case, the cleanup of remaining connection resources will occur in iscsi_target_do_login_rx() from delayed workqueue process context once the failure is detected. Reported-by: Mike Christie Reviewed-by: Mike Christie Tested-by: Mike Christie Cc: Mike Christie Reported-by: Hannes Reinecke Cc: Hannes Reinecke Cc: Sagi Grimberg Cc: Varun Prakash Cc: # v3.12+ Signed-off-by: Nicholas Bellinger --- drivers/target/iscsi/iscsi_target_nego.c | 194 +++++++++++++++++++++---------- include/target/iscsi/iscsi_target_core.h | 1 + 2 files changed, 133 insertions(+), 62 deletions(-) (limited to 'include') diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c index 7ccc9c1cbfd1..6f88b31242b0 100644 --- a/drivers/target/iscsi/iscsi_target_nego.c +++ b/drivers/target/iscsi/iscsi_target_nego.c @@ -493,14 +493,60 @@ static void iscsi_target_restore_sock_callbacks(struct iscsi_conn *conn) static int iscsi_target_do_login(struct iscsi_conn *, struct iscsi_login *); -static bool iscsi_target_sk_state_check(struct sock *sk) +static bool __iscsi_target_sk_check_close(struct sock *sk) { if (sk->sk_state == TCP_CLOSE_WAIT || sk->sk_state == TCP_CLOSE) { - pr_debug("iscsi_target_sk_state_check: TCP_CLOSE_WAIT|TCP_CLOSE," + pr_debug("__iscsi_target_sk_check_close: TCP_CLOSE_WAIT|TCP_CLOSE," "returning FALSE\n"); - return false; + return true; } - return true; + return false; +} + +static bool iscsi_target_sk_check_close(struct iscsi_conn *conn) +{ + bool state = false; + + if (conn->sock) { + struct sock *sk = conn->sock->sk; + + read_lock_bh(&sk->sk_callback_lock); + state = (__iscsi_target_sk_check_close(sk) || + test_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags)); + read_unlock_bh(&sk->sk_callback_lock); + } + return state; +} + +static bool iscsi_target_sk_check_flag(struct iscsi_conn *conn, unsigned int flag) +{ + bool state = false; + + if (conn->sock) { + struct sock *sk = conn->sock->sk; + + read_lock_bh(&sk->sk_callback_lock); + state = test_bit(flag, &conn->login_flags); + read_unlock_bh(&sk->sk_callback_lock); + } + return state; +} + +static bool iscsi_target_sk_check_and_clear(struct iscsi_conn *conn, unsigned int flag) +{ + bool state = false; + + if (conn->sock) { + struct sock *sk = conn->sock->sk; + + write_lock_bh(&sk->sk_callback_lock); + state = (__iscsi_target_sk_check_close(sk) || + test_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags)); + if (!state) + clear_bit(flag, &conn->login_flags); + write_unlock_bh(&sk->sk_callback_lock); + } + return state; } static void iscsi_target_login_drop(struct iscsi_conn *conn, struct iscsi_login *login) @@ -540,6 +586,20 @@ static void iscsi_target_do_login_rx(struct work_struct *work) pr_debug("entering iscsi_target_do_login_rx, conn: %p, %s:%d\n", conn, current->comm, current->pid); + /* + * If iscsi_target_do_login_rx() has been invoked by ->sk_data_ready() + * before initial PDU processing in iscsi_target_start_negotiation() + * has completed, go ahead and retry until it's cleared. + * + * Otherwise if the TCP connection drops while this is occuring, + * iscsi_target_start_negotiation() will detect the failure, call + * cancel_delayed_work_sync(&conn->login_work), and cleanup the + * remaining iscsi connection resources from iscsi_np process context. + */ + if (iscsi_target_sk_check_flag(conn, LOGIN_FLAGS_INITIAL_PDU)) { + schedule_delayed_work(&conn->login_work, msecs_to_jiffies(10)); + return; + } spin_lock(&tpg->tpg_state_lock); state = (tpg->tpg_state == TPG_STATE_ACTIVE); @@ -547,26 +607,12 @@ static void iscsi_target_do_login_rx(struct work_struct *work) if (!state) { pr_debug("iscsi_target_do_login_rx: tpg_state != TPG_STATE_ACTIVE\n"); - iscsi_target_restore_sock_callbacks(conn); - iscsi_target_login_drop(conn, login); - iscsit_deaccess_np(np, tpg, tpg_np); - return; + goto err; } - if (conn->sock) { - struct sock *sk = conn->sock->sk; - - read_lock_bh(&sk->sk_callback_lock); - state = iscsi_target_sk_state_check(sk); - read_unlock_bh(&sk->sk_callback_lock); - - if (!state) { - pr_debug("iscsi_target_do_login_rx, TCP state CLOSE\n"); - iscsi_target_restore_sock_callbacks(conn); - iscsi_target_login_drop(conn, login); - iscsit_deaccess_np(np, tpg, tpg_np); - return; - } + if (iscsi_target_sk_check_close(conn)) { + pr_debug("iscsi_target_do_login_rx, TCP state CLOSE\n"); + goto err; } conn->login_kworker = current; @@ -584,34 +630,29 @@ static void iscsi_target_do_login_rx(struct work_struct *work) flush_signals(current); conn->login_kworker = NULL; - if (rc < 0) { - iscsi_target_restore_sock_callbacks(conn); - iscsi_target_login_drop(conn, login); - iscsit_deaccess_np(np, tpg, tpg_np); - return; - } + if (rc < 0) + goto err; pr_debug("iscsi_target_do_login_rx after rx_login_io, %p, %s:%d\n", conn, current->comm, current->pid); rc = iscsi_target_do_login(conn, login); if (rc < 