From f1940d4e9cbe6208e7e77e433c587af108152a17 Mon Sep 17 00:00:00 2001 From: Vitaly Kuznetsov Date: Tue, 31 Aug 2021 16:39:16 +0200 Subject: [PATCH 1/3] Drivers: hv: vmbus: Fix kernel crash upon unbinding a device from uio_hv_generic driver The following crash happens when a never-used device is unbound from uio_hv_generic driver: kernel BUG at mm/slub.c:321! invalid opcode: 0000 [#1] SMP PTI CPU: 0 PID: 4001 Comm: bash Kdump: loaded Tainted: G X --------- --- 5.14.0-0.rc2.23.el9.x86_64 #1 Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090008 12/07/2018 RIP: 0010:__slab_free+0x1d5/0x3d0 ... Call Trace: ? pick_next_task_fair+0x18e/0x3b0 ? __cond_resched+0x16/0x40 ? vunmap_pmd_range.isra.0+0x154/0x1c0 ? __vunmap+0x22d/0x290 ? hv_ringbuffer_cleanup+0x36/0x40 [hv_vmbus] kfree+0x331/0x380 ? hv_uio_remove+0x43/0x60 [uio_hv_generic] hv_ringbuffer_cleanup+0x36/0x40 [hv_vmbus] vmbus_free_ring+0x21/0x60 [hv_vmbus] hv_uio_remove+0x4f/0x60 [uio_hv_generic] vmbus_remove+0x23/0x30 [hv_vmbus] __device_release_driver+0x17a/0x230 device_driver_detach+0x3c/0xa0 unbind_store+0x113/0x130 ... The problem appears to be that we free 'ring_info->pkt_buffer' twice: first, when the device is unbound from in-kernel driver (netvsc in this case) and second from hv_uio_remove(). Normally, ring buffer is supposed to be re-initialized from hv_uio_open() but this happens when UIO device is being opened and this is not guaranteed to happen. Generally, it is OK to call hv_ringbuffer_cleanup() twice for the same channel (which is being handed over between in-kernel drivers and UIO) even if we didn't call hv_ringbuffer_init() in between. We, however, need to avoid kfree() call for an already freed pointer. Fixes: adae1e931acd ("Drivers: hv: vmbus: Copy packets sent by Hyper-V out of the ring buffer") Signed-off-by: Vitaly Kuznetsov Reviewed-by: Andrea Parri Reviewed-by: Michael Kelley Link: https://lore.kernel.org/r/20210831143916.144983-1-vkuznets@redhat.com Signed-off-by: Wei Liu --- drivers/hv/ring_buffer.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/hv/ring_buffer.c b/drivers/hv/ring_buffer.c index 2aee356840a2..314015d9e912 100644 --- a/drivers/hv/ring_buffer.c +++ b/drivers/hv/ring_buffer.c @@ -245,6 +245,7 @@ void hv_ringbuffer_cleanup(struct hv_ring_buffer_info *ring_info) mutex_unlock(&ring_info->ring_buffer_mutex); kfree(ring_info->pkt_buffer); + ring_info->pkt_buffer = NULL; ring_info->pkt_buffer_size = 0; } From 7ad9bb9d0f357dcab5eb9a0f28d1c8983c48434c Mon Sep 17 00:00:00 2001 From: Wei Liu Date: Fri, 10 Sep 2021 18:57:13 +0000 Subject: [PATCH 2/3] asm-generic/hyperv: provide cpumask_to_vpset_noself This is a new variant which removes `self' cpu from the vpset. It will be used in Hyper-V enlightened IPI code. Signed-off-by: Wei Liu Reviewed-by: Michael Kelley Link: https://lore.kernel.org/r/20210910185714.299411-2-wei.liu@kernel.org --- include/asm-generic/mshyperv.h | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) diff --git a/include/asm-generic/mshyperv.h b/include/asm-generic/mshyperv.h index 9a000ba2bb75..9a134806f1d5 100644 --- a/include/asm-generic/mshyperv.h +++ b/include/asm-generic/mshyperv.h @@ -184,10 +184,12 @@ static inline int hv_cpu_number_to_vp_number(int cpu_number) return hv_vp_index[cpu_number]; } -static inline int cpumask_to_vpset(struct hv_vpset *vpset, - const struct cpumask *cpus) +static inline int __cpumask_to_vpset(struct hv_vpset *vpset, + const struct cpumask *cpus, + bool exclude_self) { int cpu, vcpu, vcpu_bank, vcpu_offset, nr_bank = 1; + int this_cpu = smp_processor_id(); /* valid_bank_mask can represent up to 64 banks */ if (hv_max_vp_index / 64 >= 64) @@ -205,6 +207,8 @@ static inline int cpumask_to_vpset(struct hv_vpset *vpset, * Some banks may end up being empty but this is acceptable. */ for_each_cpu(cpu, cpus) { + if (exclude_self && cpu == this_cpu) + continue; vcpu = hv_cpu_number_to_vp_number(cpu); if (vcpu == VP_INVAL) return -1; @@ -219,6 +223,19 @@ static inline int cpumask_to_vpset(struct hv_vpset *vpset, return nr_bank; } +static inline int cpumask_to_vpset(struct hv_vpset *vpset, + const struct cpumask *cpus) +{ + return __cpumask_to_vpset(vpset, cpus, false); +} + +static inline int cpumask_to_vpset_noself(struct hv_vpset *vpset, + const struct cpumask *cpus) +{ + WARN_ON_ONCE(preemptible()); + return __cpumask_to_vpset(vpset, cpus, true); +} + void hyperv_report_panic(struct pt_regs *regs, long err, bool in_die); bool hv_is_hyperv_initialized(void); bool hv_is_hibernation_supported(void); From