- Linux-stable-mirror - lists.linaro.org

[PATCH v3] Bluetooth: btusb: mediatek: Avoid btusb_mtk_claim_iso_intf() NULL deref

by Douglas Anderson

In btusb_mtk_setup(), we set `btmtk_data->isopkt_intf` to: usb_ifnum_to_if(data->udev, MTK_ISO_IFNUM) That function can return NULL in some cases. Even when it returns NULL, though, we still go on to call btusb_mtk_claim_iso_intf(). As of commit e9087e828827 ("Bluetooth: btusb: mediatek: Add locks for usb_driver_claim_interface()"), calling btusb_mtk_claim_iso_intf() when `btmtk_data->isopkt_intf` is NULL will cause a crash because we'll end up passing a bad pointer to device_lock(). Prior to that commit we'd pass the NULL pointer directly to usb_driver_claim_interface() which would detect it and return an error, which was handled. Resolve the crash in btusb_mtk_claim_iso_intf() by adding a NULL check at the start of the function. This makes the code handle a NULL `btmtk_data->isopkt_intf` the same way it did before the problematic commit (just with a slight change to the error message printed). Reported-by: IncogCyberpunk <incogcyberpunk(a)proton.me> Closes: http://lore.kernel.org/r/a380d061-479e-4713-bddd-1d6571ca7e86@leemhuis.info Fixes: e9087e828827 ("Bluetooth: btusb: mediatek: Add locks for usb_driver_claim_interface()") Cc: stable(a)vger.kernel.org Tested-by: IncogCyberpunk <incogcyberpunk(a)proton.me> Signed-off-by: Douglas Anderson <dianders(a)chromium.org> --- Changes in v3: - Added Cc to stable. - Added IncogCyberpunk Tested-by tag. - v2: https://patch.msgid.link/20251119092641.v2.1.I1ae7aebc967e52c7c4be7aa65fbd8… Changes in v2: - Rebase atop commit 529ac8e706c3 ("Bluetooth: ... mtk iso interface") - v1: https://patch.msgid.link/20251119085354.1.I1ae7aebc967e52c7c4be7aa65fbd8173… drivers/bluetooth/btusb.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c index fcc62e2fb641..683ac02e964b 100644 --- a/drivers/bluetooth/btusb.c +++ b/drivers/bluetooth/btusb.c @@ -2751,6 +2751,11 @@ static void btusb_mtk_claim_iso_intf(struct btusb_data *data) if (!btmtk_data) return; + if (!btmtk_data->isopkt_intf) { + bt_dev_err(data->hdev, "Can't claim NULL iso interface"); + return; + } + /* * The function usb_driver_claim_interface() is documented to need * locks held if it's not called from a probe routine. The code here -- 2.52.0.rc1.455.g30608eb744-goog

2 days, 2 hours

6
6
0 0

[PATCHv2 1/2] kernel/kexec: Change the prototype of kimage_map_segment()