0) { - iscsi_target_restore_sock_callbacks(conn); - iscsi_target_login_drop(conn, login); - iscsit_deaccess_np(np, tpg, tpg_np); + goto err; } else if (!rc) { - if (conn->sock) { - struct sock *sk = conn->sock->sk; - - write_lock_bh(&sk->sk_callback_lock); - clear_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags); - write_unlock_bh(&sk->sk_callback_lock); - } + if (iscsi_target_sk_check_and_clear(conn, LOGIN_FLAGS_READ_ACTIVE)) + goto err; } else if (rc == 1) { iscsi_target_nego_release(conn); iscsi_post_login_handler(np, conn, zero_tsih); iscsit_deaccess_np(np, tpg, tpg_np); } + return; + +err: + iscsi_target_restore_sock_callbacks(conn); + iscsi_target_login_drop(conn, login); + iscsit_deaccess_np(np, tpg, tpg_np); } static void iscsi_target_do_cleanup(struct work_struct *work) @@ -659,31 +700,54 @@ static void iscsi_target_sk_state_change(struct sock *sk) orig_state_change(sk); return; } + state = __iscsi_target_sk_check_close(sk); + pr_debug("__iscsi_target_sk_close_change: state: %d\n", state); + if (test_bit(LOGIN_FLAGS_READ_ACTIVE, &conn->login_flags)) { pr_debug("Got LOGIN_FLAGS_READ_ACTIVE=1 sk_state_change" " conn: %p\n", conn); + if (state) + set_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags); write_unlock_bh(&sk->sk_callback_lock); orig_state_change(sk); return; } - if (test_and_set_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags)) { + if (test_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags)) { pr_debug("Got LOGIN_FLAGS_CLOSED=1 sk_state_change conn: %p\n", conn); write_unlock_bh(&sk->sk_callback_lock); orig_state_change(sk); return; } + /* + * If the TCP connection has dropped, go ahead and set LOGIN_FLAGS_CLOSED, + * but only queue conn->login_work -> iscsi_target_do_login_rx() + * processing if LOGIN_FLAGS_INITIAL_PDU has already been cleared. + * + * When iscsi_target_do_login_rx() runs, iscsi_target_sk_check_close() + * will detect the dropped TCP connection from delayed workqueue context. + * + * If LOGIN_FLAGS_INITIAL_PDU is still set, which means the initial + * iscsi_target_start_negotiation() is running, iscsi_target_do_login() + * via iscsi_target_sk_check_close() or iscsi_target_start_negotiation() + * via iscsi_target_sk_check_and_clear() is responsible for detecting the + * dropped TCP connection in iscsi_np process context, and cleaning up + * the remaining iscsi connection resources. + */ + if (state) { + pr_debug("iscsi_target_sk_state_change got failed state\n"); + set_bit(LOGIN_FLAGS_CLOSED, &conn->login_flags); + state = test_bit(LOGIN_FLAGS_INITIAL_PDU, &conn->login_flags); + write_unlock_bh(&sk->sk_callback_lock); - state = iscsi_target_sk_state_check(sk); - write_unlock_bh(&sk->sk_callback_lock); - - pr_debug("iscsi_target_sk_state_change: state: %d\n", state); + orig_state_change(sk); - if (!state) { - pr_debug("iscsi_target_sk_state_change got failed state\n"); - schedule_delayed_work(&conn->login_cleanup_work, 0); + if (!state) + schedule_delayed_work(&conn->login_work, 0); return; } + write_unlock_bh(&sk->sk_callback_lock); + orig_state_change(sk); } @@ -946,6 +1010,15 @@ static int iscsi_target_do_login(struct iscsi_conn *conn, struct iscsi_login *lo if (iscsi_target_handle_csg_one(conn, login) < 0) return -1; if (login_rsp->flags & ISCSI_FLAG_LOGIN_TRANSIT) { + /* + * Check to make sure the TCP connection has not + * dropped asynchronously while session reinstatement + * was occuring in this kthread context, before + * transitioning to full feature phase operation. + */ + if (iscsi_target_sk_check_close(conn)) + return -1; + login->tsih = conn->sess->tsih; login->login_complete = 1; iscsi_target_restore_sock_callbacks(conn); @@ -972,21 +1045,6 @@ static int iscsi_target_do_login(struct iscsi_conn *conn, struct iscsi_login *lo break; } - if (conn->sock) { - struct sock *sk = conn->sock->sk; - bool state; - - read_lock_bh(&sk->sk_callback_lock); - state = iscsi_target_sk_state_check(sk); - read_unlock_bh(&sk->sk_callback_lock); - - if (!state) { - pr_debug("iscsi_target_do_login() failed state for" - " conn: %p\n", conn); - return -1; - } - } - return 0; } @@ -1255,10 +1313,22 @@ int iscsi_target_start_negotiation( write_lock_bh(&sk->sk_callback_lock); set_bit(LOGIN_FLAGS_READY, &conn->login_flags); + set_bit(LOGIN_FLAGS_INITIAL_PDU, &conn->login_flags); write_unlock_bh(&sk->sk_callback_lock); } - + /* + * If iscsi_target_do_login returns zero to signal more PDU + * exchanges are required to complete the login, go ahead and + * clear LOGIN_FLAGS_INITIAL_PDU but only if the TCP connection + * is still active. + * + * Otherwise if TCP connection dropped asynchronously, go ahead + * and perform connection cleanup now. + */ ret = iscsi_target_do_login(conn, login); + if (!ret && iscsi_target_sk_check_and_clear(conn, LOGIN_FLAGS_INITIAL_PDU)) + ret = -1; + if (ret < 