dfb5c1e12c28e35e4d4e5bc8022b0e9d585b89a7 Mon Sep 17 00:00:00 2001 From: Wei Liu Date: Fri, 10 Sep 2021 18:57:14 +0000 Subject: [PATCH 3/3] x86/hyperv: remove on-stack cpumask from hv_send_ipi_mask_allbutself It is not a good practice to allocate a cpumask on stack, given it may consume up to 1 kilobytes of stack space if the kernel is configured to have 8192 cpus. The internal helper functions __send_ipi_mask{,_ex} need to loop over the provided mask anyway, so it is not too difficult to skip `self' there. We can thus do away with the on-stack cpumask in hv_send_ipi_mask_allbutself. Adjust call sites of __send_ipi_mask as needed. Reported-by: Linus Torvalds Suggested-by: Michael Kelley Suggested-by: Linus Torvalds Fixes: 68bb7bfb7985d ("X86/Hyper-V: Enable IPI enlightenments") Signed-off-by: Wei Liu Reviewed-by: Michael Kelley Link: https://lore.kernel.org/r/20210910185714.299411-3-wei.liu@kernel.org --- arch/x86/hyperv/hv_apic.c | 43 +++++++++++++++++++++++---------------- 1 file changed, 26 insertions(+), 17 deletions(-) diff --git a/arch/x86/hyperv/hv_apic.c b/arch/x86/hyperv/hv_apic.c index 90e682a92820..32a1ad356c18 100644 --- a/arch/x86/hyperv/hv_apic.c +++ b/arch/x86/hyperv/hv_apic.c @@ -99,7 +99,8 @@ static void hv_apic_eoi_write(u32 reg, u32 val) /* * IPI implementation on Hyper-V. */ -static bool __send_ipi_mask_ex(const struct cpumask *mask, int vector) +static bool __send_ipi_mask_ex(const struct cpumask *mask, int vector, + bool exclude_self) { struct hv_send_ipi_ex **arg; struct hv_send_ipi_ex *ipi_arg; @@ -123,7 +124,10 @@ static bool __send_ipi_mask_ex(const struct cpumask *mask, int vector) if (!cpumask_equal(mask, cpu_present_mask)) { ipi_arg->vp_set.format = HV_GENERIC_SET_SPARSE_4K; - nr_bank = cpumask_to_vpset(&(ipi_arg->vp_set), mask); + if (exclude_self) + nr_bank = cpumask_to_vpset_noself(&(ipi_arg->vp_set), mask); + else + nr_bank = cpumask_to_vpset(&(ipi_arg->vp_set), mask); } if (nr_bank < 0) goto ipi_mask_ex_done; @@ -138,15 +142,25 @@ ipi_mask_ex_done: return hv_result_success(status); } -static bool __send_ipi_mask(const struct cpumask *mask, int vector) +static bool __send_ipi_mask(const struct cpumask *mask, int vector, + bool exclude_self) { - int cur_cpu, vcpu; + int cur_cpu, vcpu, this_cpu = smp_processor_id(); struct hv_send_ipi ipi_arg; u64 status; + unsigned int weight; trace_hyperv_send_ipi_mask(mask, vector); - if (cpumask_empty(mask)) + weight = cpumask_weight(mask); + + /* + * Do nothing if + * 1. the mask is empty + * 2. the mask only contains self when exclude_self is true + */ + if (weight == 0 || + (exclude_self && weight == 1 && cpumask_test_cpu(this_cpu, mask))) return true; if (!hv_hypercall_pg) @@ -172,6 +186,8 @@ static bool __send_ipi_mask(const struct cpumask *mask, int vector) ipi_arg.cpu_mask = 0; for_each_cpu(cur_cpu, mask) { + if (exclude_self && cur_cpu == this_cpu) + continue; vcpu = hv_cpu_number_to_vp_number(cur_cpu); if (vcpu == VP_INVAL) return false; @@ -191,7 +207,7 @@ static bool __send_ipi_mask(const struct cpumask *mask, int vector) return hv_result_success(status); do_ex_hypercall: - return __send_ipi_mask_ex(mask, vector); + return __send_ipi_mask_ex(mask, vector, exclude_self); } static bool __send_ipi_one(int cpu, int vector) @@ -208,7 +224,7 @@ static bool __send_ipi_one(int cpu, int vector) return false; if (vp >= 64) - return __send_ipi_mask_ex(cpumask_of(cpu), vector); + return __send_ipi_mask_ex(cpumask_of(cpu), vector, false); status = hv_do_fast_hypercall16(HVCALL_SEND_IPI, vector, BIT_ULL(vp)); return hv_result_success(status); @@ -222,20 +238,13 @@ static void hv_send_ipi(int cpu, int vector) static void hv_send_ipi_mask(const struct cpumask *mask, int vector) { - if (!__send_ipi_mask(mask, vector)) + if (!__send_ipi_mask(mask, vector, false)) orig_apic.send_IPI_mask(mask, vector); } static void hv_send_ipi_mask_allbutself(const struct cpumask *mask, int vector) { - unsigned int this_cpu = smp_processor_id(); - struct cpumask new_mask; - const struct cpumask *local_mask; - - cpumask_copy(&new_mask, mask); - cpumask_clear_cpu(this_cpu, &new_mask); - local_mask = &new_mask; - if (!__send_ipi_mask(local_mask, vector)) + if (!__send_ipi_mask(mask, vector, true)) orig_apic.send_IPI_mask_allbutself(mask, vector); } @@ -246,7 +255,7 @@ static void hv_send_ipi_allbutself(int vector) static void hv_send_ipi_all(int vector) { - if (!__send_ipi_mask(cpu_online_mask, vector)) + if (!__send_ipi_mask(cpu_online_mask, vector, false)) orig_apic.send_IPI_all(vector); }