by Pingfan Liu

The kexec segment index will be required to extract the corresponding information for that segment in kimage_map_segment(). Additionally, kexec_segment already holds the kexec relocation destination address and size. Therefore, the prototype of kimage_map_segment() can be changed. Fixes: 07d24902977e ("kexec: enable CMA based contiguous allocation") Signed-off-by: Pingfan Liu <piliu(a)redhat.com> Cc: Andrew Morton <akpm(a)linux-foundation.org> Cc: Baoquan He <bhe(a)redhat.com> Cc: Mimi Zohar <zohar(a)linux.ibm.com> Cc: Roberto Sassu <roberto.sassu(a)huawei.com> Cc: Alexander Graf <graf(a)amazon.com> Cc: Steven Chen <chenste(a)linux.microsoft.com> Cc: <stable(a)vger.kernel.org> To: kexec(a)lists.infradead.org To: linux-integrity(a)vger.kernel.org --- include/linux/kexec.h | 4 ++-- kernel/kexec_core.c | 9 ++++++--- security/integrity/ima/ima_kexec.c | 4 +--- 3 files changed, 9 insertions(+), 8 deletions(-) diff --git a/include/linux/kexec.h b/include/linux/kexec.h index ff7e231b0485..8a22bc9b8c6c 100644 --- a/include/linux/kexec.h +++ b/include/linux/kexec.h @@ -530,7 +530,7 @@ extern bool kexec_file_dbg_print; #define kexec_dprintk(fmt, arg...) \ do { if (kexec_file_dbg_print) pr_info(fmt, ##arg); } while (0) -extern void *kimage_map_segment(struct kimage *image, unsigned long addr, unsigned long size); +extern void *kimage_map_segment(struct kimage *image, int idx); extern void kimage_unmap_segment(void *buffer); #else /* !CONFIG_KEXEC_CORE */ struct pt_regs; @@ -540,7 +540,7 @@ static inline void __crash_kexec(struct pt_regs *regs) { } static inline void crash_kexec(struct pt_regs *regs) { } static inline int kexec_should_crash(struct task_struct *p) { return 0; } static inline int kexec_crash_loaded(void) { return 0; } -static inline void *kimage_map_segment(struct kimage *image, unsigned long addr, unsigned long size) +static inline void *kimage_map_segment(struct kimage *image, int idx) { return NULL; } static inline void kimage_unmap_segment(void *buffer) { } #define kexec_in_progress false diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index fa00b239c5d9..9a1966207041 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -960,17 +960,20 @@ int kimage_load_segment(struct kimage *image, int idx) return result; } -void *kimage_map_segment(struct kimage *image, - unsigned long addr, unsigned long size) +void *kimage_map_segment(struct kimage *image, int idx) { + unsigned long addr, size, eaddr; unsigned long src_page_addr, dest_page_addr = 0; - unsigned long eaddr = addr + size; kimage_entry_t *ptr, entry; struct page **src_pages; unsigned int npages; void *vaddr = NULL; int i; + addr = image->segment[idx].mem; + size = image->segment[idx].memsz; + eaddr = addr + size; + /* * Collect the source pages and map them in a contiguous VA range. */ diff --git a/security/integrity/ima/ima_kexec.c b/security/integrity/ima/ima_kexec.c index 7362f68f2d8b..5beb69edd12f 100644 --- a/security/integrity/ima/ima_kexec.c +++ b/security/integrity/ima/ima_kexec.c @@ -250,9 +250,7 @@ void ima_kexec_post_load(struct kimage *image) if (!image->ima_buffer_addr) return; - ima_kexec_buffer = kimage_map_segment(image, - image->ima_buffer_addr, - image->ima_buffer_size); + ima_kexec_buffer = kimage_map_segment(image, image->ima_segment_index); if (!ima_kexec_buffer) { pr_err("Could not map measurements buffer.\n"); return; -- 2.49.0

2 days, 2 hours

3
17
0 0

[PATCH] LoongArch: Fix arch_dup_task_struct() for CONFIG_RANDSTRUCT

by Huacai Chen

Now the optimized version of arch_dup_task_struct() for LoongArch assumes 'thread' is the last member of 'task_struct'. But this is not true if CONFIG_RANDSTRUCT is enabled after Linux-6.16. So fix the arch_dup_task_struct() function for CONFIG_RANDSTRUCT by copying the whole 'task_struct'. Cc: stable(a)vger.kernel.org Signed-off-by: Huacai Chen <chenhuacai(a)loongson.cn> --- arch/loongarch/kernel/process.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/arch/loongarch/kernel/process.c b/arch/loongarch/kernel/process.c index efd9edf65603..d1e04f9e0f79 100644 --- a/arch/loongarch/kernel/process.c +++ b/arch/loongarch/kernel/process.c @@ -130,6 +130,11 @@ int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src) preempt_enable(); + if (IS_ENABLED(CONFIG_RANDSTRUCT)) { + memcpy(dst, src, sizeof(struct task_struct)); + return 0; + } + if (!used_math()) memcpy(dst, src, offsetof(struct task_struct, thread.fpu.fpr)); else -- 2.47.3