0) { cancel_delayed_work_sync(&conn->login_work); cancel_delayed_work_sync(&conn->login_cleanup_work); diff --git a/include/target/iscsi/iscsi_target_core.h b/include/target/iscsi/iscsi_target_core.h index 275581d483dd..5f17fb770477 100644 --- a/include/target/iscsi/iscsi_target_core.h +++ b/include/target/iscsi/iscsi_target_core.h @@ -557,6 +557,7 @@ struct iscsi_conn { #define LOGIN_FLAGS_READ_ACTIVE 1 #define LOGIN_FLAGS_CLOSED 2 #define LOGIN_FLAGS_READY 4 +#define LOGIN_FLAGS_INITIAL_PDU 8 unsigned long login_flags; struct delayed_work login_work; struct delayed_work login_cleanup_work; -- cgit v1.2.3 From 1410a90ae449061b7e1ae19d275148f36948801b Mon Sep 17 00:00:00 2001 From: Max Gurtovoy Date: Sun, 28 May 2017 10:53:10 +0300 Subject: net/mlx5: Define interface bits for fencing UMR wqe HW can implement UMR wqe re-transmission in various ways. Thus, add HCA cap to distinguish the needed fence for UMR to make sure that the wqe wouldn't fail on mkey checks. Signed-off-by: Max Gurtovoy Acked-by: Leon Romanovsky Reviewed-by: Christoph Hellwig Signed-off-by: Doug Ledford --- include/linux/mlx5/mlx5_ifc.h | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 32de0724b400..edafedb7b509 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -766,6 +766,12 @@ enum { MLX5_CAP_PORT_TYPE_ETH = 0x1, }; +enum { + MLX5_CAP_UMR_FENCE_STRONG = 0x0, + MLX5_CAP_UMR_FENCE_SMALL = 0x1, + MLX5_CAP_UMR_FENCE_NONE = 0x2, +}; + struct mlx5_ifc_cmd_hca_cap_bits { u8 reserved_at_0[0x80]; @@ -875,7 +881,9 @@ struct mlx5_ifc_cmd_hca_cap_bits { u8 reserved_at_202[0x1]; u8 ipoib_enhanced_offloads[0x1]; u8 ipoib_basic_offloads[0x1]; - u8 reserved_at_205[0xa]; + u8 reserved_at_205[0x5]; + u8 umr_fence[0x2]; + u8 reserved_at_20c[0x3]; u8 drain_sigerr[0x1]; u8 cmdif_checksum[0x2]; u8 sigerr_cqe[0x1]; -- cgit v1.2.3 From 233c1955835bd8649003be9bb3d8e79788b08be1 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Sun, 14 May 2017 15:49:57 +0300 Subject: RDMA/netlink: Reduce exposure of RDMA netlink functions RDMA netlink is part of ib_core, hence ibnl_chk_listeners(), ibnl_init() and ibnl_cleanup() don't need to be published in public header file. Let's remove EXPORT_SYMBOL from ibnl_chk_listeners() and move all these functions to private header file. CC: Yuval Shaia Signed-off-by: Leon Romanovsky Reviewed-by: Yuval Shaia Signed-off-by: Doug Ledford --- drivers/infiniband/core/core_priv.h | 10 ++++++++++ drivers/infiniband/core/netlink.c | 2 +- include/rdma/rdma_netlink.h | 10 ---------- 3 files changed, 11 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h index cb7d372e4bdf..d92ab4eaa8f3 100644 --- a/drivers/infiniband/core/core_priv.h +++ b/drivers/infiniband/core/core_priv.h @@ -169,6 +169,16 @@ void ib_mad_cleanup(void); int ib_sa_init(void); void ib_sa_cleanup(void); +int ibnl_init(void); +void ibnl_cleanup(void); + +/** + * Check if there are any listeners to the netlink group + * @group: the netlink group ID + * Returns 0 on success or a negative for no listeners. + */ +int ibnl_chk_listeners(unsigned int group); + int ib_nl_handle_resolve_resp(struct sk_buff *skb, struct netlink_callback *cb); int ib_nl_handle_set_timeout(struct sk_buff *skb, diff --git a/drivers/infiniband/core/netlink.c b/drivers/infiniband/core/netlink.c index b784055423c8..94931c474d41 100644 --- a/drivers/infiniband/core/netlink.c +++ b/drivers/infiniband/core/netlink.c @@ -37,6 +37,7 @@ #include #include #include +#include "core_priv.h" struct ibnl_client { struct list_head list; @@ -55,7 +56,6 @@ int ibnl_chk_listeners(unsigned int group) return -1; return 0; } -EXPORT_SYMBOL(ibnl_chk_listeners); int ibnl_add_client(int index, int nops, const struct ibnl_client_cbs cb_table[]) diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h index 585266144329..348c102cb5f6 100644 --- a/include/rdma/rdma_netlink.h +++ b/include/rdma/rdma_netlink.h @@ -10,9 +10,6 @@ struct ibnl_client_cbs { struct module *module; }; -int ibnl_init(void); -void ibnl_cleanup(void); - /** * Add a a client to the list of IB netlink exporters. * @index: Index of the added client @@ -77,11 +74,4 @@ int ibnl_unicast(struct sk_buff *skb, struct nlmsghdr *nlh, int ibnl_multicast(struct sk_buff *skb, struct nlmsghdr *nlh, unsigned int group, gfp_t flags); -/** - * Check if there are any listeners to the netlink group - * @group: the netlink group ID - * Returns 0 on success or a negative for no listeners. - */ -int ibnl_chk_listeners(unsigned int group); - #endif /* _RDMA_NETLINK_H */ -- cgit v1.2.3 From d3957b86a40612826ef935f474b31359d66cbdca Mon Sep 17 00:00:00 2001 From: Majd Dibbiny Date: Sun, 21 May 2017 19:09:54 +0300 Subject: RDMA/SA: Fix kernel panic in CMA request handler flow Commit 9fdca4da4d8c (IB/SA: Split struct sa_path_rec based on IB and ROCE specific fields) moved the service_id to be specific attribute for IB and OPA SA Path Record, and thus wasn't assigned for