From da142f3d373a6ddaca0119615a8db2175ddc4121 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 5 Dec 2025 15:26:55 -0800 Subject: KVM: Remove subtle "struct kvm_stats_desc" pseudo-overlay Remove KVM's internal pseudo-overlay of kvm_stats_desc, which subtly aliases the flexible name[] in the uAPI definition with a fixed-size array of the same name. The unusual embedded structure results in compiler warnings due to -Wflex-array-member-not-at-end, and also necessitates an extra level of dereferencing in KVM. To avoid the "overlay", define the uAPI structure to have a fixed-size name when building for the kernel. Opportunistically clean up the indentation for the stats macros, and replace spaces with tabs. No functional change intended. Reported-by: Gustavo A. R. Silva Closes: https://lore.kernel.org/all/aPfNKRpLfhmhYqfP@kspp Acked-by: Marc Zyngier Acked-by: Christian Borntraeger [..] Acked-by: Anup Patel Reviewed-by: Bibo Mao Acked-by: Gustavo A. R. Silva Link: https://patch.msgid.link/20251205232655.445294-1-seanjc@google.com Signed-off-by: Sean Christopherson --- include/linux/kvm_host.h | 83 ++++++++++++++++++++---------------------------- 1 file changed, 35 insertions(+), 48 deletions(-) (limited to 'include/linux') diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index d93f75b05ae2..7428d9949382 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -1927,56 +1927,43 @@ enum kvm_stat_kind { struct kvm_stat_data { struct kvm *kvm; - const struct _kvm_stats_desc *desc; + const struct kvm_stats_desc *desc; enum kvm_stat_kind kind; }; -struct _kvm_stats_desc { - struct kvm_stats_desc desc; - char name[KVM_STATS_NAME_SIZE]; -}; - -#define STATS_DESC_COMMON(type, unit, base, exp, sz, bsz) \ - .flags = type | unit | base | \ - BUILD_BUG_ON_ZERO(type & ~KVM_STATS_TYPE_MASK) | \ - BUILD_BUG_ON_ZERO(unit & ~KVM_STATS_UNIT_MASK) | \ - BUILD_BUG_ON_ZERO(base & ~KVM_STATS_BASE_MASK), \ - .exponent = exp, \ - .size = sz, \ +#define STATS_DESC_COMMON(type, unit, base, exp, sz, bsz) \ + .flags = type | unit | base | \ + BUILD_BUG_ON_ZERO(type & ~KVM_STATS_TYPE_MASK) | \ + BUILD_BUG_ON_ZERO(unit & ~KVM_STATS_UNIT_MASK) | \ + BUILD_BUG_ON_ZERO(base & ~KVM_STATS_BASE_MASK), \ + .exponent = exp, \ + .size = sz, \ .bucket_size = bsz -#define VM_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vm_stat, generic.stat) \ - }, \ - .name = #stat, \ - } -#define VCPU_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vcpu_stat, generic.stat) \ - }, \ - .name = #stat, \ - } -#define VM_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vm_stat, stat) \ - }, \ - .name = #stat, \ - } -#define VCPU_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ - { \ - { \ - STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ - .offset = offsetof(struct kvm_vcpu_stat, stat) \ - }, \ - .name = #stat, \ - } +#define VM_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vm_stat, generic.stat), \ + .name = #stat, \ +} +#define VCPU_GENERIC_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vcpu_stat, generic.stat), \ + .name = #stat, \ +} +#define VM_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vm_stat, stat), \ + .name = #stat, \ +} +#define VCPU_STATS_DESC(stat, type, unit, base, exp, sz, bsz) \ +{ \ + STATS_DESC_COMMON(type, unit, base, exp, sz, bsz), \ + .offset = offsetof(struct kvm_vcpu_stat, stat), \ + .name = #stat, \ +} /* SCOPE: VM, VM_GENERIC, VCPU, VCPU_GENERIC */ #define STATS_DESC(SCOPE, stat, type, unit, base, exp, sz, bsz) \ SCOPE##_STATS_DESC(stat, type, unit, base, exp, sz, bsz) @@ -2053,7 +2040,7 @@ struct _kvm_stats_desc { STATS_DESC_IBOOLEAN(VCPU_GENERIC, blocking) ssize_t kvm_stats_read(char *id, const struct kvm_stats_header *header, - const struct _kvm_stats_desc *desc, + const struct kvm_stats_desc *desc, void *stats, size_t size_stats, char __user *user_buffer, size_t size, loff_t *offset); @@ -2098,9 +2085,9 @@ static inline void kvm_stats_log_hist_update(u64 *data, size_t size, u64 value) extern const struct kvm_stats_header kvm_vm_stats_header; -extern const struct _kvm_stats_desc kvm_vm_stats_desc[]; +extern const struct kvm_stats_desc kvm_vm_stats_desc[]; extern const struct kvm_stats_header kvm_vcpu_stats_header; -extern const struct _kvm_stats_desc kvm_vcpu_stats_desc[]; +extern const struct kvm_stats_desc kvm_vcpu_stats_desc[]; #ifdef CONFIG_KVM_GENERIC_MMU_NOTIFIER static inline int mmu_invalidate_retry(struct kvm *kvm, unsigned long mmu_seq) -- cgit v1.2.3 From 2d28ed588f8d7d0d41b0a4fad7f0d05e4bbf1797 Mon Sep 17 00:00:00 2001 From: Axel Rasmussen Date: Tue, 24 Feb 2026 16:24:34 -0800 Subject: Revert "ptdesc: remove references to folios from __pagetable_ctor() and pagetable_dtor()" This change swapped out mod_node_page_state for lruvec_stat_add_folio. But, these two APIs are not interchangeable: the lruvec version also increments memcg stats, in addition to "global" pgdat stats. So after this change, the "pagetables" memcg stat in memory.stat always yields "0", which is a userspace visible regression. I tried to look for a refactor where we add a variant of lruvec_stat_mod_folio which takes a pgdat and a memcg instead of a folio, to try to adhere to the spirit of the original patch. But at the end of the day this just means we have to call folio_memcg(ptdesc_folio(ptdesc)) anyway, which doesn't really accomplish much. This regression is visible in master as well as 6.18 stable, so CC stable too. Link: https://lkml.kernel.org/r/20260225002434.2953895-1-axelrasmussen@google.com Fixes: f0c92726e89f ("ptdesc: remove references to folios from __pagetable_ctor() and pagetable_dtor()") Signed-off-by: Axel Rasmussen Acked-by: Shakeel Butt Acked-by: Johannes Weiner Reviewed-by: Vishal Moola (Oracle) Cc: David Hildenbrand Cc: Liam Howlett Cc: Lorenzo Stoakes Cc: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Mike Rapoport Cc: Suren Baghdasaryan Cc: Vlastimil Babka Cc: Roman Gushchin Cc: Muchun Song Cc: Signed-off-by: Andrew Morton --- include/linux/mm.h | 17 ++++++----------- 1 file changed, 6 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 5be3d8a8f806..abb4963c1f06 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -3514,26 +3514,21 @@ static inline bool ptlock_init(struct ptdesc *ptdesc) { return true; } static inline void ptlock_free(struct ptdesc *ptdesc) {} #endif /* defined(CONFIG_SPLIT_PTE_PTLOCKS) */ -static inline unsigned long ptdesc_nr_pages(const struct ptdesc *ptdesc) -{ - return compound_nr(ptdesc_page(ptdesc)); -} - static inline void __pagetable_ctor(struct ptdesc *ptdesc) { - pg_data_t *pgdat = NODE_DATA(memdesc_nid(ptdesc->pt_flags)); + struct folio *folio = ptdesc_folio(ptdesc); - __SetPageTable(ptdesc_page(ptdesc)); - mod_node_page_state(pgdat, NR_PAGETABLE, ptdesc_nr_pages(ptdesc)); + __folio_set_pgtable(folio); + lruvec_stat_add_folio(folio, NR_PAGETABLE); } static inline void pagetable_dtor(struct ptdesc *ptdesc) { - pg_data_t *pgdat = NODE_DATA(memdesc_nid(ptdesc->pt_flags)); + struct folio *folio = ptdesc_folio(ptdesc); ptlock_free(ptdesc); - __ClearPageTable(ptdesc_page(ptdesc)); - mod_node_page_state(pgdat, NR_PAGETABLE, -ptdesc_nr_pages(ptdesc)); + __folio_clear_pgtable(folio); + lruvec_stat_sub_folio(folio, NR_PAGETABLE); } static inline void pagetable_dtor_free(struct ptdesc *ptdesc) -- cgit v1.2.3 From 7392f8e4ea632622b2cd2086675ba022db238b3a Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 1 Mar 2026 16:52:29 -0800 Subject: uaccess: correct kernel-doc parameter format Use the correct kernel-doc function parameter format to avoid kernel-doc warnings: Warning: include/linux/uaccess.h:814 function parameter 'uptr' not described in 'scoped_user_rw_access_size' Warning: include/linux/uaccess.h:826 function parameter 'uptr' not described in 'scoped_user_rw_access' Link: https://lkml.kernel.org/r/20260302005229.3471955-1-rdunlap@infradead.org Signed-off-by: Randy Dunlap Reviewed-by: Andrew Morton Signed-off-by: Andrew Morton --- include/linux/uaccess.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h index 1f3804245c06..001cfef21b61 100644 --- a/include/linux/uaccess.h +++ b/include/linux/uaccess.h @@ -806,7 +806,7 @@ for (bool done = false; !done; done = true) \ /** * scoped_user_rw_access_size - Start a scoped user read/write access with given size - * @uptr Pointer to the user space address to read from and write to + * @uptr: Pointer to the user space address to read from and write to * @size: Size of the access starting from @uptr * @elbl: Error label to goto when the access region is rejected * @@ -817,7 +817,7 @@ for (bool done = false; !done; done = true) \ /** * scoped_user_rw_access - Start a scoped user read/write access - * @uptr Pointer to the user space address to read from and write to + * @uptr: Pointer to the user space address to read from and write to * @elbl: Error label to goto when the access region is rejected * * The size of the access starting from @uptr is determined via sizeof(*@uptr)). -- cgit v1.2.3 From 599b4e290c8766b19378d85d4310c6ec8f90ade4 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 1 Mar 2026 16:52:22 -0800 Subject: mm/mmu_notifier: clean up mmu_notifier.h kernel-doc Eliminate kernel-doc warnings in mmu_notifier.h: - add a missing struct short description - use the correct format for function parameters - add missing function return comment sections Warning: include/linux/mmu_notifier.h:236 missing initial short description on line: * struct mmu_interval_notifier_ops Warning: include/linux/mmu_notifier.h:325 function parameter 'interval_sub' not described in 'mmu_interval_set_seq' Warning: include/linux/mmu_notifier.h:325 function parameter 'cur_seq' not described in 'mmu_interval_set_seq' Warning: include/linux/mmu_notifier.h:346 function parameter 'interval_sub' not described in 'mmu_interval_read_retry' Warning: include/linux/mmu_notifier.h:346 function parameter 'seq' not described in 'mmu_interval_read_retry' Warning: include/linux/mmu_notifier.h:346 No description found for return value of 'mmu_interval_read_retry' Warning: include/linux/mmu_notifier.h:370 function parameter 'interval_sub' not described in 'mmu_interval_check_retry' Warning: include/linux/mmu_notifier.h:370 function parameter 'seq' not described in 'mmu_interval_check_retry' Warning: include/linux/mmu_notifier.h:370 No description found for return value of 'mmu_interval_check_retry' Link: https://lkml.kernel.org/r/20260302005222.3470783-1-rdunlap@infradead.org Signed-off-by: Randy Dunlap Reviewed-by: Jason Gunthorpe Cc: David Hildenbrand Cc: "Liam R. Howlett" Cc: Lorenzo Stoakes Cc: Michal Hocko Cc: Mike Rapoport Cc: Randy Dunlap Cc: Suren Baghdasaryan Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- include/linux/mmu_notifier.h | 31 ++++++++++++++++--------------- 1 file changed, 16 insertions(+), 15 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h index 07a2bbaf86e9..8450e18a87c2 100644 --- a/include/linux/mmu_notifier.h +++ b/include/linux/mmu_notifier.h @@ -234,7 +234,7 @@ struct mmu_notifier { }; /** - * struct mmu_interval_notifier_ops + * struct mmu_interval_notifier_ops - callback for range notification * @invalidate: Upon return the caller must stop using any SPTEs within this * range. This function can sleep. Return false only if sleeping * was required but mmu_notifier_range_blockable(range) is false. @@ -309,8 +309,8 @@ void mmu_interval_notifier_remove(struct mmu_interval_notifier *interval_sub); /** * mmu_interval_set_seq - Save the invalidation sequence - * @interval_sub - The subscription passed to invalidate - * @cur_seq - The cur_seq passed to the invalidate() callback + * @interval_sub: The subscription passed to invalidate + * @cur_seq: The cur_seq passed to the invalidate() callback * * This must be called unconditionally from the invalidate callback of a * struct mmu_interval_notifier_ops under the same lock that is used to call @@ -329,8 +329,8 @@ mmu_interval_set_seq(struct mmu_interval_notifier *interval_sub, /** * mmu_interval_read_retry - End a read side critical section against a VA range - * interval_sub: The subscription - * seq: The return of the paired mmu_interval_read_begin() + * @interval_sub: The subscription + * @seq: The return of the paired mmu_interval_read_begin() * * This MUST be called under a user provided lock that is also held * unconditionally by op->invalidate() when it calls mmu_interval_set_seq(). @@ -338,7 +338,7 @@ mmu_interval_set_seq(struct mmu_interval_notifier *interval_sub, * Each call should be paired with a single mmu_interval_read_begin() and * should be used to conclude the read side. * - * Returns true if an invalidation collided with this critical section, and + * Returns: true if an invalidation collided with this critical section, and * the caller should retry. */ static inline bool @@ -350,20 +350,21 @@ mmu_interval_read_retry(struct mmu_interval_notifier *interval_sub, /** * mmu_interval_check_retry - Test if a collision has occurred - * interval_sub: The subscription - * seq: The return of the matching mmu_interval_read_begin() + * @interval_sub: The subscription + * @seq: The return of the matching mmu_interval_read_begin() * * This can be used in the critical section between mmu_interval_read_begin() - * and mmu_interval_read_retry(). A return of true indicates an invalidation - * has collided with this critical region and a future - * mmu_interval_read_retry() will return true. - * - * False is not reliable and only suggests a collision may not have - * occurred. It can be called many times and does not have to hold the user - * provided lock. + * and mmu_interval_read_retry(). * * This call can be used as part of loops and other expensive operations to * expedite a retry. + * It can be called many times and does not have to hold the user + * provided lock. + * + * Returns: true indicates an invalidation has collided with this critical + * region and a future mmu_interval_read_retry() will return true. + * False is not reliable and only suggests a collision may not have + * occurred. */ static inline bool mmu_interval_check_retry(struct mmu_interval_notifier *interval_sub, -- cgit v1.2.3 From 55f854dd5bdd8e19b936a00ef1f8d776ac32c7b0 Mon Sep 17 00:00:00 2001 From: Laurent Vivier Date: Wed, 4 Mar 2026 14:43:38 +0100 Subject: qmi_wwan: allow max_mtu above hard_mtu to control rx_urb_size Commit c7159e960f14 ("usbnet: limit max_mtu based on device's hard_mtu") capped net->max_mtu to the device's hard_mtu in usbnet_probe(). While this correctly prevents oversized packets on standard USB network devices, it breaks the qmi_wwan driver. qmi_wwan relies on userspace (e.g. ModemManager) setting a large MTU on the wwan0 interface to configure rx_urb_size via usbnet_change_mtu(). QMI modems negotiate USB transfer sizes of 16,383 or 32,767 bytes, and the USB receive buffers must be sized accordingly. With max_mtu capped to hard_mtu (~1500 bytes), userspace can no longer raise the MTU, the receive buffers remain small, and download speeds drop from >300 Mbps to ~0.8 Mbps. Introduce a FLAG_NOMAXMTU driver flag that allows individual usbnet drivers to opt out of the max_mtu cap. Set this flag in qmi_wwan's driver_info structures to restore the previous behavior for QMI devices, while keeping the safety fix in place for all other usbnet drivers. Fixes: c7159e960f14 ("usbnet: limit max_mtu based on device's hard_mtu") Cc: stable@vger.kernel.org Link: https://lore.kernel.org/lkml/CAPh3n803k8JcBPV5qEzUB-oKzWkAs-D5CU7z=Vd_nLRCr5ZqQg@mail.gmail.com/ Reported-by: Koen Vandeputte Tested-by: Daniele Palmas Signed-off-by: Laurent Vivier Link: https://patch.msgid.link/20260304134338.1785002-1-lvivier@redhat.com Signed-off-by: Jakub Kicinski --- drivers/net/usb/qmi_wwan.c | 4 ++-- drivers/net/usb/usbnet.c | 7 ++++--- include/linux/usb/usbnet.h | 1 + 3 files changed, 7 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c index 3a4985b582cb..05acac10cd2b 100644 --- a/drivers/net/usb/qmi_wwan.c +++ b/drivers/net/usb/qmi_wwan.c @@ -928,7 +928,7 @@ err: static const struct driver_info qmi_wwan_info = { .description = "WWAN/QMI device", - .flags = FLAG_WWAN | FLAG_SEND_ZLP, + .flags = FLAG_WWAN | FLAG_NOMAXMTU | FLAG_SEND_ZLP, .bind = qmi_wwan_bind, .unbind = qmi_wwan_unbind, .manage_power = qmi_wwan_manage_power, @@ -937,7 +937,7 @@ static const struct driver_info qmi_wwan_info = { static const struct driver_info qmi_wwan_info_quirk_dtr = { .description = "WWAN/QMI device", - .flags = FLAG_WWAN | FLAG_SEND_ZLP, + .flags = FLAG_WWAN | FLAG_NOMAXMTU | FLAG_SEND_ZLP, .bind = qmi_wwan_bind, .unbind = qmi_wwan_unbind, .manage_power = qmi_wwan_manage_power, diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c index ed86ba87ca4e..b72ba0803392 100644 --- a/drivers/net/usb/usbnet.c +++ b/drivers/net/usb/usbnet.c @@ -1829,11 +1829,12 @@ usbnet_probe(struct usb_interface *udev, const struct usb_device_id *prod) if ((dev->driver_info->flags & FLAG_NOARP) != 0) net->flags |= IFF_NOARP; - if (net->max_mtu > (dev->hard_mtu - net->hard_header_len)) + if ((dev->driver_info->flags & FLAG_NOMAXMTU) == 0 && + net->max_mtu > (dev->hard_mtu - net->hard_header_len)) net->max_mtu = dev->hard_mtu - net->hard_header_len; - if (net->mtu > net->max_mtu) - net->mtu = net->max_mtu; + if (net->mtu > (dev->hard_mtu - net->hard_header_len)) + net->mtu = dev->hard_mtu - net->hard_header_len; } else if (!info->in || !info->out) status = usbnet_get_endpoints(dev, udev); diff --git a/include/linux/usb/usbnet.h b/include/linux/usb/usbnet.h index b0e84896e6ac..bbf799ccf3b3 100644 --- a/include/linux/usb/usbnet.h +++ b/include/linux/usb/usbnet.h @@ -132,6 +132,7 @@ struct driver_info { #define FLAG_MULTI_PACKET 0x2000 #define FLAG_RX_ASSEMBLE 0x4000 /* rx packets may span >1 frames */ #define FLAG_NOARP 0x8000 /* device can't do ARP */ +#define FLAG_NOMAXMTU 0x10000 /* allow max_mtu above hard_mtu */ /* init device ... can sleep, or cause probe() failure */ int (*bind)(struct usbnet *, struct usb_interface *); -- cgit v1.2.3 From 6f1a9140ecda3baba3d945b9a6155af4268aafc4 Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Sat, 7 Mar 2026 00:01:34 +0800 Subject: net: add xmit recursion limit to tunnel xmit functions Tunnel xmit functions (iptunnel_xmit, ip6tunnel_xmit) lack their own recursion limit. When a bond device in broadcast mode has GRE tap interfaces as slaves, and those GRE tunnels route back through the bond, multicast/broadcast traffic triggers infinite recursion between bond_xmit_broadcast() and ip_tunnel_xmit()/ip6_tnl_xmit(), causing kernel stack overflow. The existing XMIT_RECURSION_LIMIT (8) in the no-qdisc path is not sufficient because tunnel recursion involves route lookups and full IP output, consuming much more stack per level. Use a lower limit of 4 (IP_TUNNEL_RECURSION_LIMIT) to prevent overflow. Add recursion detection using dev_xmit_recursion helpers directly in iptunnel_xmit() and ip6tunnel_xmit() to cover all IPv4/IPv6 tunnel paths including UDP encapsulated tunnels (VXLAN, Geneve, etc.). Move dev_xmit_recursion helpers from net/core/dev.h to public header include/linux/netdevice.h so they can be used by tunnel code. BUG: KASAN: stack-out-of-bounds in blake2s.constprop.0+0xe7/0x160 Write of size 32 at addr ffff88810033fed0 by task kworker/0:1/11 Workqueue: mld mld_ifc_work Call Trace: __build_flow_key.constprop.0 (net/ipv4/route.c:515) ip_rt_update_pmtu (net/ipv4/route.c:1073) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:84) ip_tunnel_xmit (net/ipv4/ip_tunnel.c:847) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) ip_finish_output2 (net/ipv4/ip_output.c:237) ip_output (net/ipv4/ip_output.c:438) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:86) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) ip_finish_output2 (net/ipv4/ip_output.c:237) ip_output (net/ipv4/ip_output.c:438) iptunnel_xmit (net/ipv4/ip_tunnel_core.c:86) ip_tunnel_xmit (net/ipv4/ip_tunnel.c:847) gre_tap_xmit (net/ipv4/ip_gre.c:779) dev_hard_start_xmit (net/core/dev.c:3887) sch_direct_xmit (net/sched/sch_generic.c:347) __dev_queue_xmit (net/core/dev.c:4802) bond_dev_queue_xmit (drivers/net/bonding/bond_main.c:312) bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5279) bond_start_xmit (drivers/net/bonding/bond_main.c:5530) dev_hard_start_xmit (net/core/dev.c:3887) __dev_queue_xmit (net/core/dev.c:4841) mld_sendpack mld_ifc_work process_one_work worker_thread Fixes: 745e20f1b626 ("net: add a recursion limit in xmit path") Reported-by: Xiang Mei Signed-off-by: Weiming Shi Link: https://patch.msgid.link/20260306160133.3852900-2-bestswngs@gmail.com Signed-off-by: Paolo Abeni --- include/linux/netdevice.h | 32 ++++++++++++++++++++++++++++++++ include/net/ip6_tunnel.h | 12 ++++++++++++ include/net/ip_tunnels.h | 7 +++++++ net/core/dev.h | 35 ----------------------------------- net/ipv4/ip_tunnel_core.c | 13 +++++++++++++ 5 files changed, 64 insertions(+), 35 deletions(-) (limited to 'include/linux') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 67e25f6d15a4..ae269a2e7f4d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3576,17 +3576,49 @@ struct page_pool_bh { }; DECLARE_PER_CPU(struct page_pool_bh, system_page_pool); +#define XMIT_RECURSION_LIMIT 8 + #ifndef CONFIG_PREEMPT_RT static inline int dev_recursion_level(void) { return this_cpu_read(softnet_data.xmit.recursion); } + +static inline bool dev_xmit_recursion(void) +{ + return unlikely(__this_cpu_read(softnet_data.xmit.recursion) > + XMIT_RECURSION_LIMIT); +} + +static inline void dev_xmit_recursion_inc(void) +{ + __this_cpu_inc(softnet_data.xmit.recursion); +} + +static inline void dev_xmit_recursion_dec(void) +{ + __this_cpu_dec(softnet_data.xmit.recursion); +} #else static inline int dev_recursion_level(void) { return current->net_xmit.recursion; } +static inline bool dev_xmit_recursion(void) +{ + return unlikely(current->net_xmit.recursion > XMIT_RECURSION_LIMIT); +} + +static inline void dev_xmit_recursion_inc(void) +{ + current->net_xmit.recursion++; +} + +static inline void dev_xmit_recursion_dec(void) +{ + current->net_xmit.recursion--; +} #endif void __netif_schedule(struct Qdisc *q); diff --git a/include/net/ip6_tunnel.h b/include/net/ip6_tunnel.h index 120db2865811..1253cbb4b0a4 100644 --- a/include/net/ip6_tunnel.h +++ b/include/net/ip6_tunnel.h @@ -156,6 +156,16 @@ static inline void ip6tunnel_xmit(struct sock *sk, struct sk_buff *skb, { int pkt_len, err; + if (dev_recursion_level() > IP_TUNNEL_RECURSION_LIMIT) { + net_crit_ratelimited("Dead loop on virtual device %s, fix it urgently!\n", + dev->name); + DEV_STATS_INC(dev, tx_errors); + kfree_skb(skb); + return; + } + + dev_xmit_recursion_inc(); + memset(skb->cb, 0, sizeof(struct inet6_skb_parm)); IP6CB(skb)->flags = ip6cb_flags; pkt_len = skb->len - skb_inner_network_offset(skb); @@ -166,6 +176,8 @@ static inline void ip6tunnel_xmit(struct sock *sk, struct sk_buff *skb, pkt_len = -1; iptunnel_xmit_stats(dev, pkt_len); } + + dev_xmit_recursion_dec(); } #endif #endif diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h index 4021e6a73e32..80662f812080 100644 --- a/include/net/ip_tunnels.h +++ b/include/net/ip_tunnels.h @@ -27,6 +27,13 @@ #include #endif +/* Recursion limit for tunnel xmit to detect routing loops. + * Unlike XMIT_RECURSION_LIMIT (8) used in the no-qdisc path, tunnel + * recursion involves route lookups and full IP output, consuming much + * more stack per level, so a lower limit is needed. + */ +#define IP_TUNNEL_RECURSION_LIMIT 4 + /* Keep error state on tunnel for 30 sec */ #define IPTUNNEL_ERR_TIMEO (30*HZ) diff --git a/net/core/dev.h b/net/core/dev.h index 98793a738f43..781619e76b3e 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -366,41 +366,6 @@ static inline void napi_assert_will_not_race(const struct napi_struct *napi) void kick_defer_list_purge(unsigned int cpu); -#define XMIT_RECURSION_LIMIT 8 - -#ifndef CONFIG_PREEMPT_RT -static inline bool dev_xmit_recursion(void) -{ - return unlikely(__this_cpu_read(softnet_data.xmit.recursion) > - XMIT_RECURSION_LIMIT); -} - -static inline void dev_xmit_recursion_inc(void) -{ - __this_cpu_inc(softnet_data.xmit.recursion); -} - -static inline void dev_xmit_recursion_dec(void) -{ - __this_cpu_dec(softnet_data.xmit.recursion); -} -#else -static inline bool dev_xmit_recursion(void) -{ - return unlikely(current->net_xmit.recursion > XMIT_RECURSION_LIMIT); -} - -static inline void dev_xmit_recursion_inc(void) -{ - current->net_xmit.recursion++; -} - -static inline void dev_xmit_recursion_dec(void) -{ - current->net_xmit.recursion--; -} -#endif - int dev_set_hwtstamp_phylib(struct net_device *dev, struct kernel_hwtstamp_config *cfg, struct netlink_ext_ack *extack); diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c index 2e61ac137128..b1b6bf949f65 100644 --- a/net/ipv4/ip_tunnel_core.c +++ b/net/ipv4/ip_tunnel_core.c @@ -58,6 +58,17 @@ void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb, struct iphdr *iph; int err; + if (dev_recursion_level() > IP_TUNNEL_RECURSION_LIMIT) { + net_crit_ratelimited("Dead loop on virtual device %s, fix it urgently!\n", + dev->name); + DEV_STATS_INC(dev, tx_errors); + ip_rt_put(rt); + kfree_skb(skb); + return; + } + + dev_xmit_recursion_inc(); + skb_scrub_packet(skb, xnet); skb_clear_hash_if_not_l4(skb); @@ -88,6 +99,8 @@ void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb, pkt_len = 0; iptunnel_xmit_stats(dev, pkt_len); } + + dev_xmit_recursion_dec(); } EXPORT_SYMBOL_GPL(iptunnel_xmit); -- cgit v1.2.3 From fa655a9ca73f7df32b8ca4d14ce11742f9578288 Mon Sep 17 00:00:00 2001 From: Thorsten Blum Date: Tue, 3 Mar 2026 22:31:01 +0100 Subject: nvme: Annotate struct nvme_dhchap_key with __counted_by Add the __counted_by() compiler attribute to the flexible array member 'key' to improve access bounds-checking via CONFIG_UBSAN_BOUNDS and CONFIG_FORTIFY_SOURCE. Reviewed-by: Christoph Hellwig Signed-off-by: Thorsten Blum Signed-off-by: Keith Busch --- include/linux/nvme-auth.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/nvme-auth.h b/include/linux/nvme-auth.h index 60e069a6757f..e75c29c51464 100644 --- a/include/linux/nvme-auth.h +++ b/include/linux/nvme-auth.h @@ -11,7 +11,7 @@ struct nvme_dhchap_key { size_t len; u8 hash; - u8 key[]; + u8 key[] __counted_by(len); }; u32 nvme_auth_get_seqnum(void); -- cgit v1.2.3 From 22fd7f7fed2ae3702f90d1985c326354e86b9c75 Mon Sep 17 00:00:00 2001 From: Muhammad Amirul Asyraf Mohamad Jamian Date: Thu, 5 Mar 2026 01:31:51 -0800 Subject: firmware: stratix10-svc: Add Multi SVC clients support In the current implementation, SVC client drivers such as socfpga-hwmon, intel_fcs, stratix10-soc, stratix10-rsu each send an SMC command that triggers a single thread in the stratix10-svc driver. Upon receiving a callback, the initiating client driver sends a stratix10-svc-done signal, terminating the thread without waiting for other pending SMC commands to complete. This leads to a timeout issue in the firmware SVC mailbox service when multiple client drivers send SMC commands concurrently. To resolve this issue, a dedicated thread is now created per channel. The stratix10-svc driver will support up to the number of channels defined by SVC_NUM_CHANNEL. Thread synchronization is handled using a mutex to prevent simultaneous issuance of SMC commands by multiple threads. SVC_NUM_DATA_IN_FIFO is reduced from 32 to 8, since each channel now has its own dedicated FIFO and the SDM processes commands one at a time. 8 entries per channel is sufficient while keeping the total aggregate capacity the same (4 channels x 8 = 32 entries). Additionally, a thread task is now validated before invoking kthread_stop when the user aborts, ensuring safe termination. Timeout values have also been adjusted to accommodate the increased load from concurrent client driver activity. Fixes: 7ca5ce896524 ("firmware: add Intel Stratix10 service layer driver") Cc: stable@vger.kernel.org Signed-off-by: Ang Tien Sung Signed-off-by: Fong, Yan Kei Signed-off-by: Muhammad Amirul Asyraf Mohamad Jamian Link: https://lore.kernel.org/all/20260305093151.2678-1-muhammad.amirul.asyraf.mohamad.jamian@altera.com Signed-off-by: Dinh Nguyen --- drivers/firmware/stratix10-svc.c | 228 ++++++++++++--------- .../linux/firmware/intel/stratix10-svc-client.h | 8 +- 2 files changed, 130 insertions(+), 106 deletions(-) (limited to 'include/linux') diff --git a/drivers/firmware/stratix10-svc.c b/drivers/firmware/stratix10-svc.c index 6f5c298582ab..e9e35d67ef96 100644 --- a/drivers/firmware/stratix10-svc.c +++ b/drivers/firmware/stratix10-svc.c @@ -37,15 +37,14 @@ * service layer will return error to FPGA manager when timeout occurs, * timeout is set to 30 seconds (30 * 1000) at Intel Stratix10 SoC. */ -#define SVC_NUM_DATA_IN_FIFO 32 +#define SVC_NUM_DATA_IN_FIFO 8 #define SVC_NUM_CHANNEL 4 -#define FPGA_CONFIG_DATA_CLAIM_TIMEOUT_MS 200 +#define FPGA_CONFIG_DATA_CLAIM_TIMEOUT_MS 2000 #define FPGA_CONFIG_STATUS_TIMEOUT_SEC 30 #define BYTE_TO_WORD_SIZE 4 /* stratix10 service layer clients */ #define STRATIX10_RSU "stratix10-rsu" -#define INTEL_FCS "intel-fcs" /* Maximum number of SDM client IDs. */ #define MAX_SDM_CLIENT_IDS 16 @@ -105,11 +104,9 @@ struct stratix10_svc_chan; /** * struct stratix10_svc - svc private data * @stratix10_svc_rsu: pointer to stratix10 RSU device - * @intel_svc_fcs: pointer to the FCS device */ struct stratix10_svc { struct platform_device *stratix10_svc_rsu; - struct platform_device *intel_svc_fcs; }; /** @@ -251,12 +248,10 @@ struct stratix10_async_ctrl { * @num_active_client: number of active service client * @node: list management * @genpool: memory pool pointing to the memory region - * @task: pointer to the thread task which handles SMC or HVC call - * @svc_fifo: a queue for storing service message data * @complete_status: state for completion - * @svc_fifo_lock: protect access to service message data queue * @invoke_fn: function to issue secure monitor call or hypervisor call * @svc: manages the list of client svc drivers + * @sdm_lock: only allows a single command single response to SDM * @actrl: async control structure * * This struct is used to create communication channels for service clients, to @@ -269,12 +264,10 @@ struct stratix10_svc_controller { int num_active_client; struct list_head node; struct gen_pool *genpool; - struct task_struct *task; - struct kfifo svc_fifo; struct completion complete_status; - spinlock_t svc_fifo_lock; svc_invoke_fn *invoke_fn; struct stratix10_svc *svc; + struct mutex sdm_lock; struct stratix10_async_ctrl actrl; }; @@ -283,6 +276,9 @@ struct stratix10_svc_controller { * @ctrl: pointer to service controller which is the provider of this channel * @scl: pointer to service client which owns the channel * @name: service client name associated with the channel + * @task: pointer to the thread task which handles SMC or HVC call + * @svc_fifo: a queue for storing service message data (separate fifo for every channel) + * @svc_fifo_lock: protect access to service message data queue (locking pending fifo) * @lock: protect access to the channel * @async_chan: reference to asynchronous channel object for this channel * @@ -293,6 +289,9 @@ struct stratix10_svc_chan { struct stratix10_svc_controller *ctrl; struct stratix10_svc_client *scl; char *name; + struct task_struct *task; + struct kfifo svc_fifo; + spinlock_t svc_fifo_lock; spinlock_t lock; struct stratix10_async_chan *async_chan; }; @@ -527,10 +526,10 @@ static void svc_thread_recv_status_ok(struct stratix10_svc_data *p_data, */ static int svc_normal_to_secure_thread(void *data) { - struct stratix10_svc_controller - *ctrl = (struct stratix10_svc_controller *)data; - struct stratix10_svc_data *pdata; - struct stratix10_svc_cb_data *cbdata; + struct stratix10_svc_chan *chan = (struct stratix10_svc_chan *)data; + struct stratix10_svc_controller *ctrl = chan->ctrl; + struct stratix10_svc_data *pdata = NULL; + struct stratix10_svc_cb_data *cbdata = NULL; struct arm_smccc_res res; unsigned long a0, a1, a2, a3, a4, a5, a6, a7; int ret_fifo = 0; @@ -555,12 +554,12 @@ static int svc_normal_to_secure_thread(void *data) a6 = 0; a7 = 0; - pr_debug("smc_hvc_shm_thread is running\n"); + pr_debug("%s: %s: Thread is running!\n", __func__, chan->name); while (!kthread_should_stop()) { - ret_fifo = kfifo_out_spinlocked(&ctrl->svc_fifo, + ret_fifo = kfifo_out_spinlocked(&chan->svc_fifo, pdata, sizeof(*pdata), - &ctrl->svc_fifo_lock); + &chan->svc_fifo_lock); if (!ret_fifo) continue; @@ -569,9 +568,25 @@ static int svc_normal_to_secure_thread(void *data) (unsigned int)pdata->paddr, pdata->command, (unsigned int)pdata->size); + /* SDM can only process one command at a time */ + pr_debug("%s: %s: Thread is waiting for mutex!\n", + __func__, chan->name); + if (mutex_lock_interruptible(&ctrl->sdm_lock)) { + /* item already dequeued; notify client to unblock it */ + cbdata->status = BIT(SVC_STATUS_ERROR); + cbdata->kaddr1 = NULL; + cbdata->kaddr2 = NULL; + cbdata->kaddr3 = NULL; + if (pdata->chan->scl) + pdata->chan->scl->receive_cb(pdata->chan->scl, + cbdata); + break; + } + switch (pdata->command) { case COMMAND_RECONFIG_DATA_CLAIM: svc_thread_cmd_data_claim(ctrl, pdata, cbdata); + mutex_unlock(&ctrl->sdm_lock); continue; case COMMAND_RECONFIG: a0 = INTEL_SIP_SMC_FPGA_CONFIG_START; @@ -700,10 +715,11 @@ static int svc_normal_to_secure_thread(void *data) break; default: pr_warn("it shouldn't happen\n"); - break; + mutex_unlock(&ctrl->sdm_lock); + continue; } - pr_debug("%s: before SMC call -- a0=0x%016x a1=0x%016x", - __func__, + pr_debug("%s: %s: before SMC call -- a0=0x%016x a1=0x%016x", + __func__, chan->name, (unsigned int)a0, (unsigned int)a1); pr_debug(" a2=0x%016x\n", (unsigned int)a2); @@ -712,8 +728,8 @@ static int svc_normal_to_secure_thread(void *data) pr_debug(" a5=0x%016x\n", (unsigned int)a5); ctrl->invoke_fn(a0, a1, a2, a3, a4, a5, a6, a7, &res); - pr_debug("%s: after SMC call -- res.a0=0x%016x", - __func__, (unsigned int)res.a0); + pr_debug("%s: %s: after SMC call -- res.a0=0x%016x", + __func__, chan->name, (unsigned int)res.a0); pr_debug(" res.a1=0x%016x, res.a2=0x%016x", (unsigned int)res.a1, (unsigned int)res.a2); pr_debug(" res.a3=0x%016x\n", (unsigned int)res.a3); @@ -728,6 +744,7 @@ static int svc_normal_to_secure_thread(void *data) cbdata->kaddr2 = NULL; cbdata->kaddr3 = NULL; pdata->chan->scl->receive_cb(pdata->chan->scl, cbdata); + mutex_unlock(&ctrl->sdm_lock); continue; } @@ -801,6 +818,8 @@ static int svc_normal_to_secure_thread(void *data) break; } + + mutex_unlock(&ctrl->sdm_lock); } kfree(cbdata); @@ -1696,22 +1715,33 @@ int stratix10_svc_send(struct stratix10_svc_chan *chan, void *msg) if (!p_data) return -ENOMEM; - /* first client will create kernel thread */ - if (!chan->ctrl->task) { - chan->ctrl->task = - kthread_run_on_cpu(svc_normal_to_secure_thread, - (void *)chan->ctrl, - cpu, "svc_smc_hvc_thread"); - if (IS_ERR(chan->ctrl->task)) { + /* first caller creates the per-channel kthread */ + if (!chan->task) { + struct task_struct *task; + + task = kthread_run_on_cpu(svc_normal_to_secure_thread, + (void *)chan, + cpu, "svc_smc_hvc_thread"); + if (IS_ERR(task)) { dev_err(chan->ctrl->dev, "failed to create svc_smc_hvc_thread\n"); kfree(p_data); return -EINVAL; } + + spin_lock(&chan->lock); + if (chan->task) { + /* another caller won the race; discard our thread */ + spin_unlock(&chan->lock); + kthread_stop(task); + } else { + chan->task = task; + spin_unlock(&chan->lock); + } } - pr_debug("%s: sent P-va=%p, P-com=%x, P-size=%u\n", __func__, - p_msg->payload, p_msg->command, + pr_debug("%s: %s: sent P-va=%p, P-com=%x, P-size=%u\n", __func__, + chan->name, p_msg->payload, p_msg->command, (unsigned int)p_msg->payload_length); if (list_empty(&svc_data_mem)) { @@ -1747,12 +1777,16 @@ int stratix10_svc_send(struct stratix10_svc_chan *chan, void *msg) p_data->arg[2] = p_msg->arg[2]; p_data->size = p_msg->payload_length; p_data->chan = chan; - pr_debug("%s: put to FIFO pa=0x%016x, cmd=%x, size=%u\n", __func__, - (unsigned int)p_data->paddr, p_data->command, - (unsigned int)p_data->size); - ret = kfifo_in_spinlocked(&chan->ctrl->svc_fifo, p_data, + pr_debug("%s: %s: put to FIFO pa=0x%016x, cmd=%x, size=%u\n", + __func__, + chan->name, + (unsigned int)p_data->paddr, + p_data->command, + (unsigned int)p_data->size); + + ret = kfifo_in_spinlocked(&chan->svc_fifo, p_data, sizeof(*p_data), - &chan->ctrl->svc_fifo_lock); + &chan->svc_fifo_lock); kfree(p_data); @@ -1773,11 +1807,12 @@ EXPORT_SYMBOL_GPL(stratix10_svc_send); */ void stratix10_svc_done(struct stratix10_svc_chan *chan) { - /* stop thread when thread is running AND only one active client */ - if (chan->ctrl->task && chan->ctrl->num_active_client <= 1) { - pr_debug("svc_smc_hvc_shm_thread is stopped\n"); - kthread_stop(chan->ctrl->task); - chan->ctrl->task = NULL; + /* stop thread when thread is running */ + if (chan->task) { + pr_debug("%s: %s: svc_smc_hvc_shm_thread is stopping\n", + __func__, chan->name); + kthread_stop(chan->task); + chan->task = NULL; } } EXPORT_SYMBOL_GPL(stratix10_svc_done); @@ -1817,8 +1852,8 @@ void *stratix10_svc_allocate_memory(struct stratix10_svc_chan *chan, pmem->paddr = pa; pmem->size = s; list_add_tail(&pmem->node, &svc_data_mem); - pr_debug("%s: va=%p, pa=0x%016x\n", __func__, - pmem->vaddr, (unsigned int)pmem->paddr); + pr_debug("%s: %s: va=%p, pa=0x%016x\n", __func__, + chan->name, pmem->vaddr, (unsigned int)pmem->paddr); return (void *)va; } @@ -1855,6 +1890,13 @@ static const struct of_device_id stratix10_svc_drv_match[] = { {}, }; +static const char * const chan_names[SVC_NUM_CHANNEL] = { + SVC_CLIENT_FPGA, + SVC_CLIENT_RSU, + SVC_CLIENT_FCS, + SVC_CLIENT_HWMON +}; + static int stratix10_svc_drv_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; @@ -1862,11 +1904,11 @@ static int stratix10_svc_drv_probe(struct platform_device *pdev) struct stratix10_svc_chan *chans; struct gen_pool *genpool; struct stratix10_svc_sh_memory *sh_memory; - struct stratix10_svc *svc; + struct stratix10_svc *svc = NULL; svc_invoke_fn *invoke_fn; size_t fifo_size; - int ret; + int ret, i = 0; /* get SMC or HVC function */ invoke_fn = get_invoke_func(dev); @@ -1905,8 +1947,8 @@ static int stratix10_svc_drv_probe(struct platform_device *pdev) controller->num_active_client = 0; controller->chans = chans; controller->genpool = genpool; - controller->task = NULL; controller->invoke_fn = invoke_fn; + INIT_LIST_HEAD(&controller->node); init_completion(&controller->complete_status); ret = stratix10_svc_async_init(controller); @@ -1917,32 +1959,20 @@ static int stratix10_svc_drv_probe(struct platform_device *pdev) } fifo_size = sizeof(struct stratix10_svc_data) * SVC_NUM_DATA_IN_FIFO; - ret = kfifo_alloc(&controller->svc_fifo, fifo_size, GFP_KERNEL); - if (ret) { - dev_err(dev, "failed to allocate FIFO\n"); - goto err_async_exit; - } - spin_lock_init(&controller->svc_fifo_lock); - - chans[0].scl = NULL; - chans[0].ctrl = controller; - chans[0].name = SVC_CLIENT_FPGA; - spin_lock_init(&chans[0].lock); + mutex_init(&controller->sdm_lock); - chans[1].scl = NULL; - chans[1].ctrl = controller; - chans[1].name = SVC_CLIENT_RSU; - spin_lock_init(&chans[1].lock); - - chans[2].scl = NULL; - chans[2].ctrl = controller; - chans[2].name = SVC_CLIENT_FCS; - spin_lock_init(&chans[2].lock); - - chans[3].scl = NULL; - chans[3].ctrl = controller; - chans[3].name = SVC_CLIENT_HWMON; - spin_lock_init(&chans[3].lock); + for (i = 0; i < SVC_NUM_CHANNEL; i++) { + chans[i].scl = NULL; + chans[i].ctrl = controller; + chans[i].name = (char *)chan_names[i]; + spin_lock_init(&chans[i].lock); + ret = kfifo_alloc(&chans[i].svc_fifo, fifo_size, GFP_KERNEL); + if (ret) { + dev_err(dev, "failed to allocate FIFO %d\n", i); + goto err_free_fifos; + } + spin_lock_init(&chans[i].svc_fifo_lock); + } list_add_tail(&controller->node, &svc_ctrl); platform_set_drvdata(pdev, controller); @@ -1951,7 +1981,7 @@ static int stratix10_svc_drv_probe(struct platform_device *pdev) svc = devm_kzalloc(dev, sizeof(*svc), GFP_KERNEL); if (!svc) { ret = -ENOMEM; - goto err_free_kfifo; + goto err_free_fifos; } controller->svc = svc; @@ -1959,51 +1989,43 @@ static int stratix10_svc_drv_probe(struct platform_device *pdev) if (!svc->stratix10_svc_rsu) { dev_err(dev, "failed to allocate %s device\n", STRATIX10_RSU); ret = -ENOMEM; - goto err_free_kfifo; + goto err_free_fifos; } ret = platform_device_add(svc->stratix10_svc_rsu); - if (ret) { - platform_device_put(svc->stratix10_svc_rsu); - goto err_free_kfifo; - } - - svc->intel_svc_fcs = platform_device_alloc(INTEL_FCS, 1); - if (!svc->intel_svc_fcs) { - dev_err(dev, "failed to allocate %s device\n", INTEL_FCS); - ret = -ENOMEM; - goto err_unregister_rsu_dev; - } - - ret = platform_device_add(svc->intel_svc_fcs); - if (ret) { - platform_device_put(svc->intel_svc_fcs); - goto err_unregister_rsu_dev; - } + if (ret) + goto err_put_device; ret = of_platform_default_populate(dev_of_node(dev), NULL, dev); if (ret) - goto err_unregister_fcs_dev; + goto err_unregister_rsu_dev; pr_info("Intel Service Layer Driver Initialized\n"); return 0; -err_unregister_fcs_dev: - platform_device_unregister(svc->intel_svc_fcs); err_unregister_rsu_dev: platform_device_unregister(svc->stratix10_svc_rsu); -err_free_kfifo: - kfifo_free(&controller->svc_fifo); -err_async_exit: + goto err_free_fifos; +err_put_device: + platform_device_put(svc->stratix10_svc_rsu); +err_free_fifos: + /* only remove from list if list_add_tail() was reached */ + if (!list_empty(&controller->node)) + list_del(&controller->node); + /* free only the FIFOs that were successfully allocated */ + while (i--) + kfifo_free(&chans[i].svc_fifo); stratix10_svc_async_exit(controller); err_destroy_pool: gen_pool_destroy(genpool); + return ret; } static void stratix10_svc_drv_remove(struct platform_device *pdev) { + int i; struct stratix10_svc_controller *ctrl = platform_get_drvdata(pdev); struct stratix10_svc *svc = ctrl->svc; @@ -2011,14 +2033,16 @@ static void stratix10_svc_drv_remove(struct platform_device *pdev) of_platform_depopulate(ctrl->dev); - platform_device_unregister(svc->intel_svc_fcs); platform_device_unregister(svc->stratix10_svc_rsu); - kfifo_free(&ctrl->svc_fifo); - if (ctrl->task) { - kthread_stop(ctrl->task); - ctrl->task = NULL; + for (i = 0; i < SVC_NUM_CHANNEL; i++) { + if (ctrl->chans[i].task) { + kthread_stop(ctrl->chans[i].task); + ctrl->chans[i].task = NULL; + } + kfifo_free(&ctrl->chans[i].svc_fifo); } + if (ctrl->genpool) gen_pool_destroy(ctrl->genpool); list_del(&ctrl->node); diff --git a/include/linux/firmware/intel/stratix10-svc-client.h b/include/linux/firmware/intel/stratix10-svc-client.h index d290060f4c73..91013161e9db 100644 --- a/include/linux/firmware/intel/stratix10-svc-client.h +++ b/include/linux/firmware/intel/stratix10-svc-client.h @@ -68,12 +68,12 @@ * timeout value used in Stratix10 FPGA manager driver. * timeout value used in RSU driver */ -#define SVC_RECONFIG_REQUEST_TIMEOUT_MS 300 -#define SVC_RECONFIG_BUFFER_TIMEOUT_MS 720 -#define SVC_RSU_REQUEST_TIMEOUT_MS 300 +#define SVC_RECONFIG_REQUEST_TIMEOUT_MS 5000 +#define SVC_RECONFIG_BUFFER_TIMEOUT_MS 5000 +#define SVC_RSU_REQUEST_TIMEOUT_MS 2000 #define SVC_FCS_REQUEST_TIMEOUT_MS 2000 #define SVC_COMPLETED_TIMEOUT_MS 30000 -#define SVC_HWMON_REQUEST_TIMEOUT_MS 300 +#define SVC_HWMON_REQUEST_TIMEOUT_MS 2000 struct stratix10_svc_chan; -- cgit v1.2.3 From b2e48c429ec54715d16fefa719dd2fbded2e65be Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 10 Mar 2026 21:28:53 +0100 Subject: sched/mmcid: Prevent CID stalls due to concurrent forks A newly forked task is accounted as MMCID user before the task is visible in the process' thread list and the global task list. This creates the following problem: CPU1 CPU2 fork() sched_mm_cid_fork(tnew1) tnew1->mm.mm_cid_users++; tnew1->mm_cid.cid = getcid() -> preemption fork() sched_mm_cid_fork(tnew2) tnew2->mm.mm_cid_users++; // Reaches the per CPU threshold mm_cid_fixup_tasks_to_cpus() for_each_other(current, p) .... As tnew1 is not visible yet, this fails to fix up the already allocated CID of tnew1. As a consequence a subsequent schedule in might fail to acquire a (transitional) CID and the machine stalls. Move the invocation of sched_mm_cid_fork() after the new task becomes visible in the thread and the task list to prevent this. This also makes it symmetrical vs. exit() where the task is removed as CID user before the task is removed from the thread and task lists. Fixes: fbd0e71dc370 ("sched/mmcid: Provide CID ownership mode fixup functions") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Tested-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20260310202525.969061974@kernel.org --- include/linux/sched.h | 2 -- kernel/fork.c | 2 -- kernel/sched/core.c | 22 +++++++++++++++------- 3 files changed, 15 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index a7b4a980eb2f..5a5d3dbc9cdf 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2354,7 +2354,6 @@ static __always_inline void alloc_tag_restore(struct alloc_tag *tag, struct allo #ifdef CONFIG_SCHED_MM_CID void sched_mm_cid_before_execve(struct task_struct *t); void sched_mm_cid_after_execve(struct task_struct *t); -void sched_mm_cid_fork(struct task_struct *t); void sched_mm_cid_exit(struct task_struct *t); static __always_inline int task_mm_cid(struct task_struct *t) { @@ -2363,7 +2362,6 @@ static __always_inline int task_mm_cid(struct task_struct *t) #else static inline void sched_mm_cid_before_execve(struct task_struct *t) { } static inline void sched_mm_cid_after_execve(struct task_struct *t) { } -static inline void sched_mm_cid_fork(struct task_struct *t) { } static inline void sched_mm_cid_exit(struct task_struct *t) { } static __always_inline int task_mm_cid(struct task_struct *t) { diff --git a/kernel/fork.c b/kernel/fork.c index 65113a304518..7febf4c2889e 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1586,7 +1586,6 @@ static int copy_mm(u64 clone_flags, struct task_struct *tsk) tsk->mm = mm; tsk->active_mm = mm; - sched_mm_cid_fork(tsk); return 0; } @@ -2498,7 +2497,6 @@ bad_fork_cleanup_namespaces: exit_nsproxy_namespaces(p); bad_fork_cleanup_mm: if (p->mm) { - sched_mm_cid_exit(p); mm_clear_owner(p->mm, p); mmput(p->mm); } diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b7f77c165a6e..d25427855b5d 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4729,8 +4729,11 @@ void sched_cancel_fork(struct task_struct *p) scx_cancel_fork(p); } +static void sched_mm_cid_fork(struct task_struct *t); + void sched_post_fork(struct task_struct *p) { + sched_mm_cid_fork(p); uclamp_post_fork(p); scx_post_fork(p); } @@ -10646,12 +10649,13 @@ static void mm_cid_do_fixup_tasks_to_cpus(struct mm_struct *mm) * possible switch back to per task mode happens either in the * deferred handler function or in the next fork()/exit(). * - * The caller has already transferred. The newly incoming task is - * already accounted for, but not yet visible. + * The caller has already transferred so remove it from the users + * count. The incoming task is already visible and has mm_cid.active, + * but has task::mm_cid::cid == UNSET. Still it needs to be accounted + * for. Concurrent fork()s might add more threads, but all of them have + * task::mm_cid::active = 0, so they don't affect the accounting here. */ - users = mm->mm_cid.users - 2; - if (!users) - return; + users = mm->mm_cid.users - 1; guard(rcu)(); for_other_threads(current, t) { @@ -10688,12 +10692,15 @@ static bool sched_mm_cid_add_user(struct task_struct *t, struct mm_struct *mm) return mm_update_max_cids(mm); } -void sched_mm_cid_fork(struct task_struct *t) +static void sched_mm_cid_fork(struct task_struct *t) { struct mm_struct *mm = t->mm; bool percpu; - WARN_ON_ONCE(!mm || t->mm_cid.cid != MM_CID_UNSET); + if (!mm) + return; + + WARN_ON_ONCE(t->mm_cid.cid != MM_CID_UNSET); guard(mutex)(&mm->mm_cid.mutex); scoped_guard(raw_spinlock_irq, &mm->mm_cid.lock) { @@ -10885,6 +10892,7 @@ void mm_init_cid(struct mm_struct *mm, struct task_struct *p) } #else /* CONFIG_SCHED_MM_CID */ static inline void mm_update_cpus_allowed(struct mm_struct *mm, const struct cpumask *affmsk) { } +static inline void sched_mm_cid_fork(struct task_struct *t) { } #endif /* !CONFIG_SCHED_MM_CID */ static DEFINE_PER_CPU(struct sched_change_ctx, sched_change_ctx); -- cgit v1.2.3 From 192d852129b1b7c4f0ddbab95d0de1efd5ee1405 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 10 Mar 2026 21:29:09 +0100 Subject: sched/mmcid: Avoid full tasklist walks Chasing vfork()'ed tasks on a CID ownership mode switch requires a full task list walk, which is obviously expensive on large systems. Avoid that by keeping a list of tasks using a mm MMCID entity in mm::mm_cid and walk this list instead. This removes the proven to be flaky counting logic and avoids a full task list walk in the case of vfork()'ed tasks. Fixes: fbd0e71dc370 ("sched/mmcid: Provide CID ownership mode fixup functions") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Tested-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20260310202526.183824481@kernel.org --- include/linux/rseq_types.h | 6 +++++- kernel/fork.c | 1 + kernel/sched/core.c | 54 +++++++++++----------------------------------- 3 files changed, 18 insertions(+), 43 deletions(-) (limited to 'include/linux') diff --git a/include/linux/rseq_types.h b/include/linux/rseq_types.h index da5fa6f40294..0b42045988db 100644 --- a/include/linux/rseq_types.h +++ b/include/linux/rseq_types.h @@ -133,10 +133,12 @@ struct rseq_data { }; * @active: MM CID is active for the task * @cid: The CID associated to the task either permanently or * borrowed from the CPU + * @node: Queued in the per MM MMCID list */ struct sched_mm_cid { unsigned int active; unsigned int cid; + struct hlist_node node; }; /** @@ -157,6 +159,7 @@ struct mm_cid_pcpu { * @work: Regular work to handle the affinity mode change case * @lock: Spinlock to protect against affinity setting which can't take @mutex * @mutex: Mutex to serialize forks and exits related to this mm + * @user_list: List of the MM CID users of a MM * @nr_cpus_allowed: The number of CPUs in the per MM allowed CPUs map. The map * is growth only. * @users: The number of tasks sharing this MM. Separate from mm::mm_users @@ -177,13 +180,14 @@ struct mm_mm_cid { raw_spinlock_t lock; struct mutex mutex; + struct hlist_head user_list; /* Low frequency modified */ unsigned int nr_cpus_allowed; unsigned int users; unsigned int pcpu_thrs; unsigned int update_deferred; -}____cacheline_aligned_in_smp; +} ____cacheline_aligned; #else /* CONFIG_SCHED_MM_CID */ struct mm_mm_cid { }; struct sched_mm_cid { }; diff --git a/kernel/fork.c b/kernel/fork.c index 7febf4c2889e..bc2bf58b93b6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1000,6 +1000,7 @@ static struct task_struct *dup_task_struct(struct task_struct *orig, int node) #ifdef CONFIG_SCHED_MM_CID tsk->mm_cid.cid = MM_CID_UNSET; tsk->mm_cid.active = 0; + INIT_HLIST_NODE(&tsk->mm_cid.node); #endif return tsk; diff --git a/kernel/sched/core.c b/kernel/sched/core.c index f56156f91d08..496dff740dca 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -10620,13 +10620,10 @@ static inline void mm_cid_transit_to_cpu(struct task_struct *t, struct mm_cid_pc } } -static bool mm_cid_fixup_task_to_cpu(struct task_struct *t, struct mm_struct *mm) +static void mm_cid_fixup_task_to_cpu(struct task_struct *t, struct mm_struct *mm) { /* Remote access to mm::mm_cid::pcpu requires rq_lock */ guard(task_rq_lock)(t); - /* If the task is not active it is not in the users count */ - if (!t->mm_cid.active) - return false; if (cid_on_task(t->mm_cid.cid)) { /* If running on the CPU, put the CID in transit mode, otherwise drop it */ if (task_rq(t)->curr == t) @@ -10634,51 +10631,21 @@ static bool mm_cid_fixup_task_to_cpu(struct task_struct *t, struct mm_struct *mm else mm_unset_cid_on_task(t); } - return true; } -static void mm_cid_do_fixup_tasks_to_cpus(struct mm_struct *mm) +static void mm_cid_fixup_tasks_to_cpus(void) { - struct task_struct *p, *t; - unsigned int users; - - /* - * This can obviously race with a concurrent affinity change, which - * increases the number of allowed CPUs for this mm, but that does - * not affect the mode and only changes the CID constraints. A - * possible switch back to per task mode happens either in the - * deferred handler function or in the next fork()/exit(). - * - * The caller has already transferred so remove it from the users - * count. The incoming task is already visible and has mm_cid.active, - * but has task::mm_cid::cid == UNSET. Still it needs to be accounted - * for. Concurrent fork()s might add more threads, but all of them have - * task::mm_cid::active = 0, so they don't affect the accounting here. - */ - users = mm->mm_cid.users - 1; - - guard(rcu)(); - for_other_threads(current, t) { - if (mm_cid_fixup_task_to_cpu(t, mm)) - users--; - } + struct mm_struct *mm = current->mm; + struct task_struct *t; - if (!users) - return; + lockdep_assert_held(&mm->mm_cid.mutex); - /* Happens only for VM_CLONE processes. */ - for_each_process_thread(p, t) { - if (t == current || t->mm != mm) - continue; - mm_cid_fixup_task_to_cpu(t, mm); + hlist_for_each_entry(t, &mm->mm_cid.user_list, mm_cid.node) { + /* Current has already transferred before invoking the fixup. */ + if (t != current) + mm_cid_fixup_task_to_cpu(t, mm); } -} - -static void mm_cid_fixup_tasks_to_cpus(void) -{ - struct mm_struct *mm = current->mm; - mm_cid_do_fixup_tasks_to_cpus(mm); mm_cid_complete_transit(mm, MM_CID_ONCPU); } @@ -10687,6 +10654,7 @@ static bool sched_mm_cid_add_user(struct task_struct *t, struct mm_struct *mm) lockdep_assert_held(&mm->mm_cid.lock); t->mm_cid.active = 1; + hlist_add_head(&t->mm_cid.node, &mm->mm_cid.user_list); mm->mm_cid.users++; return mm_update_max_cids(mm); } @@ -10744,6 +10712,7 @@ static bool sched_mm_cid_remove_user(struct task_struct *t) /* Clear the transition bit */ t->mm_cid.cid = cid_from_transit_cid(t->mm_cid.cid); mm_unset_cid_on_task(t); + hlist_del_init(&t->mm_cid.node); t->mm->mm_cid.users--; return mm_update_max_cids(t->mm); } @@ -10886,6 +10855,7 @@ void mm_init_cid(struct mm_struct *mm, struct task_struct *p) mutex_init(&mm->mm_cid.mutex); mm->mm_cid.irq_work = IRQ_WORK_INIT_HARD(mm_cid_irq_work); INIT_WORK(&mm->mm_cid.work, mm_cid_work_fn); + INIT_HLIST_HEAD(&mm->mm_cid.user_list); cpumask_copy(mm_cpus_allowed(mm), &p->cpus_mask); bitmap_zero(mm_cidmask(mm), num_possible_cpus()); } -- cgit v1.2.3 From 416909962e7cdf29fd01ac523c953f37708df93d Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Tue, 17 Feb 2026 22:07:47 -0500 Subject: USB: usbcore: Introduce usb_bulk_msg_killable() The synchronous message API in usbcore (usb_control_msg(), usb_bulk_msg(), and so on) uses uninterruptible waits. However, drivers may call these routines in the context of a user thread, which means it ought to be possible to at least kill them. For this reason, introduce a new usb_bulk_msg_killable() function which behaves the same as usb_bulk_msg() except for using wait_for_completion_killable_timeout() instead of wait_for_completion_timeout(). The same can be done later for usb_control_msg() later on, if it turns out to be needed. Signed-off-by: Alan Stern Suggested-by: Oliver Neukum Link: https://lore.kernel.org/linux-usb/3acfe838-6334-4f6d-be7c-4bb01704b33d@rowland.harvard.edu/ Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") CC: stable@vger.kernel.org Link: https://patch.msgid.link/248628b4-cc83-4e81-a620-3ce4e0376d41@rowland.harvard.edu Signed-off-by: Greg Kroah-Hartman --- drivers/usb/core/message.c | 79 ++++++++++++++++++++++++++++++++++++++++------ include/linux/usb.h | 5 +-- 2 files changed, 72 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c index ea970ddf8879..d97ec7e8c280 100644 --- a/drivers/usb/core/message.c +++ b/drivers/usb/core/message.c @@ -42,16 +42,17 @@ static void usb_api_blocking_completion(struct urb *urb) /* - * Starts urb and waits for completion or timeout. Note that this call - * is NOT interruptible. Many device driver i/o requests should be - * interruptible and therefore these drivers should implement their - * own interruptible routines. + * Starts urb and waits for completion or timeout. + * Whether or not the wait is killable depends on the flag passed in. + * For example, compare usb_bulk_msg() and usb_bulk_msg_killable(). */ -static int usb_start_wait_urb(struct urb *urb, int timeout, int *actual_length) +static int usb_start_wait_urb(struct urb *urb, int timeout, int *actual_length, + bool killable) { struct api_context ctx; unsigned long expire; int retval; + long rc; init_completion(&ctx.done); urb->context = &ctx; @@ -61,12 +62,21 @@ static int usb_start_wait_urb(struct urb *urb, int timeout, int *actual_length) goto out; expire = timeout ? msecs_to_jiffies(timeout) : MAX_SCHEDULE_TIMEOUT; - if (!wait_for_completion_timeout(&ctx.done, expire)) { + if (killable) + rc = wait_for_completion_killable_timeout(&ctx.done, expire); + else + rc = wait_for_completion_timeout(&ctx.done, expire); + if (rc <= 0) { usb_kill_urb(urb); - retval = (ctx.status == -ENOENT ? -ETIMEDOUT : ctx.status); + if (ctx.status != -ENOENT) + retval = ctx.status; + else if (rc == 0) + retval = -ETIMEDOUT; + else + retval = rc; dev_dbg(&urb->dev->dev, - "%s timed out on ep%d%s len=%u/%u\n", + "%s timed out or killed on ep%d%s len=%u/%u\n", current->comm, usb_endpoint_num(&urb->ep->desc), usb_urb_dir_in(urb) ? "in" : "out", @@ -100,7 +110,7 @@ static int usb_internal_control_msg(struct usb_device *usb_dev, usb_fill_control_urb(urb, usb_dev, pipe, (unsigned char *)cmd, data, len, usb_api_blocking_completion, NULL); - retv = usb_start_wait_urb(urb, timeout, &length); + retv = usb_start_wait_urb(urb, timeout, &length, false); if (retv < 0) return retv; else @@ -385,10 +395,59 @@ int usb_bulk_msg(struct usb_device *usb_dev, unsigned int pipe, usb_fill_bulk_urb(urb, usb_dev, pipe, data, len, usb_api_blocking_completion, NULL); - return usb_start_wait_urb(urb, timeout, actual_length); + return usb_start_wait_urb(urb, timeout, actual_length, false); } EXPORT_SYMBOL_GPL(usb_bulk_msg); +/** + * usb_bulk_msg_killable - Builds a bulk urb, sends it off and waits for completion in a killable state + * @usb_dev: pointer to the usb device to send the message to + * @pipe: endpoint "pipe" to send the message to + * @data: pointer to the data to send + * @len: length in bytes of the data to send + * @actual_length: pointer to a location to put the actual length transferred + * in bytes + * @timeout: time in msecs to wait for the message to complete before + * timing out (if 0 the wait is forever) + * + * Context: task context, might sleep. + * + * This function is just like usb_blk_msg() except that it waits in a + * killable state. + * + * Return: + * If successful, 0. Otherwise a negative error number. The number of actual + * bytes transferred will be stored in the @actual_length parameter. + * + */ +int usb_bulk_msg_killable(struct usb_device *usb_dev, unsigned int pipe, + void *data, int len, int *actual_length, int timeout) +{ + struct urb *urb; + struct usb_host_endpoint *ep; + + ep = usb_pipe_endpoint(usb_dev, pipe); + if (!ep || len < 0) + return -EINVAL; + + urb = usb_alloc_urb(0, GFP_KERNEL); + if (!urb) + return -ENOMEM; + + if ((ep->desc.bmAttributes & USB_ENDPOINT_XFERTYPE_MASK) == + USB_ENDPOINT_XFER_INT) { + pipe = (pipe & ~(3 << 30)) | (PIPE_INTERRUPT << 30); + usb_fill_int_urb(urb, usb_dev, pipe, data, len, + usb_api_blocking_completion, NULL, + ep->desc.bInterval); + } else + usb_fill_bulk_urb(urb, usb_dev, pipe, data, len, + usb_api_blocking_completion, NULL); + + return usb_start_wait_urb(urb, timeout, actual_length, true); +} +EXPORT_SYMBOL_GPL(usb_bulk_msg_killable); + /*-------------------------------------------------------------------*/ static void sg_clean(struct usb_sg_request *io) diff --git a/include/linux/usb.h b/include/linux/usb.h index fbfcc70b07fb..57ceeb02a7cb 100644 --- a/include/linux/usb.h +++ b/include/linux/usb.h @@ -1868,8 +1868,9 @@ extern int usb_control_msg(struct usb_device *dev, unsigned int pipe, extern int usb_interrupt_msg(struct usb_device *usb_dev, unsigned int pipe, void *data, int len, int *actual_length, int timeout); extern int usb_bulk_msg(struct usb_device *usb_dev, unsigned int pipe, - void *data, int len, int *actual_length, - int timeout); + void *data, int len, int *actual_length, int timeout); +extern int usb_bulk_msg_killable(struct usb_device *usb_dev, unsigned int pipe, + void *data, int len, int *actual_length, int timeout); /* wrappers around usb_control_msg() for the most common standard requests */ int usb_control_msg_send(struct usb_device *dev, __u8 endpoint, __u8 request, -- cgit v1.2.3 From 1015c27a5e1a63efae2b18a9901494474b4d1dc3 Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Tue, 17 Feb 2026 22:10:32 -0500 Subject: USB: core: Limit the length of unkillable synchronous timeouts The usb_control_msg(), usb_bulk_msg(), and usb_interrupt_msg() APIs in usbcore allow unlimited timeout durations. And since they use uninterruptible waits, this leaves open the possibility of hanging a task for an indefinitely long time, with no way to kill it short of unplugging the target device. To prevent this sort of problem, enforce a maximum limit on the length of these unkillable timeouts. The limit chosen here, somewhat arbitrarily, is 60 seconds. On many systems (although not all) this is short enough to avoid triggering the kernel's hung-task detector. In addition, clear up the ambiguity of negative timeout values by treating them the same as 0, i.e., using the maximum allowed timeout. Signed-off-by: Alan Stern Link: https://lore.kernel.org/linux-usb/3acfe838-6334-4f6d-be7c-4bb01704b33d@rowland.harvard.edu/ Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") CC: stable@vger.kernel.org Link: https://patch.msgid.link/15fc9773-a007-47b0-a703-df89a8cf83dd@rowland.harvard.edu Signed-off-by: Greg Kroah-Hartman --- drivers/usb/core/message.c | 27 +++++++++++++-------------- include/linux/usb.h | 3 +++ 2 files changed, 16 insertions(+), 14 deletions(-) (limited to 'include/linux') diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c index d97ec7e8c280..2ab120ce2fa8 100644 --- a/drivers/usb/core/message.c +++ b/drivers/usb/core/message.c @@ -45,6 +45,8 @@ static void usb_api_blocking_completion(struct urb *urb) * Starts urb and waits for completion or timeout. * Whether or not the wait is killable depends on the flag passed in. * For example, compare usb_bulk_msg() and usb_bulk_msg_killable(). + * + * For non-killable waits, we enforce a maximum limit on the timeout value. */ static int usb_start_wait_urb(struct urb *urb, int timeout, int *actual_length, bool killable) @@ -61,7 +63,9 @@ static int usb_start_wait_urb(struct urb *urb, int timeout, int *actual_length, if (unlikely(retval)) goto out; - expire = timeout ? msecs_to_jiffies(timeout) : MAX_SCHEDULE_TIMEOUT; + if (!killable && (timeout <= 0 || timeout > USB_MAX_SYNCHRONOUS_TIMEOUT)) + timeout = USB_MAX_SYNCHRONOUS_TIMEOUT; + expire = (timeout > 0) ? msecs_to_jiffies(timeout) : MAX_SCHEDULE_TIMEOUT; if (killable) rc = wait_for_completion_killable_timeout(&ctx.done, expire); else @@ -127,8 +131,7 @@ static int usb_internal_control_msg(struct usb_device *usb_dev, * @index: USB message index value * @data: pointer to the data to send * @size: length in bytes of the data to send - * @timeout: time in msecs to wait for the message to complete before timing - * out (if 0 the wait is forever) + * @timeout: time in msecs to wait for the message to complete before timing out * * Context: task context, might sleep. * @@ -183,8 +186,7 @@ EXPORT_SYMBOL_GPL(usb_control_msg); * @index: USB message index value * @driver_data: pointer to the data to send * @size: length in bytes of the data to send - * @timeout: time in msecs to wait for the message to complete before timing - * out (if 0 the wait is forever) + * @timeout: time in msecs to wait for the message to complete before timing out * @memflags: the flags for memory allocation for buffers * * Context: !in_interrupt () @@ -242,8 +244,7 @@ EXPORT_SYMBOL_GPL(usb_control_msg_send); * @index: USB message index value * @driver_data: pointer to the data to be filled in by the message * @size: length in bytes of the data to be received - * @timeout: time in msecs to wait for the message to complete before timing - * out (if 0 the wait is forever) + * @timeout: time in msecs to wait for the message to complete before timing out * @memflags: the flags for memory allocation for buffers * * Context: !in_interrupt () @@ -314,8 +315,7 @@ EXPORT_SYMBOL_GPL(usb_control_msg_recv); * @len: length in bytes of the data to send * @actual_length: pointer to a location to put the actual length transferred * in bytes - * @timeout: time in msecs to wait for the message to complete before - * timing out (if 0 the wait is forever) + * @timeout: time in msecs to wait for the message to complete before timing out * * Context: task context, might sleep. * @@ -347,8 +347,7 @@ EXPORT_SYMBOL_GPL(usb_interrupt_msg); * @len: length in bytes of the data to send * @actual_length: pointer to a location to put the actual length transferred * in bytes - * @timeout: time in msecs to wait for the message to complete before - * timing out (if 0 the wait is forever) + * @timeout: time in msecs to wait for the message to complete before timing out * * Context: task context, might sleep. * @@ -408,12 +407,12 @@ EXPORT_SYMBOL_GPL(usb_bulk_msg); * @actual_length: pointer to a location to put the actual length transferred * in bytes * @timeout: time in msecs to wait for the message to complete before - * timing out (if 0 the wait is forever) + * timing out (if <= 0, the wait is as long as possible) * * Context: task context, might sleep. * - * This function is just like usb_blk_msg() except that it waits in a - * killable state. + * This function is just like usb_blk_msg(), except that it waits in a + * killable state and there is no limit on the timeout length. * * Return: * If successful, 0. Otherwise a negative error number. The number of actual diff --git a/include/linux/usb.h b/include/linux/usb.h index 57ceeb02a7cb..04277af4bb9d 100644 --- a/include/linux/usb.h +++ b/include/linux/usb.h @@ -1862,6 +1862,9 @@ void usb_free_noncoherent(struct usb_device *dev, size_t size, * SYNCHRONOUS CALL SUPPORT * *-------------------------------------------------------------------*/ +/* Maximum value allowed for timeout in synchronous routines below */ +#define USB_MAX_SYNCHRONOUS_TIMEOUT 60000 /* ms */ + extern int usb_control_msg(struct usb_device *dev, unsigned int pipe, __u8 request, __u8 requesttype, __u16 value, __u16 index, void *data, __u16 size, int timeout); -- cgit v1.2.3 From 9f6a983cfa22ac662c86e60816d3a357d4b551e9 Mon Sep 17 00:00:00 2001 From: Jie Deng Date: Fri, 27 Feb 2026 16:49:31 +0800 Subject: usb: core: new quirk to handle devices with zero configurations Some USB devices incorrectly report bNumConfigurations as 0 in their device descriptor, which causes the USB core to reject them during enumeration. logs: usb 1-2: device descriptor read/64, error -71 usb 1-2: no configurations usb 1-2: can't read configurations, error -22 However, these devices actually work correctly when treated as having a single configuration. Add a new quirk USB_QUIRK_FORCE_ONE_CONFIG to handle such devices. When this quirk is set, assume the device has 1 configuration instead of failing with -EINVAL. This quirk is applied to the device with VID:PID 5131:2007 which exhibits this behavior. Signed-off-by: Jie Deng Link: https://patch.msgid.link/20260227084931.1527461-1-dengjie03@kylinos.cn Signed-off-by: Greg Kroah-Hartman --- Documentation/admin-guide/kernel-parameters.txt | 3 +++ drivers/usb/core/config.c | 6 +++++- drivers/usb/core/quirks.c | 5 +++++ include/linux/usb/quirks.h | 3 +++ 4 files changed, 16 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt index cb850e5290c2..7d907efe9f49 100644 --- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -8183,6 +8183,9 @@ Kernel parameters p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT (Reduce timeout of the SET_ADDRESS request from 5000 ms to 500 ms); + q = USB_QUIRK_FORCE_ONE_CONFIG (Device + claims zero configurations, + forcing to 1); Example: quirks=0781:5580:bk,0a5c:5834:gij usbhid.mousepoll= diff --git a/drivers/usb/core/config.c b/drivers/usb/core/config.c index 1cd5fa61dc76..6a1fd967e0a6 100644 --- a/drivers/usb/core/config.c +++ b/drivers/usb/core/config.c @@ -927,7 +927,11 @@ int usb_get_configuration(struct usb_device *dev) dev->descriptor.bNumConfigurations = ncfg = USB_MAXCONFIG; } - if (ncfg < 1) { + if (ncfg < 1 && dev->quirks & USB_QUIRK_FORCE_ONE_CONFIG) { + dev_info(ddev, "Device claims zero configurations, forcing to 1\n"); + dev->descriptor.bNumConfigurations = 1; + ncfg = 1; + } else if (ncfg < 1) { dev_err(ddev, "no configurations\n"); return -EINVAL; } diff --git a/drivers/usb/core/quirks.c b/drivers/usb/core/quirks.c index e347236d83e8..7bd408db05f4 100644 --- a/drivers/usb/core/quirks.c +++ b/drivers/usb/core/quirks.c @@ -140,6 +140,8 @@ static int quirks_param_set(const char *value, const struct kernel_param *kp) case 'p': flags |= USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT; break; + case 'q': + flags |= USB_QUIRK_FORCE_ONE_CONFIG; /* Ignore unrecognized flag characters */ } } @@ -589,6 +591,9 @@ static const struct usb_device_id usb_quirk_list[] = { /* VCOM device */ { USB_DEVICE(0x4296, 0x7570), .driver_info = USB_QUIRK_CONFIG_INTF_STRINGS }, + /* Noji-MCS SmartCard Reader */ + { USB_DEVICE(0x5131, 0x2007), .driver_info = USB_QUIRK_FORCE_ONE_CONFIG }, + /* INTEL VALUE SSD */ { USB_DEVICE(0x8086, 0xf1a5), .driver_info = USB_QUIRK_RESET_RESUME }, diff --git a/include/linux/usb/quirks.h b/include/linux/usb/quirks.h index 2f7bd2fdc616..b3cc7beab4a3 100644 --- a/include/linux/usb/quirks.h +++ b/include/linux/usb/quirks.h @@ -78,4 +78,7 @@ /* skip BOS descriptor request */ #define USB_QUIRK_NO_BOS BIT(17) +/* Device claims zero configurations, forcing to 1 */ +#define USB_QUIRK_FORCE_ONE_CONFIG BIT(18) + #endif /* __LINUX_USB_QUIRKS_H */ -- cgit v1.2.3 From 96189080265e6bb5dde3a4afbaf947af493e3f82 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 9 Mar 2026 14:21:37 -0600 Subject: io_uring: ensure ctx->rings is stable for task work flags manipulation If DEFER_TASKRUN | SETUP_TASKRUN is used and task work is added while the ring is being resized, it's possible for the OR'ing of IORING_SQ_TASKRUN to happen in the small window of swapping into the new rings and the old rings being freed. Prevent this by adding a 2nd ->rings pointer, ->rings_rcu, which is protected by RCU. The task work flags manipulation is inside RCU already, and if the resize ring freeing is done post an RCU synchronize, then there's no need to add locking to the fast path of task work additions. Note: this is only done for DEFER_TASKRUN, as that's the only setup mode that supports ring resizing. If this ever changes, then they too need to use the io_ctx_mark_taskrun() helper. Link: https://lore.kernel.org/io-uring/20260309062759.482210-1-naup96721@gmail.com/ Cc: stable@vger.kernel.org Fixes: 79cfe9e59c2a ("io_uring/register: add IORING_REGISTER_RESIZE_RINGS") Reported-by: Hao-Yu Yang Suggested-by: Pavel Begunkov Signed-off-by: Jens Axboe --- include/linux/io_uring_types.h | 1 + io_uring/io_uring.c | 2 ++ io_uring/register.c | 11 +++++++++++ io_uring/tw.c | 22 ++++++++++++++++++++-- 4 files changed, 34 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h index 3e4a82a6f817..dd1420bfcb73 100644 --- a/include/linux/io_uring_types.h +++ b/include/linux/io_uring_types.h @@ -388,6 +388,7 @@ struct io_ring_ctx { * regularly bounce b/w CPUs. */ struct { + struct io_rings __rcu *rings_rcu; struct llist_head work_llist; struct llist_head retry_llist; unsigned long check_cq; diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index ccab8562d273..20fdc442e014 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -2066,6 +2066,7 @@ static void io_rings_free(struct io_ring_ctx *ctx) io_free_region(ctx->user, &ctx->sq_region); io_free_region(ctx->user, &ctx->ring_region); ctx->rings = NULL; + RCU_INIT_POINTER(ctx->rings_rcu, NULL); ctx->sq_sqes = NULL; } @@ -2703,6 +2704,7 @@ static __cold int io_allocate_scq_urings(struct io_ring_ctx *ctx, if (ret) return ret; ctx->rings = rings = io_region_get_ptr(&ctx->ring_region); + rcu_assign_pointer(ctx->rings_rcu, rings); if (!(ctx->flags & IORING_SETUP_NO_SQARRAY)) ctx->sq_array = (u32 *)((char *)rings + rl->sq_array_offset); diff --git a/io_uring/register.c b/io_uring/register.c index a839b22fd392..5f3eb018fb32 100644 --- a/io_uring/register.c +++ b/io_uring/register.c @@ -633,7 +633,15 @@ overflow: ctx->sq_entries = p->sq_entries; ctx->cq_entries = p->cq_entries; + /* + * Just mark any flag we may have missed and that the application + * should act on unconditionally. Worst case it'll be an extra + * syscall. + */ + atomic_or(IORING_SQ_TASKRUN | IORING_SQ_NEED_WAKEUP, &n.rings->sq_flags); ctx->rings = n.rings; + rcu_assign_pointer(ctx->rings_rcu, n.rings); + ctx->sq_sqes = n.sq_sqes; swap_old(ctx, o, n, ring_region); swap_old(ctx, o, n, sq_region); @@ -642,6 +650,9 @@ overflow: out: spin_unlock(&ctx->completion_lock); mutex_unlock(&ctx->mmap_lock); + /* Wait for concurrent io_ctx_mark_taskrun() */ + if (to_free == &o) + synchronize_rcu_expedited(); io_register_free_rings(ctx, to_free); if (ctx->sq_data) diff --git a/io_uring/tw.c b/io_uring/tw.c index 1ee2b8ab07c8..2f2b4ac4b126 100644 --- a/io_uring/tw.c +++ b/io_uring/tw.c @@ -152,6 +152,21 @@ void tctx_task_work(struct callback_head *cb) WARN_ON_ONCE(ret); } +/* + * Sets IORING_SQ_TASKRUN in the sq_flags shared with userspace, using the + * RCU protected rings pointer to be safe against concurrent ring resizing. + */ +static void io_ctx_mark_taskrun(struct io_ring_ctx *ctx) +{ + lockdep_assert_in_rcu_read_lock(); + + if (ctx->flags & IORING_SETUP_TASKRUN_FLAG) { + struct io_rings *rings = rcu_dereference(ctx->rings_rcu); + + atomic_or(IORING_SQ_TASKRUN, &rings->sq_flags); + } +} + void io_req_local_work_add(struct io_kiocb *req, unsigned flags) { struct io_ring_ctx *ctx = req->ctx; @@ -206,8 +221,7 @@ void io_req_local_work_add(struct io_kiocb *req, unsigned flags) */ if (!head) { - if (ctx->flags & IORING_SETUP_TASKRUN_FLAG) - atomic_or(IORING_SQ_TASKRUN, &ctx->rings->sq_flags); + io_ctx_mark_taskrun(ctx); if (ctx->has_evfd) io_eventfd_signal(ctx, false); } @@ -231,6 +245,10 @@ void io_req_normal_work_add(struct io_kiocb *req) if (!llist_add(&req->io_task_work.node, &tctx->task_list)) return; + /* + * Doesn't need to use ->rings_rcu, as resizing isn't supported for + * !DEFER_TASKRUN. + */ if (ctx->flags & IORING_SETUP_TASKRUN_FLAG) atomic_or(IORING_SQ_TASKRUN, &ctx->rings->sq_flags); -- cgit v1.2.3 From 94a4b1f959989de9c54d43c3a102fb1ee92e1414 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sat, 7 Mar 2026 17:50:53 -0300 Subject: ipv6: move the disable_ipv6_mod knob to core code From: Jakub Kicinski Make sure disable_ipv6_mod itself is not part of the IPv6 module, in case core code wants to refer to it. We will remove support for IPv6=m soon, this change helps make fixes we commit before that less messy. Link: https://patch.msgid.link/20260307-net-nd_tbl_fixes-v4-1-e2677e85628c@suse.com Signed-off-by: Jakub Kicinski --- include/linux/ipv6.h | 7 ++++++- net/ipv4/af_inet.c | 6 ++++++ net/ipv6/af_inet6.c | 8 -------- 3 files changed, 12 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index 443053a76dcf..a7421382a916 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -333,7 +333,12 @@ struct tcp6_timewait_sock { }; #if IS_ENABLED(CONFIG_IPV6) -bool ipv6_mod_enabled(void); +extern int disable_ipv6_mod; + +static inline bool ipv6_mod_enabled(void) +{ + return disable_ipv6_mod == 0; +} static inline struct ipv6_pinfo *inet6_sk(const struct sock *__sk) { diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 8036e76aa1e4..c7731e300a44 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -124,6 +124,12 @@ #include +/* Keep the definition of IPv6 disable here for now, to avoid annoying linker + * issues in case IPv6=m + */ +int disable_ipv6_mod; +EXPORT_SYMBOL(disable_ipv6_mod); + /* The inetsw table contains everything that inet_create needs to * build a new socket. */ diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 23cc9b4cb2f1..4cbd45b68088 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -86,8 +86,6 @@ struct ipv6_params ipv6_defaults = { .autoconf = 1, }; -static int disable_ipv6_mod; - module_param_named(disable, disable_ipv6_mod, int, 0444); MODULE_PARM_DESC(disable, "Disable IPv6 module such that it is non-functional"); @@ -97,12 +95,6 @@ MODULE_PARM_DESC(disable_ipv6, "Disable IPv6 on all interfaces"); module_param_named(autoconf, ipv6_defaults.autoconf, int, 0444); MODULE_PARM_DESC(autoconf, "Enable IPv6 address autoconfiguration on all interfaces"); -bool ipv6_mod_enabled(void) -{ - return disable_ipv6_mod == 0; -} -EXPORT_SYMBOL_GPL(ipv6_mod_enabled); - static struct ipv6_pinfo *inet6_sk_generic(struct sock *sk) { const int offset = sk->sk_prot->ipv6_pinfo_offset; -- cgit v1.2.3