2 days, 2 hours

1
0
0 0

[PATCH] ksmbd: ipc: fix use-after-free in ipc_msg_send_request

by Qianchang Zhao

ipc_msg_send_request() waits for a generic netlink reply using an ipc_msg_table_entry on the stack. The generic netlink handler (handle_generic_event()/handle_response()) fills entry->response under ipc_msg_table_lock, but ipc_msg_send_request() used to validate and free entry->response without holding the same lock. Under high concurrency this allows a race where handle_response() is copying data into entry->response while ipc_msg_send_request() has just freed it, leading to a slab-use-after-free reported by KASAN in handle_generic_event(): BUG: KASAN: slab-use-after-free in handle_generic_event+0x3c4/0x5f0 [ksmbd] Write of size 12 at addr ffff888198ee6e20 by task pool/109349 ... Freed by task: kvfree ipc_msg_send_request [ksmbd] ksmbd_rpc_open -> ksmbd_session_rpc_open [ksmbd] Fix by: - Taking ipc_msg_table_lock in ipc_msg_send_request() while validating entry->response, freeing it when invalid, and removing the entry from ipc_msg_table. - Returning the final entry->response pointer to the caller only after the hash entry is removed under the lock. - Returning NULL in the error path, preserving the original API semantics. This makes all accesses to entry->response consistent with handle_response(), which already updates and fills the response buffer under ipc_msg_table_lock, and closes the race that allowed the UAF. Reported-by: Qianchang Zhao <pioooooooooip(a)gmail.com> Reported-by: Zhitong Liu <liuzhitong1993(a)gmail.com> Cc: stable(a)vger.kernel.org Signed-off-by: Qianchang Zhao <pioooooooooip(a)gmail.com> --- fs/smb/server/transport_ipc.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/fs/smb/server/transport_ipc.c b/fs/smb/server/transport_ipc.c index 46f87fd1c..7b1a060da 100644 --- a/fs/smb/server/transport_ipc.c +++ b/fs/smb/server/transport_ipc.c @@ -532,6 +532,7 @@ static int ipc_validate_msg(struct ipc_msg_table_entry *entry) static void *ipc_msg_send_request(struct ksmbd_ipc_msg *msg, unsigned int handle) { struct ipc_msg_table_entry entry; + void *response = NULL; int ret; if ((int)handle < 0) @@ -553,6 +554,8 @@ static void *ipc_msg_send_request(struct ksmbd_ipc_msg *msg, unsigned int handle ret = wait_event_interruptible_timeout(entry.wait, entry.response != NULL, IPC_WAIT_TIMEOUT); + + down_write(&ipc_msg_table_lock); if (entry.response) { ret = ipc_validate_msg(&entry); if (ret) { @@ -560,11 +563,19 @@ static void *ipc_msg_send_request(struct ksmbd_ipc_msg *msg, unsigned int handle entry.response = NULL; } } + + response = entry.response; + hash_del(&entry.ipc_table_hlist); + up_write(&ipc_msg_table_lock); + + return response; + out: down_write(&ipc_msg_table_lock); hash_del(&entry.ipc_table_hlist); up_write(&ipc_msg_table_lock); - return entry.response; + + return NULL; } static int ksmbd_ipc_heartbeat_request(void) -- 2.34.1

2 days, 3 hours

2
1
0 0

[PATCH] smb: client: fix memory leak in cifs_construct_tcon()

by Paulo Alcantara

When having a multiuser mount with domain= specified and using cifscreds, cifs_set_cifscreds() will end up setting @ctx->domainname, so it needs to be freed before leaving cifs_construct_tcon(). This fixes the following memory leak reported by kmemleak: mount.cifs //srv/share /mnt -o domain=ZELDA,multiuser,... su - testuser cifscreds add -d ZELDA -u testuser ... ls /mnt/1 ... umount /mnt echo scan > /sys/kernel/debug/kmemleak cat /sys/kernel/debug/kmemleak unreferenced object 0xffff8881203c3f08 (size 8): comm "ls", pid 5060, jiffies 4307222943 hex dump (first 8 bytes): 5a 45 4c 44 41 00 cc cc ZELDA... backtrace (crc d109a8cf): __kmalloc_node_track_caller_noprof+0x572/0x710 kstrdup+0x3a/0x70 cifs_sb_tlink+0x1209/0x1770 [cifs] cifs_get_fattr+0xe1/0xf50 [cifs] cifs_get_inode_info+0xb5/0x240 [cifs] cifs_revalidate_dentry_attr+0x2d1/0x470 [cifs] cifs_getattr+0x28e/0x450 [cifs] vfs_getattr_nosec+0x126/0x180 vfs_statx+0xf6/0x220 do_statx+0xab/0x110 __x64_sys_statx+0xd5/0x130 do_syscall_64+0xbb/0x380 entry_SYSCALL_64_after_hwframe+0x77/0x7f Fixes: f2aee329a68f ("cifs: set domainName when a domain-key is used in multiuser") Signed-off-by: Paulo Alcantara (Red Hat) <pc(a)manguebit.org> Cc: Jay Shin <jaeshin(a)redhat.com> Cc: stable(a)vger.kernel.org Cc: linux-cifs(a)vger.kernel.org --- fs/smb/client/connect.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/smb/client/connect.c b/fs/smb/client/connect.c index 55cb4b0cbd48..2f94d93b95e9 100644 --- a/fs/smb/client/connect.c +++ b/fs/smb/client/connect.c @@ -4451,6 +4451,7 @@ cifs_construct_tcon(struct cifs_sb_info *cifs_sb, kuid_t fsuid) out: kfree(ctx->username); + kfree(ctx->domainname); kfree_sensitive(ctx->password); kfree(origin_fullpath); kfree(ctx); -- 2.51.1