RoCE. This caused to the following kernel panic in the CMA request handler flow: [ 27.074594] BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 [ 27.074731] IP: __radix_tree_lookup+0x1d/0xe0 ... [ 27.075356] Workqueue: ib_cm cm_work_handler [ib_cm] [ 27.075401] task: ffff88022e3b8000 task.stack: ffffc90001298000 [ 27.075449] RIP: 0010:__radix_tree_lookup+0x1d/0xe0 ... [ 27.075979] Call Trace: [ 27.076015] radix_tree_lookup+0xd/0x10 [ 27.076055] cma_ps_find+0x59/0x70 [rdma_cm] [ 27.076097] cma_id_from_event+0xd2/0x470 [rdma_cm] [ 27.076144] ? ib_init_ah_from_path+0x39a/0x590 [ib_core] [ 27.076193] cma_req_handler+0x25/0x480 [rdma_cm] [ 27.076237] cm_process_work+0x25/0x120 [ib_cm] [ 27.076280] ? cm_get_bth_pkey.isra.62+0x3c/0xa0 [ib_cm] [ 27.076350] cm_req_handler+0xb03/0xd40 [ib_cm] [ 27.076430] ? sched_clock_cpu+0x11/0xb0 [ 27.076478] cm_work_handler+0x194/0x1588 [ib_cm] [ 27.076525] process_one_work+0x160/0x410 [ 27.076565] worker_thread+0x137/0x4a0 [ 27.076614] kthread+0x112/0x150 [ 27.076684] ? max_active_store+0x60/0x60 [ 27.077642] ? kthread_park+0x90/0x90 [ 27.078530] ret_from_fork+0x2c/0x40 This patch moves it back to the common SA Path Record structure and removes the redundant setter and getter. Tested on Connect-IB and Connect-X4 in Infiniband and RoCE respectively. Fixes: 9fdca4da4d8c (IB/SA: Split struct sa_path_rec based on IB ands ROCE specific fields) Signed-off-by: Majd Dibbiny Reviewed-by: Parav Pandit Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/core/cm.c | 4 ++-- drivers/infiniband/core/cma.c | 13 ++++++------- drivers/infiniband/core/sa_query.c | 6 +++--- drivers/infiniband/ulp/srp/ib_srp.c | 2 +- include/rdma/ib_sa.h | 25 +++---------------------- 5 files changed, 15 insertions(+), 35 deletions(-) (limited to 'include') diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index 1844770f3ae8..2b4d613a3474 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -1429,7 +1429,7 @@ static void cm_format_paths_from_req(struct cm_req_msg *req_msg, primary_path->packet_life_time = cm_req_get_primary_local_ack_timeout(req_msg); primary_path->packet_life_time -= (primary_path->packet_life_time > 0); - sa_path_set_service_id(primary_path, req_msg->service_id); + primary_path->service_id = req_msg->service_id; if (req_msg->alt_local_lid) { alt_path->dgid = req_msg->alt_local_gid; @@ -1452,7 +1452,7 @@ static void cm_format_paths_from_req(struct cm_req_msg *req_msg, alt_path->packet_life_time = cm_req_get_alt_local_ack_timeout(req_msg); alt_path->packet_life_time -= (alt_path->packet_life_time > 0); - sa_path_set_service_id(alt_path, req_msg->service_id); + alt_path->service_id = req_msg->service_id; } } diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 91b7a2fe5a55..31bb82d8ecd7 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -1140,7 +1140,7 @@ static void cma_save_ib_info(struct sockaddr *src_addr, ib->sib_pkey = path->pkey; ib->sib_flowinfo = path->flow_label; memcpy(&ib->sib_addr, &path->sgid, 16); - ib->sib_sid = sa_path_get_service_id(path); + ib->sib_sid = path->service_id; ib->sib_scope_id = 0; } else { ib->sib_pkey = listen_ib->sib_pkey; @@ -1274,8 +1274,7 @@ static int cma_save_req_info(const struct ib_cm_event *ib_event, memcpy(&req->local_gid, &req_param->primary_path->sgid, sizeof(req->local_gid)); req->has_gid = true; - req->service_id = - sa_path_get_service_id(req_param->primary_path); + req->service_id = req_param->primary_path->service_id; req->pkey = be16_to_cpu(req_param->primary_path->pkey); if (req->pkey != req_param->bth_pkey) pr_warn_ratelimited("RDMA CMA: got different BTH P_Key (0x%x) and primary path P_Key (0x%x)\n" @@ -1827,7 +1826,8 @@ static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id, struct rdma_route *rt; const sa_family_t ss_family = listen_id->route.addr.src_addr.ss_family; struct sa_path_rec *path = ib_event->param.req_rcvd.primary_path; - const __be64 service_id = sa_path_get_service_id(path); + const __be64 service_id = + ib_event->param.req_rcvd.primary_path->service_id; int ret; id = rdma_create_id(listen_id->route.addr.dev_addr.net, @@ -2345,9 +2345,8 @@ static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms, path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr)); path_rec.numb_path = 1; path_rec.reversible = 1; - sa_path_set_service_id(&path_rec, - rdma_get_service_id(&id_priv->id, - cma_dst_addr(id_priv))); + path_rec.service_id = rdma_get_service_id(&id_priv->id, + cma_dst_addr(id_priv)); comp_mask = IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID | IB_SA_PATH_REC_PKEY | IB_SA_PATH_REC_NUMB_PATH | diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c index e335b09c022e..fb7aec4047c8 100644 --- a/drivers/infiniband/core/sa_query.c +++ b/drivers/infiniband/core/sa_query.c @@ -194,7 +194,7 @@ static u32 tid; .field_name = "sa_path_rec:" #field static