2 days, 7 hours

2
2
0 0

Re: [PATCH] e1000: fix OOB in e1000_tbi_should_accept()

by Tony Nguyen

On 11/23/2025 7:45 PM, Guangshuo Li wrote: > Hi Tony, all, > > Thanks for the review. As suggested by Tony, I’ll keep the declarations > at the top and place the bounds checks before assigning last_byte. I’ll > send a v2 with the following change: > > static bool e1000_tbi_should_accept(struct e1000_adapter *adapter, > u8 status, u8 errors, > u32 length, const u8 *data) > { > struct e1000_hw *hw = &adapter->hw; > u8 last_byte; > > /* Guard against OOB on data[length - 1] */ > if (unlikely(!length)) > return false; > > /* Upper bound: length must not exceed rx_buffer_len */ > if (unlikely(length > adapter->rx_buffer_len)) > return false; > > last_byte = data[length - 1]; > > /* existing logic follows ... */ > } > Please let me know if further adjustments are preferred. This looks along the lines of what I was expecting. Thanks, Tony > Best regards, > Guangshuo Li

2 days, 8 hours

1
0
0 0

[patch V2 1/3] x86/msi: Make irq_retrigger() functional for posted MSI

by Thomas Gleixner

From: Thomas Gleixner <tglx(a)linutronix.de> Luigi reported that retriggering a posted MSI interrupt does not work correctly. The reason is that the retrigger happens at the vector domain by sending an IPI to the actual vector on the target CPU. That works correctly exactly once because the posted MSI interrupt chip does not issue an EOI as that's only required for the posted MSI notification vector itself. As a consequence the vector becomes stale in the ISR, which not only affects this vector but also any lower priority vector in the affected APIC because the ISR bit is not cleared. Luigi proposed to set the vector in the remap PIR bitmap and raise the posted MSI notification vector. That works, but that still does not cure a related problem: If there is ever a stray interrupt on such a vector, then the related APIC ISR bit becomes stale due to the lack of EOI as described above. Unlikely to happen, but if it happens it's not debuggable at all. So instead of playing games with the PIR, this can be actually solved for both cases by: 1) Keeping track of the posted interrupt vector handler state 2) Implementing a posted MSI specific irq_ack() callback which checks that state. If the posted vector handler is inactive it issues an EOI, otherwise it delegates that to the posted handler. This is correct versus affinity changes and concurrent events on the posted vector as the actual handler invocation is serialized through the interrupt descriptor lock. Fixes: ed1e48ea4370 ("iommu/vt-d: Enable posted mode for device MSIs") Reported-by: Luigi Rizzo <lrizzo(a)google.com> Signed-off-by: Thomas Gleixner <tglx(a)linutronix.de> Tested-by: Luigi Rizzo <lrizzo(a)google.com> Cc: stable(a)vger.kernel.org Closes: https://lore.kernel.org/lkml/20251124104836.3685533-1-lrizzo@google.com --- V2: Use __this_cpu...() - Luigi --- arch/x86/include/asm/irq_remapping.h | 7 +++++++ arch/x86/kernel/irq.c | 23 +++++++++++++++++++++++ drivers/iommu/intel/irq_remapping.c | 8 ++++---- 3 files changed, 34 insertions(+), 4 deletions(-) --- a/arch/x86/include/asm/irq_remapping.h +++ b/arch/x86/include/asm/irq_remapping.h @@ -87,4 +87,11 @@ static inline void panic_if_irq_remap(co } #endif /* CONFIG_IRQ_REMAP */ + +#ifdef CONFIG_X86_POSTED_MSI +void intel_ack_posted_msi_irq(struct irq_data *irqd); +#else +#define intel_ack_posted_msi_irq NULL +#endif + #endif /* __X86_IRQ_REMAPPING_H */ --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -396,6 +396,7 @@ DEFINE_IDTENTRY_SYSVEC_SIMPLE(sysvec_kvm /* Posted Interrupt Descriptors for coalesced MSIs to be posted */ DEFINE_PER_CPU_ALIGNED(struct pi_desc, posted_msi_pi_desc); +static DEFINE_PER_CPU_CACHE_HOT(bool, posted_msi_handler_active); void intel_posted_msi_init(void) { @@ -413,6 +414,25 @@ void intel_posted_msi_init(void) this_cpu_write(posted_msi_pi_desc.ndst, destination); } +void intel_ack_posted_msi_irq(struct irq_data *irqd) +{ + irq_move_irq(irqd); + + /* + * Handle the rare case that irq_retrigger() raised the actual + * assigned vector on the target CPU, which means that it was not + * invoked via the posted MSI handler below. In that case APIC EOI + * is required as otherwise the ISR entry becomes stale and lower + * priority interrupts are never going to be delivered after that. + * + * If the posted handler invoked the device interrupt handler then + * the EOI would be premature because it would acknowledge the + * posted vector. + */ + if (unlikely(!__this_cpu_read(posted_msi_handler_active))) + apic_eoi(); +} + static __always_inline bool handle_pending_pir(unsigned long *pir, struct pt_regs *regs) { unsigned long pir_copy[NR_PIR_WORDS]; @@ -445,6 +465,8 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_posted_msi pid = this_cpu_ptr(&posted_msi_pi_desc); + /* Mark the handler active for intel_ack_posted_msi_irq() */ + __this_cpu_write(posted_msi_handler_active, true); inc_irq_stat(posted_msi_notification_count); irq_enter(); @@ -473,6 +495,7 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_posted_msi apic_eoi(); irq_exit(); + __this_cpu_write(posted_msi_handler_active, false); set_irq_regs(old_regs); } #endif /* X86_POSTED_MSI */ --- a/drivers/iommu/intel/irq_remapping.c +++ b/drivers/iommu/intel/irq_remapping.c @@ -1303,17 +1303,17 @@ static struct irq_chip intel_ir_chip = { * irq_enter(); * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * apic_eoi() @@ -1322,7 +1322,7 @@ static struct irq_chip intel_ir_chip = { */ static struct irq_chip intel_ir_chip_post_msi = { .name = "INTEL-IR-POST", - .irq_ack = irq_move_irq, + .irq_ack = intel_ack_posted_msi_irq, .irq_set_affinity = intel_ir_set_affinity, .irq_compose_msi_msg = intel_ir_compose_msi_msg, .irq_set_vcpu_affinity = intel_ir_set_vcpu_affinity,

2 days, 9 hours

1
0
0 0

[PATCH v3] KVM: x86: Add x2APIC "features" to control EOI broadcast suppression

by Khushit Shah

Add two flags for KVM_CAP_X2APIC_API to allow userspace to control support for Suppress EOI Broadcasts, which KVM completely mishandles. When x2APIC support was first added, KVM incorrectly advertised and "enabled" Suppress EOI Broadcast, without fully supporting the I/O APIC side of the equation, i.e. without adding directed EOI to KVM's in-kernel I/O APIC. That flaw was carried over to split IRQCHIP support, i.e. KVM advertised support for Suppress EOI Broadcasts irrespective of whether or not the userspace I/O APIC implementation supported directed EOIs. Even worse, KVM didn't actually suppress EOI broadcasts, i.e. userspace VMMs without support for directed EOI came to rely on the "spurious" broadcasts. KVM "fixed" the in-kernel I/O APIC implementation by completely disabling support for Suppress EOI Broadcasts in commit 0bcc3fb95b97 ("KVM: lapic: stop advertising DIRECTED_EOI when in-kernel IOAPIC is in use"), but didn't do anything to remedy userspace I/O APIC implementations. KVM's bogus handling of Suppress EOI Broadcast is problematic when the guest relies on interrupts being masked in the I/O APIC until well after the initial local APIC EOI. E.g. Windows with Credential Guard enabled handles interrupts in the following order: 1. Interrupt for L2 arrives. 2. L1 APIC EOIs the interrupt. 3. L1 resumes L2 and injects the interrupt. 4. L2 EOIs after servicing. 5. L1 performs the I/O APIC EOI. Because KVM EOIs the I/O APIC at step #2, the guest can get an interrupt storm, e.g. if the IRQ line is still asserted and userspace reacts to the EOI by re-injecting the IRQ, because the guest doesn't de-assert the line until step #4, and doesn't expect the interrupt to be re-enabled until step #5. Unfortunately, simply "fixing" the bug isn't an option, as KVM has no way of knowing if the userspace I/O APIC supports directed EOIs, i.e. suppressing EOI broadcasts would result in interrupts being stuck masked in the userspace I/O APIC due to step #5 being ignored by userspace. And fully disabling support for Suppress EOI Broadcast is also undesirable, as picking up the fix would require a guest reboot, *and* more importantly would change the virtual CPU model exposed to the guest without any buy-in from userspace. Add two flags to allow userspace to choose exactly how to solve the immediate issue, and in the long term to allow userspace to control the virtual CPU model that is exposed to the guest (KVM should never have enabled support for Suppress EOI Broadcast without a userspace opt-in). Note, Suppress EOI Broadcasts is defined only in Intel's SDM, not in AMD's APM. But the bit is writable on some AMD CPUs, e.g. Turin, and KVM's ABI is to support Directed EOI (KVM's name) irrespective of guest CPU vendor. Fixes: 7543a635aa09 ("KVM: x86: Add KVM exit for IOAPIC EOIs") Closes: https://lore.kernel.org/kvm/7D497EF1-607D-4D37-98E7-DAF95F099342@nutanix.com Cc: stable(a)vger.kernel.org Co-developed-by: Sean Christopherson <seanjc(a)google.com> Signed-off-by: Sean Christopherson <seanjc(a)google.com> Signed-off-by: Khushit Shah <khushit.shah(a)nutanix.com> --- v3: - Resend with correct patch contents; v2 mail formatting was broken. No functional changes beyond the naming and grammar feedback from v1. Testing: I ran the tests with QEMU 9.1 and a 6.12 kernel with the patch applied. - With an unmodified QEMU build, KVM's LAPIC SEOIB behavior remains unchanged. - Invoking the x2APIC API with KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK correctly suppresses LAPIC -> IOAPIC EOI broadcasts (verified via KVM tracepoints). - Invoking the x2APIC API with KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST results in SEOIB not being advertised to the guest, as expected (confirmed by checking the LAPIC LVR value inside the guest). I'll send the corresponding QEMU-side patch shortly. --- Documentation/virt/kvm/api.rst | 14 ++++++++++++-- arch/x86/include/asm/kvm_host.h | 2 ++ arch/x86/include/uapi/asm/kvm.h | 6 ++++-- arch/x86/kvm/lapic.c | 13 +++++++++++++ arch/x86/kvm/x86.c | 12 +++++++++--- 5 files changed, 40 insertions(+), 7 deletions(-) diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst index 57061fa29e6a..4141d2bd8156 100644 --- a/Documentation/virt/kvm/api.rst +++ b/Documentation/virt/kvm/api.rst @@ -7800,8 +7800,10 @@ Will return -EBUSY if a VCPU has already been created. Valid feature flags in args[0] are:: - #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) - #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) + #define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) + #define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) + #define KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK (1ULL << 2) + #define KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST (1ULL << 3) Enabling KVM_X2APIC_API_USE_32BIT_IDS changes the behavior of KVM_SET_GSI_ROUTING, KVM_SIGNAL_MSI, KVM_SET_LAPIC, and KVM_GET_LAPIC, @@ -7814,6 +7816,14 @@ as a broadcast even in x2APIC mode in order to support physical x2APIC without interrupt remapping. This is undesirable in logical mode, where 0xff represents CPUs 0-7 in cluster 0. +Setting KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK overrides +KVM's quirky behavior of not actually suppressing EOI broadcasts for split IRQ +chips when support for Suppress EOI Broadcasts is advertised to the guest. + +Setting KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST disables support for +Suppress EOI Broadcasts entirely, i.e. instructs KVM to NOT advertise support +to the guest and thus disallow enabling EOI broadcast suppression in SPIV. + 7.8 KVM_CAP_S390_USER_INSTR0 ---------------------------- diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 48598d017d6f..f6fdc0842c05 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -1480,6 +1480,8 @@ struct kvm_arch { bool x2apic_format; bool x2apic_broadcast_quirk_disabled; + bool disable_ignore_suppress_eoi_broadcast_quirk; + bool x2apic_disable_suppress_eoi_broadcast; bool has_mapped_host_mmio; bool guest_can_read_msr_platform_info; diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h index d420c9c066d4..7255385b6d80 100644 --- a/arch/x86/include/uapi/asm/kvm.h +++ b/arch/x86/include/uapi/asm/kvm.h @@ -913,8 +913,10 @@ struct kvm_sev_snp_launch_finish { __u64 pad1[4]; }; -#define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) -#define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) +#define KVM_X2APIC_API_USE_32BIT_IDS (1ULL << 0) +#define KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK (1ULL << 1) +#define KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK (1ULL << 2) +#define KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST (1ULL << 3) struct kvm_hyperv_eventfd { __u32 conn_id; diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 0ae7f913d782..cf8a2162872b 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -562,6 +562,7 @@ void kvm_apic_set_version(struct kvm_vcpu *vcpu) * IOAPIC. */ if (guest_cpu_cap_has(vcpu, X86_FEATURE_X2APIC) && + !vcpu->kvm->arch.x2apic_disable_suppress_eoi_broadcast && !ioapic_in_kernel(vcpu->kvm)) v |= APIC_LVR_DIRECTED_EOI; kvm_lapic_set_reg(apic, APIC_LVR, v); @@ -1517,6 +1518,18 @@ static void kvm_ioapic_send_eoi(struct kvm_lapic *apic, int vector) /* Request a KVM exit to inform the userspace IOAPIC. */ if (irqchip_split(apic->vcpu->kvm)) { + /* + * Don't exit to userspace if the guest has enabled Directed + * EOI, a.k.a. Suppress EOI Broadcasts, in which case the local + * APIC doesn't broadcast EOIs (the guest must EOI the target + * I/O APIC(s) directly). Ignore the suppression if userspace + * has NOT disabled KVM's quirk (KVM advertised support for + * Suppress EOI Broadcasts without actually suppressing EOIs). + */ + if ((kvm_lapic_get_reg(apic, APIC_SPIV) & APIC_SPIV_DIRECTED_EOI) && + apic->vcpu->kvm->arch.disable_ignore_suppress_eoi_broadcast_quirk) + return; + apic->vcpu->arch.pending_ioapic_eoi = vector; kvm_make_request(KVM_REQ_IOAPIC_EOI_EXIT, apic->vcpu); return; diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index c9c2aa6f4705..e1b6fe783615 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -121,8 +121,11 @@ static u64 __read_mostly efer_reserved_bits = ~((u64)EFER_SCE); #define KVM_CAP_PMU_VALID_MASK KVM_PMU_CAP_DISABLE -#define KVM_X2APIC_API_VALID_FLAGS (KVM_X2APIC_API_USE_32BIT_IDS | \ - KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK) +#define KVM_X2APIC_API_VALID_FLAGS \ + (KVM_X2APIC_API_USE_32BIT_IDS | \ + KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK | \ + KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK | \ + KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST) static void update_cr8_intercept(struct kvm_vcpu *vcpu); static void process_nmi(struct kvm_vcpu *vcpu); @@ -6782,7 +6785,10 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm, kvm->arch.x2apic_format = true; if (cap->args[0] & KVM_X2APIC_API_DISABLE_BROADCAST_QUIRK) kvm->arch.x2apic_broadcast_quirk_disabled = true; - + if (cap->args[0] & KVM_X2APIC_API_DISABLE_IGNORE_SUPPRESS_EOI_BROADCAST_QUIRK) + kvm->arch.disable_ignore_suppress_eoi_broadcast_quirk = true; + if (cap->args[0] & KVM_X2APIC_API_DISABLE_SUPPRESS_EOI_BROADCAST) + kvm->arch.x2apic_disable_suppress_eoi_broadcast = true; r = 0; break; case KVM_CAP_X86_DISABLE_EXITS: -- 2.39.3