const struct ib_field path_rec_table[] = { - { PATH_REC_FIELD(ib.service_id), + { PATH_REC_FIELD(service_id), .offset_words = 0, .offset_bits = 0, .size_bits = 64 }, @@ -296,7 +296,7 @@ static const struct ib_field path_rec_table[] = { .field_name = "sa_path_rec:" #field static const struct ib_field opa_path_rec_table[] = { - { OPA_PATH_REC_FIELD(opa.service_id), + { OPA_PATH_REC_FIELD(service_id), .offset_words = 0, .offset_bits = 0, .size_bits = 64 }, @@ -774,7 +774,7 @@ static void ib_nl_set_path_rec_attrs(struct sk_buff *skb, /* Now build the attributes */ if (comp_mask & IB_SA_PATH_REC_SERVICE_ID) { - val64 = be64_to_cpu(sa_path_get_service_id(sa_rec)); + val64 = be64_to_cpu(sa_rec->service_id); nla_put(skb, RDMA_NLA_F_MANDATORY | LS_NLA_TYPE_SERVICE_ID, sizeof(val64), &val64); } diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 4306285fb155..2354c742caa1 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -320,7 +320,7 @@ static int srp_new_cm_id(struct srp_rdma_ch *ch) ch->path.sgid = target->sgid; ch->path.dgid = target->orig_dgid; ch->path.pkey = target->pkey; - sa_path_set_service_id(&ch->path, target->service_id); + ch->path.service_id = target->service_id; return 0; } diff --git a/include/rdma/ib_sa.h b/include/rdma/ib_sa.h index f5f70e345318..355b81f4242d 100644 --- a/include/rdma/ib_sa.h +++ b/include/rdma/ib_sa.h @@ -158,7 +158,6 @@ enum sa_path_rec_type { }; struct sa_path_rec_ib { - __be64 service_id; __be16 dlid; __be16 slid; u8 raw_traffic; @@ -174,7 +173,6 @@ struct sa_path_rec_roce { }; struct sa_path_rec_opa { - __be64 service_id; __be32 dlid; __be32 slid; u8 raw_traffic; @@ -189,6 +187,7 @@ struct sa_path_rec_opa { struct sa_path_rec { union ib_gid dgid; union ib_gid sgid; + __be64 service_id; /* reserved */ __be32 flow_label; u8 hop_limit; @@ -262,7 +261,7 @@ static inline void path_conv_opa_to_ib(struct sa_path_rec *ib, ib->ib.dlid = htons(ntohl(opa->opa.dlid)); ib->ib.slid = htons(ntohl(opa->opa.slid)); } - ib->ib.service_id = opa->opa.service_id; + ib->service_id = opa->service_id; ib->ib.raw_traffic = opa->opa.raw_traffic; } @@ -281,7 +280,7 @@ static inline void path_conv_ib_to_opa(struct sa_path_rec *opa, } opa->opa.slid = slid; opa->opa.dlid = dlid; - opa->opa.service_id = ib->ib.service_id; + opa->service_id = ib->service_id; opa->opa.raw_traffic = ib->ib.raw_traffic; } @@ -591,15 +590,6 @@ static inline bool sa_path_is_roce(struct sa_path_rec *rec) (rec->rec_type == SA_PATH_REC_TYPE_ROCE_V2)); } -static inline void sa_path_set_service_id(struct sa_path_rec *rec, - __be64 service_id) -{ - if (rec->rec_type == SA_PATH_REC_TYPE_IB) - rec->ib.service_id = service_id; - else if (rec->rec_type == SA_PATH_REC_TYPE_OPA) - rec->opa.service_id = service_id; -} - static inline void sa_path_set_slid(struct sa_path_rec *rec, __be32 slid) { if (rec->rec_type == SA_PATH_REC_TYPE_IB) @@ -625,15 +615,6 @@ static inline void sa_path_set_raw_traffic(struct sa_path_rec *rec, rec->opa.raw_traffic = raw_traffic; } -static inline __be64 sa_path_get_service_id(struct sa_path_rec *rec) -{ - if (rec->rec_type == SA_PATH_REC_TYPE_IB) - return rec->ib.service_id; - else if (rec->rec_type == SA_PATH_REC_TYPE_OPA) - return rec->opa.service_id; - return 0; -} - static inline __be32 sa_path_get_slid(struct sa_path_rec *rec) { if (rec->rec_type == SA_PATH_REC_TYPE_IB) -- cgit v1.2.3 From 1bde33e051233f0ed93a8bc67137016ab38c3d2d Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Fri, 2 Jun 2017 14:46:13 -0700 Subject: include/linux/gfp.h: fix ___GFP_NOLOCKDEP value Igor Stoppa has noticed that __GFP_NOLOCKDEP can use a lower bit. At the time commit 7e7844226f10 ("lockdep: allow to disable reclaim lockup detection") was written we still had __GFP_OTHER_NODE but I have removed it in commit 41b6167e8f74 ("mm: get rid of __GFP_OTHER_NODE") and forgot to lower the bit value. The current value is outside of __GFP_BITS_SHIFT so it cannot be used actually. Fixes: 7e7844226f10 ("lockdep: allow to disable reclaim lockup detection") Signed-off-by: Michal Hocko Reported-by: Igor Stoppa Acked-by: Vlastimil Babka Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/gfp.