2 days, 9 hours

2
1
0 0

Re: [PATCH] fuse: fix io-uring list corruption for terminated non-committed requests

by Joanne Koong

On Tue, Nov 25, 2025 at 10:15 AM Joanne Koong <joannelkoong(a)gmail.com> wrote: > > When a request is terminated before it has been committed, the request > is not removed from the queue's list. This leaves a dangling list entry > that leads to list corruption and use-after-free issues. > > Remove the request from the queue's list for terminated non-committed > requests. > > Signed-off-by: Joanne Koong <joannelkoong(a)gmail.com> > Fixes: c090c8abae4b ("fuse: Add io-uring sqe commit and fetch support") Sorry, forgot to add the stable tag. There should be this line: Cc: stable(a)vger.kernel.org > --- > fs/fuse/dev_uring.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/fs/fuse/dev_uring.c b/fs/fuse/dev_uring.c > index 0066c9c0a5d5..7760fe4e1f9e 100644 > --- a/fs/fuse/dev_uring.c > +++ b/fs/fuse/dev_uring.c > @@ -86,6 +86,7 @@ static void fuse_uring_req_end(struct fuse_ring_ent *ent, struct fuse_req *req, > lockdep_assert_not_held(&queue->lock); > spin_lock(&queue->lock); > ent->fuse_req = NULL; > + list_del_init(&req->list); > if (test_bit(FR_BACKGROUND, &req->flags)) { > queue->active_background--; > spin_lock(&fc->bg_lock); > -- > 2.47.3 >