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 2b1a44f5bdb6..a89d37e8b387 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -41,7 +41,7 @@ struct vm_area_struct; #define ___GFP_WRITE 0x800000u #define ___GFP_KSWAPD_RECLAIM 0x1000000u #ifdef CONFIG_LOCKDEP -#define ___GFP_NOLOCKDEP 0x4000000u +#define ___GFP_NOLOCKDEP 0x2000000u #else #define ___GFP_NOLOCKDEP 0 #endif -- cgit v1.2.3 From 60b0a8c3d2480f3b57282b47b7cae7ee71c48635 Mon Sep 17 00:00:00 2001 From: Matthias Kaehlcke Date: Fri, 2 Jun 2017 14:46:16 -0700 Subject: frv: declare jiffies to be located in the .data section Commit 7c30f352c852 ("jiffies.h: declare jiffies and jiffies_64 with ____cacheline_aligned_in_smp") removed a section specification from the jiffies declaration that caused conflicts on some platforms. Unfortunately this change broke the build for frv: kernel/built-in.o: In function `__do_softirq': (.text+0x6460): relocation truncated to fit: R_FRV_GPREL12 against symbol `jiffies' defined in *ABS* section in .tmp_vmlinux1 kernel/built-in.o: In function `__do_softirq': (.text+0x6574): relocation truncated to fit: R_FRV_GPREL12 against symbol `jiffies' defined in *ABS* section in .tmp_vmlinux1 kernel/built-in.o: In function `pwq_activate_delayed_work': workqueue.c:(.text+0x15b9c): relocation truncated to fit: R_FRV_GPREL12 against symbol `jiffies' defined in *ABS* section in .tmp_vmlinux1 ... Add __jiffy_arch_data to the declaration of jiffies and use it on frv to include the section specification. For all other platforms __jiffy_arch_data (currently) has no effect. Fixes: 7c30f352c852 ("jiffies.h: declare jiffies and jiffies_64 with ____cacheline_aligned_in_smp") Link: http://lkml.kernel.org/r/20170516221333.177280-1-mka@chromium.org Signed-off-by: Matthias Kaehlcke Reported-by: Guenter Roeck Tested-by: Guenter Roeck Reviewed-by: David Howells Cc: Sudip Mukherjee Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/frv/include/asm/timex.h | 6 ++++++ include/linux/jiffies.h | 6 +++++- 2 files changed, 11 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/arch/frv/include/asm/timex.h b/arch/frv/include/asm/timex.h index a89bddefdacf..139093fab326 100644 --- a/arch/frv/include/asm/timex.h +++ b/arch/frv/include/asm/timex.h @@ -16,5 +16,11 @@ static inline cycles_t get_cycles(void) #define vxtime_lock() do {} while (0) #define vxtime_unlock() do {} while (0) +/* This attribute is used in include/linux/jiffies.h alongside with + * __cacheline_aligned_in_smp. It is assumed that __cacheline_aligned_in_smp + * for frv does not contain another section specification. + */ +#define __jiffy_arch_data __attribute__((__section__(".data"))) + #endif diff --git a/include/linux/jiffies.h b/include/linux/jiffies.h index 36872fbb815d..734377ad42e9 100644 --- a/include/linux/jiffies.h +++ b/include/linux/jiffies.h @@ -64,13 +64,17 @@ extern int register_refined_jiffies(long clock_tick_rate); /* TICK_USEC is the time between ticks in usec assuming fake USER_HZ */ #define TICK_USEC ((1000000UL + USER_HZ/2) / USER_HZ) +#ifndef __jiffy_arch_data +#define __jiffy_arch_data +#endif + /* * The 64-bit value is not atomic - you MUST NOT read it * without sampling the sequence number in jiffies_lock. * get_jiffies_64() will do this for you as appropriate. */ extern u64 __cacheline_aligned_in_smp jiffies_64; -extern unsigned long volatile __cacheline_aligned_in_smp jiffies; +extern unsigned long volatile __cacheline_aligned_in_smp __jiffy_arch_data jiffies; #if (BITS_PER_LONG < 64) u64 get_jiffies_64(void); -- cgit v1.2.3 From 9a291a7c9428155e8e623e4a3989f8be47134df5 Mon Sep 17 00:00:00 2001 From: James Morse Date: Fri, 2 Jun 2017 14:46:46 -0700 Subject: mm/hugetlb: report -EHWPOISON not -EFAULT when FOLL_HWPOISON is specified KVM uses get_user_pages() to resolve its stage2 faults. KVM sets the FOLL_HWPOISON flag causing faultin_page() to return -EHWPOISON when it finds a VM_FAULT_HWPOISON. KVM handles these hwpoison pages as a special case. (check_user_page_hwpoison()) When huge pages are involved, this doesn't work so well. get_user_pages() calls follow_hugetlb_page(), which stops early if it receives VM_FAULT_HWPOISON from hugetlb_fault(), eventually returning -EFAULT to the caller. The step to map this to -EHWPOISON based on the FOLL_ flags is missing. The hwpoison special case is skipped, and -EFAULT is returned to user-space, causing Qemu or kvmtool to exit. Instead, move this VM_FAULT_ to errno mapping code into a header file and use it from faultin_page() and follow_hugetlb_page(). With this, KVM works as expected. This isn't a problem for arm64 today as we haven't enabled MEMORY_FAILURE, but I can't see any reason this doesn't happen on x86 too, so I think this should be a fix. This doesn't apply earlier than stable's v4.11.1 due to all sorts of cleanup. [james.morse@arm.com: add vm_fault_to_errno() call to faultin_page()] suggested. Link: http://lkml.kernel.org/r/20170525171035.16359-1-james.morse@arm.com [akpm@linux-foundation.org: coding-style fixes] Link: http://lkml.kernel.org/r/20170524160900.28786-1-james.morse@arm.com Signed-off-by: James Morse Acked-by: Punit Agrawal Acked-by: Naoya Horiguchi Cc: "Kirill A . Shutemov" Cc: [4.11.1+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mm.h | 11 +++++++++++ mm/gup.c | 20 ++++++++------------ mm/hugetlb.c | 5 +++++ 3 files changed, 24 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/include/linux/mm.h b/include/linux/mm.h index 7cb17c6b97de..b892e95d4929 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2327,6 +2327,17 @@ static inline struct page *follow_page(struct vm_area_struct *vma, #define FOLL_REMOTE 0x2000 /* we are working on non-current tsk/mm */ #define FOLL_COW 0x4000 /* internal GUP flag */ +static inline int vm_fault_to_errno(int vm_fault, int foll_flags) +{ + if (vm_fault & VM_FAULT_OOM) + return -ENOMEM; + if (vm_fault & (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) + return (foll_flags & FOLL_HWPOISON) ? -EHWPOISON : -EFAULT; + if (vm_fault & (VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV)) + return -EFAULT; + return 0; +} + typedef int (*pte_fn_t)(pte_t *pte, pgtable_t token, unsigned long addr, void *data); extern int apply_to_page_range(struct mm_struct *mm, unsigned long address, diff --git a/mm/gup.c b/mm/gup.c index d9e6fddcc51f..b3c7214d710d 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -407,12 +407,10 @@ static int faultin_page(struct task_struct *tsk, struct vm_area_struct *vma, ret = handle_mm_fault(vma, address, fault_flags); if (ret & VM_FAULT_ERROR) { - if (ret & VM_FAULT_OOM) - return -ENOMEM; - if (ret & (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) - return *flags & FOLL_HWPOISON ? -EHWPOISON : -EFAULT; - if (ret & (VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV)) - return -EFAULT; + int err = vm_fault_to_errno(ret, *flags); + + if (err) + return err; BUG(); } @@ -723,12 +721,10 @@ retry: ret = handle_mm_fault(vma, address, fault_flags); major |= ret & VM_FAULT_MAJOR; if (ret & VM_FAULT_ERROR) { - if (ret & VM_FAULT_OOM) - return -ENOMEM; - if (ret & (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) - return -EHWPOISON; - if (ret & (VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV)) - return -EFAULT; + int err = vm_fault_to_errno(ret, 0); + + if (err) + return err; BUG(); } diff --git a/mm/hugetlb.c b/mm/hugetlb.c index e5828875f7bb..3eedb187e549 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4170,6 +4170,11 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, } ret = hugetlb_fault(mm, vma, vaddr, fault_flags); if (ret & VM_FAULT_ERROR) { + int err = vm_fault_to_errno(ret, flags); + + if (err) + return err; + remainder = 0; break; } -- cgit v1.2.3 From 864b9a393dcb5aed09b8fd31b9bbda0fdda99374 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Fri, 2 Jun 2017 14:46:49 -0700 Subject: mm: consider memblock reservations for deferred memory initialization sizing We have seen an early OOM killer invocation on ppc64 systems with crashkernel=4096M: kthreadd invoked oom-killer: gfp_mask=0x16040c0(GFP_KERNEL|__GFP_COMP|__GFP_NOTRACK), nodemask=7, order=0, oom_score_adj=0 kthreadd cpuset=/ mems_allowed=7 CPU: 0 PID: 2 Comm: kthreadd Not tainted 4.4.68-1.gd7fe927-default #1 Call Trace: dump_stack+0xb0/0xf0 (unreliable) dump_header+0xb0/0x258 out_of_memory+0x5f0/0x640 __alloc_pages_nodemask+0xa8c/0xc80 kmem_getpages+0x84/0x1a0 fallback_alloc+0x2a4/0x320 kmem_cache_alloc_node+0xc0/0x2e0 copy_process.isra.25+0x260/0x1b30 _do_fork+0x94/0x470 kernel_thread+0x48/0x60 kthreadd+0x264/0x330 ret_from_kernel_thread+0x5c/0xa4 Mem-Info: active_anon:0 inactive_anon:0 isolated_anon:0 active_file:0 inactive_file:0 isolated_file:0 unevictable:0 dirty:0 writeback:0 unstable:0 slab_reclaimable:5 slab_unreclaimable:73 mapped:0 shmem:0 pagetables:0 bounce:0 free:0 free_pcp:0 free_cma:0 Node 7 DMA free:0kB min:0kB low:0kB high:0kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:52428800kB managed:110016kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB slab_reclaimable:320kB slab_unreclaimable:4672kB kernel_stack:1152kB pagetables:0kB unstable:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? yes lowmem_reserve[]: 0 0 0 0 Node 7 DMA: 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB 0*8192kB 0*16384kB = 0kB 0 total pagecache pages 0 pages in swap cache Swap cache stats: add 0, delete 0, find 0/0 Free swap = 0kB Total swap = 0kB 819200 pages RAM 0 pages HighMem/MovableOnly 817481 pages reserved 0 pages cma reserved 0 pages hwpoisoned the reason is that the managed memory is too low (only 110MB) while the rest of the the 50GB is still waiting for the deferred intialization to be done. update_defer_init estimates the initial memoty to initialize to 2GB at least but it doesn't consider any memory allocated in that range. In this particular case we've had Reserving 4096MB of memory at 128MB for crashkernel (System RAM: 51200MB) so the low 2GB is mostly depleted. Fix this by considering memblock allocations in the initial static initialization estimation. Move the max_initialise to reset_deferred_meminit and implement a simple memblock_reserved_memory helper which iterates all reserved blocks and sums the size of all that start below the given address. The cumulative size is than added on top of the initial estimation. This is still not ideal because reset_deferred_meminit doesn't consider holes and so reservation might be above the initial estimation whihch we ignore but let's make the logic simpler until we really need to handle more complicated cases. Fixes: 3a80a7fa7989 ("mm: meminit: initialise a subset of struct pages if CONFIG_DEFERRED_STRUCT_PAGE_INIT is set") Link: http://lkml.kernel.org/r/20170531104010.GI27783@dhcp22.suse.cz Signed-off-by: Michal Hocko Acked-by: Mel Gorman Tested-by: Srikar Dronamraju Cc: [4.2+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/memblock.h | 8 ++++++++ include/linux/mmzone.h | 1 + mm/memblock.c | 23 +++++++++++++++++++++++ mm/page_alloc.c | 33 ++++++++++++++++++++++----------- 4 files changed, 54 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/include/linux/memblock.h b/include/linux/memblock.h index 4ce24a376262..8098695e5d8d 100644 --- a/include/linux/memblock.h +++ b/include/linux/memblock.h @@ -425,12 +425,20 @@ static inline void early_memtest(phys_addr_t start, phys_addr_t end) } #endif +extern unsigned long memblock_reserved_memory_within(phys_addr_t start_addr, + phys_addr_t end_addr); #else static inline phys_addr_t memblock_alloc(phys_addr_t size, phys_addr_t align) { return 0; } +static inline unsigned long memblock_reserved_memory_within(phys_addr_t start_addr, + phys_addr_t end_addr) +{ + return 0; +} + #endif /* CONFIG_HAVE_MEMBLOCK */ #endif /* __KERNEL__ */ diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index ebaccd4e7d8c..ef6a13b7bd3e 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -678,6 +678,7 @@ typedef struct pglist_data { * is the first PFN that needs to be initialised. */ unsigned long first_deferred_pfn; + unsigned long static_init_size; #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */ #ifdef CONFIG_TRANSPARENT_HUGEPAGE diff --git a/mm/memblock.c b/mm/memblock.c index b049c9b2dba8..7b8a5db76a2f 100644 --- a/mm/memblock.c +++ b/mm/memblock.c @@ -1739,6 +1739,29 @@ static void __init_memblock memblock_dump(struct memblock_type *type) } } +extern unsigned long __init_memblock +memblock_reserved_memory_within(phys_addr_t start_addr, phys_addr_t end_addr) +{ + struct memblock_region *rgn; + unsigned long size = 0; + int idx; + + for_each_memblock_type((&memblock.reserved), rgn) { + phys_addr_t start, end; + + if (rgn->base + rgn->size < start_addr) + continue; + if (rgn->base > end_addr) + continue; + + start = rgn->base; + end = start + rgn->size; + size += end - start; + } + + return size; +} + void __init_memblock __memblock_dump_all(void) { pr_info("MEMBLOCK configuration:\n"); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index b7a6f583a373..2302f250d6b1 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -292,6 +292,26 @@ int page_group_by_mobility_disabled __read_mostly; #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT static inline void reset_deferred_meminit(pg_data_t *pgdat) { + unsigned long max_initialise; + unsigned long reserved_lowmem; + + /* + * Initialise at least 2G of a node but also take into account that + * two large system hashes that can take up 1GB for 0.25TB/node. + */ + max_initialise = max(2UL << (30 - PAGE_SHIFT), + (pgdat->node_spanned_pages >> 8)); + + /* + * Compensate the all the memblock reservations (e.g. crash kernel) + * from the initial estimation to make sure we will initialize enough + * memory to boot. + */ + reserved_lowmem = memblock_reserved_memory_within(pgdat->node_start_pfn, + pgdat->node_start_pfn + max_initialise); + max_initialise += reserved_lowmem; + + pgdat->static_init_size = min(max_initialise, pgdat->node_spanned_pages); pgdat->first_deferred_pfn = ULONG_MAX; } @@ -314,20 +334,11 @@ static inline bool update_defer_init(pg_data_t *pgdat, unsigned long pfn, unsigned long zone_end, unsigned long *nr_initialised) { - unsigned long max_initialise; - /* Always populate low zones for address-contrained allocations */ if (zone_end < pgdat_end_pfn(pgdat)) return true; - /* - * Initialise at least 2G of a node but also take into account that - * two large system hashes that can take up 1GB for 0.25TB/node. - */ - max_initialise = max(2UL << (30 - PAGE_SHIFT), - (pgdat->node_spanned_pages >> 8)); - (*nr_initialised)++; - if ((*nr_initialised > max_initialise) && + if ((*nr_initialised > pgdat->static_init_size) && (pfn & (PAGES_PER_SECTION - 1)) == 0) { pgdat->first_deferred_pfn = pfn; return false; @@ -6138,7 +6149,6 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size, /* pg_data_t should be reset to zero when it's allocated */ WARN_ON(pgdat->nr_zones || pgdat->kswapd_classzone_idx); - reset_deferred_meminit(pgdat); pgdat->node_id = nid; pgdat->node_start_pfn = node_start_pfn; pgdat->per_cpu_nodestats = NULL; @@ -6160,6 +6170,7 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size, (unsigned long)pgdat->node_mem_map); #endif + reset_deferred_meminit(pgdat); free_area_init_core(pgdat); } -- cgit v1.2.3