2 days, 10 hours

2
1
0 0

[patch 1/3] x86/msi: Make irq_retrigger() functional for posted MSI

by Thomas Gleixner

Luigi reported that retriggering a posted MSI interrupt does not work correctly. The reason is that the retrigger happens at the vector domain by sending an IPI to the actual vector on the target CPU. That works correctly exactly once because the posted MSI interrupt chip does not issue an EOI as that's only required for the posted MSI notification vector itself. As a consequence the vector becomes stale in the ISR, which not only affects this vector but also any lower priority vector in the affected APIC because the ISR bit is not cleared. Luigi proposed to set the vector in the remap PIR bitmap and raise the posted MSI notification vector. That works, but that still does not cure a related problem: If there is ever a stray interrupt on such a vector, then the related APIC ISR bit becomes stale due to the lack of EOI as described above. Unlikely to happen, but if it happens it's not debuggable at all. So instead of playing games with the PIR, this can be actually solved for both cases by: 1) Keeping track of the posted interrupt vector handler state 2) Implementing a posted MSI specific irq_ack() callback which checks that state. If the posted vector handler is inactive it issues an EOI, otherwise it delegates that to the posted handler. This is correct versus affinity changes and concurrent events on the posted vector as the actual handler invocation is serialized through the interrupt descriptor lock. Fixes: ed1e48ea4370 ("iommu/vt-d: Enable posted mode for device MSIs") Reported-by: Luigi Rizzo <lrizzo(a)google.com> Signed-off-by: Thomas Gleixner <tglx(a)linutronix.de> Tested-by: Luigi Rizzo <lrizzo(a)google.com> Cc: stable(a)vger.kernel.org Closes: https://lore.kernel.org/lkml/20251124104836.3685533-1-lrizzo@google.com --- arch/x86/include/asm/irq_remapping.h | 7 +++++++ arch/x86/kernel/irq.c | 23 +++++++++++++++++++++++ drivers/iommu/intel/irq_remapping.c | 8 ++++---- 3 files changed, 34 insertions(+), 4 deletions(-) --- a/arch/x86/include/asm/irq_remapping.h +++ b/arch/x86/include/asm/irq_remapping.h @@ -87,4 +87,11 @@ static inline void panic_if_irq_remap(co } #endif /* CONFIG_IRQ_REMAP */ + +#ifdef CONFIG_X86_POSTED_MSI +void intel_ack_posted_msi_irq(struct irq_data *irqd); +#else +#define intel_ack_posted_msi_irq NULL +#endif + #endif /* __X86_IRQ_REMAPPING_H */ --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -397,6 +397,7 @@ DEFINE_IDTENTRY_SYSVEC_SIMPLE(sysvec_kvm /* Posted Interrupt Descriptors for coalesced MSIs to be posted */ DEFINE_PER_CPU_ALIGNED(struct pi_desc, posted_msi_pi_desc); +static DEFINE_PER_CPU_CACHE_HOT(bool, posted_msi_handler_active); void intel_posted_msi_init(void) { @@ -414,6 +415,25 @@ void intel_posted_msi_init(void) this_cpu_write(posted_msi_pi_desc.ndst, destination); } +void intel_ack_posted_msi_irq(struct irq_data *irqd) +{ + irq_move_irq(irqd); + + /* + * Handle the rare case that irq_retrigger() raised the actual + * assigned vector on the target CPU, which means that it was not + * invoked via the posted MSI handler below. In that case APIC EOI + * is required as otherwise the ISR entry becomes stale and lower + * priority interrupts are never going to be delivered after that. + * + * If the posted handler invoked the device interrupt handler then + * the EOI would be premature because it would acknowledge the + * posted vector. + */ + if (unlikely(!this_cpu_read(posted_msi_handler_active))) + apic_eoi(); +} + static __always_inline bool handle_pending_pir(unsigned long *pir, struct pt_regs *regs) { unsigned long pir_copy[NR_PIR_WORDS]; @@ -446,6 +466,8 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_posted_msi pid = this_cpu_ptr(&posted_msi_pi_desc); + /* Mark the handler active for intel_ack_posted_msi_irq() */ + this_cpu_write(posted_msi_handler_active, true); inc_irq_stat(posted_msi_notification_count); irq_enter(); @@ -474,6 +496,7 @@ DEFINE_IDTENTRY_SYSVEC(sysvec_posted_msi apic_eoi(); irq_exit(); + this_cpu_write(posted_msi_handler_active, false); set_irq_regs(old_regs); } #endif /* X86_POSTED_MSI */ --- a/drivers/iommu/intel/irq_remapping.c +++ b/drivers/iommu/intel/irq_remapping.c @@ -1303,17 +1303,17 @@ static struct irq_chip intel_ir_chip = { * irq_enter(); * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * handle_edge_irq() * irq_chip_ack_parent() - * irq_move_irq(); // No EOI + * intel_ack_posted_msi_irq(); // No EOI * handle_irq_event() * driver_handler() * apic_eoi() @@ -1322,7 +1322,7 @@ static struct irq_chip intel_ir_chip = { */ static struct irq_chip intel_ir_chip_post_msi = { .name = "INTEL-IR-POST", - .irq_ack = irq_move_irq, + .irq_ack = intel_ack_posted_msi_irq, .irq_set_affinity = intel_ir_set_affinity, .irq_compose_msi_msg = intel_ir_compose_msi_msg, .irq_set_vcpu_affinity = intel_ir_set_vcpu_affinity,

2 days, 12 hours

2
2
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror