linux-stable.git/include, branch v5.2.9 Linux kernel stable tree KVM: arm/arm64: Sync ICH_VMCR_EL2 back when about to block 2019-08-16T08:11:11+00:00 Marc Zyngier maz@kernel.org 2019-08-02T09:28:32+00:00 81ccda70dd52f445699d9eb49117df5549ab54ce commit 5eeaf10eec394b28fad2c58f1f5c3a5da0e87d1c upstream. Since commit commit 328e56647944 ("KVM: arm/arm64: vgic: Defer touching GICH_VMCR to vcpu_load/put"), we leave ICH_VMCR_EL2 (or its GICv2 equivalent) loaded as long as we can, only syncing it back when we're scheduled out. There is a small snag with that though: kvm_vgic_vcpu_pending_irq(), which is indirectly called from kvm_vcpu_check_block(), needs to evaluate the guest's view of ICC_PMR_EL1. At the point were we call kvm_vcpu_check_block(), the vcpu is still loaded, and whatever changes to PMR is not visible in memory until we do a vcpu_put(). Things go really south if the guest does the following: mov x0, #0 // or any small value masking interrupts msr ICC_PMR_EL1, x0 [vcpu preempted, then rescheduled, VMCR sampled] mov x0, #ff // allow all interrupts msr ICC_PMR_EL1, x0 wfi // traps to EL2, so samping of VMCR [interrupt arrives just after WFI] Here, the hypervisor's view of PMR is zero, while the guest has enabled its interrupts. kvm_vgic_vcpu_pending_irq() will then say that no interrupts are pending (despite an interrupt being received) and we'll block for no reason. If the guest doesn't have a periodic interrupt firing once it has blocked, it will stay there forever. To avoid this unfortuante situation, let's resync VMCR from kvm_arch_vcpu_blocking(), ensuring that a following kvm_vcpu_check_block() will observe the latest value of PMR. This has been found by booting an arm64 Linux guest with the pseudo NMI feature, and thus using interrupt priorities to mask interrupts instead of the usual PSTATE masking. Cc: stable@vger.kernel.org # 4.12 Fixes: 328e56647944 ("KVM: arm/arm64: vgic: Defer touching GICH_VMCR to vcpu_load/put") Signed-off-by: Marc Zyngier <maz@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 5eeaf10eec394b28fad2c58f1f5c3a5da0e87d1c upstream.

Since commit commit 328e56647944 ("KVM: arm/arm64: vgic: Defer
touching GICH_VMCR to vcpu_load/put"), we leave ICH_VMCR_EL2 (or
its GICv2 equivalent) loaded as long as we can, only syncing it
back when we're scheduled out.

There is a small snag with that though: kvm_vgic_vcpu_pending_irq(),
which is indirectly called from kvm_vcpu_check_block(), needs to
evaluate the guest's view of ICC_PMR_EL1. At the point were we
call kvm_vcpu_check_block(), the vcpu is still loaded, and whatever
changes to PMR is not visible in memory until we do a vcpu_put().

Things go really south if the guest does the following:

	mov x0, #0	// or any small value masking interrupts
	msr ICC_PMR_EL1, x0

	[vcpu preempted, then rescheduled, VMCR sampled]

	mov x0, #ff	// allow all interrupts
	msr ICC_PMR_EL1, x0
	wfi		// traps to EL2, so samping of VMCR

	[interrupt arrives just after WFI]

Here, the hypervisor's view of PMR is zero, while the guest has enabled
its interrupts. kvm_vgic_vcpu_pending_irq() will then say that no
interrupts are pending (despite an interrupt being received) and we'll
block for no reason. If the guest doesn't have a periodic interrupt
firing once it has blocked, it will stay there forever.

To avoid this unfortuante situation, let's resync VMCR from
kvm_arch_vcpu_blocking(), ensuring that a following kvm_vcpu_check_block()
will observe the latest value of PMR.

This has been found by booting an arm64 Linux guest with the pseudo NMI
feature, and thus using interrupt priorities to mask interrupts instead
of the usual PSTATE masking.

Cc: stable@vger.kernel.org # 4.12
Fixes: 328e56647944 ("KVM: arm/arm64: vgic: Defer touching GICH_VMCR to vcpu_load/put")
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
KVM: Fix leak vCPU's VMCS value into other pCPU 2019-08-16T08:11:11+00:00 Wanpeng Li wanpengli@tencent.com 2019-08-05T02:03:19+00:00 a3968fee8385eed72835f02f56bafae064c52f98 commit 17e433b54393a6269acbcb792da97791fe1592d8 upstream. After commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts), a five years old bug is exposed. Running ebizzy benchmark in three 80 vCPUs VMs on one 80 pCPUs Skylake server, a lot of rcu_sched stall warning splatting in the VMs after stress testing: INFO: rcu_sched detected stalls on CPUs/tasks: { 4 41 57 62 77} (detected by 15, t=60004 jiffies, g=899, c=898, q=15073) Call Trace: flush_tlb_mm_range+0x68/0x140 tlb_flush_mmu.part.75+0x37/0xe0 tlb_finish_mmu+0x55/0x60 zap_page_range+0x142/0x190 SyS_madvise+0x3cd/0x9c0 system_call_fastpath+0x1c/0x21 swait_active() sustains to be true before finish_swait() is called in kvm_vcpu_block(), voluntarily preempted vCPUs are taken into account by kvm_vcpu_on_spin() loop greatly increases the probability condition kvm_arch_vcpu_runnable(vcpu) is checked and can be true, when APICv is enabled the yield-candidate vCPU's VMCS RVI field leaks(by vmx_sync_pir_to_irr()) into spinning-on-a-taken-lock vCPU's current VMCS. This patch fixes it by checking conservatively a subset of events. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krčmář <rkrcmar@redhat.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Marc Zyngier <Marc.Zyngier@arm.com> Cc: stable@vger.kernel.org Fixes: 98f4a1467 (KVM: add kvm_arch_vcpu_runnable() test to kvm_vcpu_on_spin() loop) Signed-off-by: Wanpeng Li <wanpengli@tencent.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 17e433b54393a6269acbcb792da97791fe1592d8 upstream.

After commit d73eb57b80b (KVM: Boost vCPUs that are delivering interrupts), a
five years old bug is exposed. Running ebizzy benchmark in three 80 vCPUs VMs
on one 80 pCPUs Skylake server, a lot of rcu_sched stall warning splatting
in the VMs after stress testing:

 INFO: rcu_sched detected stalls on CPUs/tasks: { 4 41 57 62 77} (detected by 15, t=60004 jiffies, g=899, c=898, q=15073)
 Call Trace:
   flush_tlb_mm_range+0x68/0x140
   tlb_flush_mmu.part.75+0x37/0xe0
   tlb_finish_mmu+0x55/0x60
   zap_page_range+0x142/0x190
   SyS_madvise+0x3cd/0x9c0
   system_call_fastpath+0x1c/0x21

swait_active() sustains to be true before finish_swait() is called in
kvm_vcpu_block(), voluntarily preempted vCPUs are taken into account
by kvm_vcpu_on_spin() loop greatly increases the probability condition
kvm_arch_vcpu_runnable(vcpu) is checked and can be true, when APICv
is enabled the yield-candidate vCPU's VMCS RVI field leaks(by
vmx_sync_pir_to_irr()) into spinning-on-a-taken-lock vCPU's current
VMCS.

This patch fixes it by checking conservatively a subset of events.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Marc Zyngier <Marc.Zyngier@arm.com>
Cc: stable@vger.kernel.org
Fixes: 98f4a1467 (KVM: add kvm_arch_vcpu_runnable() test to kvm_vcpu_on_spin() loop)
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ALSA: compress: Fix regression on compressed capture streams 2019-08-16T08:11:04+00:00 Charles Keepax ckeepax@opensource.cirrus.com 2019-07-22T09:24:33+00:00 7989879611415f429bbb2b144bfe33a8daad0e64 [ Upstream commit 4475f8c4ab7b248991a60d9c02808dbb813d6be8 ] A previous fix to the stop handling on compressed capture streams causes some knock on issues. The previous fix updated snd_compr_drain_notify to set the state back to PREPARED for capture streams. This causes some issues however as the handling for snd_compr_poll differs between the two states and some user-space applications were relying on the poll failing after the stream had been stopped. To correct this regression whilst still fixing the original problem the patch was addressing, update the capture handling to skip the PREPARED state rather than skipping the SETUP state as it has done until now. Fixes: 4f2ab5e1d13d ("ALSA: compress: Fix stop handling on compressed capture streams") Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com> Acked-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 4475f8c4ab7b248991a60d9c02808dbb813d6be8 ]

A previous fix to the stop handling on compressed capture streams causes
some knock on issues. The previous fix updated snd_compr_drain_notify to
set the state back to PREPARED for capture streams. This causes some
issues however as the handling for snd_compr_poll differs between the
two states and some user-space applications were relying on the poll
failing after the stream had been stopped.

To correct this regression whilst still fixing the original problem the
patch was addressing, update the capture handling to skip the PREPARED
state rather than skipping the SETUP state as it has done until now.

Fixes: 4f2ab5e1d13d ("ALSA: compress: Fix stop handling on compressed capture streams")
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
nl80211: fix NL80211_HE_MAX_CAPABILITY_LEN 2019-08-16T08:11:02+00:00 John Crispin john@phrozen.org 2019-06-27T09:58:32+00:00 c60ab146fa9e99c52f2a5bbc2193241d4272e67c [ Upstream commit 5edaac063bbf1267260ad2a5b9bb803399343e58 ] NL80211_HE_MAX_CAPABILITY_LEN has changed between D2.0 and D4.0. It is now MAC (6) + PHY (11) + MCS (12) + PPE (25) = 54. Signed-off-by: John Crispin <john@phrozen.org> Link: https://lore.kernel.org/r/20190627095832.19445-1-john@phrozen.org Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 5edaac063bbf1267260ad2a5b9bb803399343e58 ]

NL80211_HE_MAX_CAPABILITY_LEN has changed between D2.0 and D4.0. It is now
MAC (6) + PHY (11) + MCS (12) + PPE (25) = 54.

Signed-off-by: John Crispin <john@phrozen.org>
Link: https://lore.kernel.org/r/20190627095832.19445-1-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
crypto: ccp - Add support for valid authsize values less than 16 2019-08-16T08:10:51+00:00 Gary R Hook gary.hook@amd.com 2019-07-30T16:05:24+00:00 9552214366b55878d9b0958f00eea7fc61ca50a2 commit 9f00baf74e4b6f79a3a3dfab44fb7bb2e797b551 upstream. AES GCM encryption allows for authsize values of 4, 8, and 12-16 bytes. Validate the requested authsize, and retain it to save in the request context. Fixes: 36cf515b9bbe2 ("crypto: ccp - Enable support for AES GCM on v5 CCPs") Cc: <stable@vger.kernel.org> Signed-off-by: Gary R Hook <gary.hook@amd.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 9f00baf74e4b6f79a3a3dfab44fb7bb2e797b551 upstream.

AES GCM encryption allows for authsize values of 4, 8, and 12-16 bytes.
Validate the requested authsize, and retain it to save in the request
context.

Fixes: 36cf515b9bbe2 ("crypto: ccp - Enable support for AES GCM on v5 CCPs")
Cc: <stable@vger.kernel.org>
Signed-off-by: Gary R Hook <gary.hook@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
compat_ioctl: pppoe: fix PPPOEIOCSFWD handling 2019-08-09T15:51:48+00:00 Arnd Bergmann arnd@arndb.de 2019-07-30T19:25:20+00:00 fb930c0055dff9744eed47c4a35514a65529f519 [ Upstream commit 055d88242a6046a1ceac3167290f054c72571cd9 ] Support for handling the PPPOEIOCSFWD ioctl in compat mode was added in linux-2.5.69 along with hundreds of other commands, but was always broken sincen only the structure is compatible, but the command number is not, due to the size being sizeof(size_t), or at first sizeof(sizeof((struct sockaddr_pppox)), which is different on 64-bit architectures. Guillaume Nault adds: And the implementation was broken until 2016 (see 29e73269aa4d ("pppoe: fix reference counting in PPPoE proxy")), and nobody ever noticed. I should probably have removed this ioctl entirely instead of fixing it. Clearly, it has never been used. Fix it by adding a compat_ioctl handler for all pppoe variants that translates the command number and then calls the regular ioctl function. All other ioctl commands handled by pppoe are compatible between 32-bit and 64-bit, and require compat_ptr() conversion. This should apply to all stable kernels. Acked-by: Guillaume Nault <g.nault@alphalink.fr> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
[ Upstream commit 055d88242a6046a1ceac3167290f054c72571cd9 ]

Support for handling the PPPOEIOCSFWD ioctl in compat mode was added in
linux-2.5.69 along with hundreds of other commands, but was always broken
sincen only the structure is compatible, but the command number is not,
due to the size being sizeof(size_t), or at first sizeof(sizeof((struct
sockaddr_pppox)), which is different on 64-bit architectures.

Guillaume Nault adds:

  And the implementation was broken until 2016 (see 29e73269aa4d ("pppoe:
  fix reference counting in PPPoE proxy")), and nobody ever noticed. I
  should probably have removed this ioctl entirely instead of fixing it.
  Clearly, it has never been used.

Fix it by adding a compat_ioctl handler for all pppoe variants that
translates the command number and then calls the regular ioctl function.

All other ioctl commands handled by pppoe are compatible between 32-bit
and 64-bit, and require compat_ptr() conversion.

This should apply to all stable kernels.

Acked-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/mlx5e: Prevent encap flow counter update async to user query 2019-08-09T15:51:46+00:00 Ariel Levkovich lariel@mellanox.com 2019-07-06T15:06:15+00:00 7c8eb11fd3729e6f3b6763eafc88098989b6280d [ Upstream commit 90bb769291161cf25a818d69cf608c181654473e ] This patch prevents a race between user invoked cached counters query and a neighbor last usage updater. The cached flow counter stats can be queried by calling "mlx5_fc_query_cached" which provides the number of bytes and packets that passed via this flow since the last time this counter was queried. It does so by reducting the last saved stats from the current, cached stats and then updating the last saved stats with the cached stats. It also provide the lastuse value for that flow. Since "mlx5e_tc_update_neigh_used_value" needs to retrieve the last usage time of encapsulation flows, it calls the flow counter query method periodically and async to user queries of the flow counter using cls_flower. This call is causing the driver to update the last reported bytes and packets from the cache and therefore, future user queries of the flow stats will return lower than expected number for bytes and packets since the last saved stats in the driver was updated async to the last saved stats in cls_flower. This causes wrong stats presentation of encapsulation flows to user. Since the neighbor usage updater only needs the lastuse stats from the cached counter, the fix is to use a dedicated lastuse query call that returns the lastuse value without synching between the cached stats and the last saved stats. Fixes: f6dfb4c3f216 ("net/mlx5e: Update neighbour 'used' state using HW flow rules counters") Signed-off-by: Ariel Levkovich <lariel@mellanox.com> Reviewed-by: Roi Dayan <roid@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
[ Upstream commit 90bb769291161cf25a818d69cf608c181654473e ]

This patch prevents a race between user invoked cached counters
query and a neighbor last usage updater.

The cached flow counter stats can be queried by calling
"mlx5_fc_query_cached" which provides the number of bytes and
packets that passed via this flow since the last time this counter
was queried.
It does so by reducting the last saved stats from the current, cached
stats and then updating the last saved stats with the cached stats.
It also provide the lastuse value for that flow.

Since "mlx5e_tc_update_neigh_used_value" needs to retrieve the
last usage time of encapsulation flows, it calls the flow counter
query method periodically and async to user queries of the flow counter
using cls_flower.
This call is causing the driver to update the last reported bytes and
packets from the cache and therefore, future user queries of the flow
stats will return lower than expected number for bytes and packets
since the last saved stats in the driver was updated async to the last
saved stats in cls_flower.

This causes wrong stats presentation of encapsulation flows to user.

Since the neighbor usage updater only needs the lastuse stats from the
cached counter, the fix is to use a dedicated lastuse query call that
returns the lastuse value without synching between the cached stats and
the last saved stats.

Fixes: f6dfb4c3f216 ("net/mlx5e: Update neighbour 'used' state using HW flow rules counters")
Signed-off-by: Ariel Levkovich <lariel@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/mlx5: Fix modify_cq_in alignment 2019-08-09T15:51:46+00:00 Edward Srouji edwards@mellanox.com 2019-07-23T07:12:55+00:00 901092bf0afb955bb667eaf9cf8dbb5fbb95c51a [ Upstream commit 7a32f2962c56d9d8a836b4469855caeee8766bd4 ] Fix modify_cq_in alignment to match the device specification. After this fix the 'cq_umem_valid' field will be in the right offset. Cc: <stable@vger.kernel.org> # 4.19 Fixes: bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits") Signed-off-by: Edward Srouji <edwards@mellanox.com> Reviewed-by: Yishai Hadas <yishaih@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
[ Upstream commit 7a32f2962c56d9d8a836b4469855caeee8766bd4 ]

Fix modify_cq_in alignment to match the device specification.
After this fix the 'cq_umem_valid' field will be in the right offset.

Cc: <stable@vger.kernel.org> # 4.19
Fixes: bd37197554eb ("net/mlx5: Update mlx5_ifc with DEVX UID bits")
Signed-off-by: Edward Srouji <edwards@mellanox.com>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
scsi: fcoe: Embed fc_rport_priv in fcoe_rport structure 2019-08-09T15:51:41+00:00 Hannes Reinecke hare@suse.de 2019-07-24T09:00:55+00:00 3d9de04646e7accf8522f5740690a365667c76b8 commit 023358b136d490ca91735ac6490db3741af5a8bd upstream. Gcc-9 complains for a memset across pointer boundaries, which happens as the code tries to allocate a flexible array on the stack. Turns out we cannot do this without relying on gcc-isms, so with this patch we'll embed the fc_rport_priv structure into fcoe_rport, can use the normal 'container_of' outcast, and will only have to do a memset over one structure. Signed-off-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 023358b136d490ca91735ac6490db3741af5a8bd upstream.

Gcc-9 complains for a memset across pointer boundaries, which happens as
the code tries to allocate a flexible array on the stack.  Turns out we
cannot do this without relying on gcc-isms, so with this patch we'll embed
the fc_rport_priv structure into fcoe_rport, can use the normal
'container_of' outcast, and will only have to do a memset over one
structure.

Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
RDMA/devices: Do not deadlock during client removal 2019-08-06T17:08:20+00:00 Jason Gunthorpe jgg@mellanox.com 2019-07-31T08:18:40+00:00 e2aa2bdc546fb65851e15c010bccd448108284f8 commit 621e55ff5b8e0ab5d1063f0eae0ef3960bef8f6e upstream. lockdep reports: WARNING: possible circular locking dependency detected modprobe/302 is trying to acquire lock: 0000000007c8919c ((wq_completion)ib_cm){+.+.}, at: flush_workqueue+0xdf/0x990 but task is already holding lock: 000000002d3d2ca9 (&device->client_data_rwsem){++++}, at: remove_client_context+0x79/0xd0 [ib_core] which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #2 (&device->client_data_rwsem){++++}: down_read+0x3f/0x160 ib_get_net_dev_by_params+0xd5/0x200 [ib_core] cma_ib_req_handler+0x5f6/0x2090 [rdma_cm] cm_process_work+0x29/0x110 [ib_cm] cm_req_handler+0x10f5/0x1c00 [ib_cm] cm_work_handler+0x54c/0x311d [ib_cm] process_one_work+0x4aa/0xa30 worker_thread+0x62/0x5b0 kthread+0x1ca/0x1f0 ret_from_fork+0x24/0x30 -> #1 ((work_completion)(&(&work->work)->work)){+.+.}: process_one_work+0x45f/0xa30 worker_thread+0x62/0x5b0 kthread+0x1ca/0x1f0 ret_from_fork+0x24/0x30 -> #0 ((wq_completion)ib_cm){+.+.}: lock_acquire+0xc8/0x1d0 flush_workqueue+0x102/0x990 cm_remove_one+0x30e/0x3c0 [ib_cm] remove_client_context+0x94/0xd0 [ib_core] disable_device+0x10a/0x1f0 [ib_core] __ib_unregister_device+0x5a/0xe0 [ib_core] ib_unregister_device+0x21/0x30 [ib_core] mlx5_ib_stage_ib_reg_cleanup+0x9/0x10 [mlx5_ib] __mlx5_ib_remove+0x3d/0x70 [mlx5_ib] mlx5_ib_remove+0x12e/0x140 [mlx5_ib] mlx5_remove_device+0x144/0x150 [mlx5_core] mlx5_unregister_interface+0x3f/0xf0 [mlx5_core] mlx5_ib_cleanup+0x10/0x3a [mlx5_ib] __x64_sys_delete_module+0x227/0x350 do_syscall_64+0xc3/0x6a4 entry_SYSCALL_64_after_hwframe+0x49/0xbe Which is due to the read side of the client_data_rwsem being obtained recursively through a work queue flush during cm client removal. The lock is being held across the remove in remove_client_context() so that the function is a fence, once it returns the client is removed. This is required so that the two callers do not proceed with destruction until the client completes removal. Instead of using client_data_rwsem use the existing device unregistration refcount and add a similar client unregistration (client->uses) refcount. This will fence the two unregistration paths without holding any locks. Cc: <stable@vger.kernel.org> Fixes: 921eab1143aa ("RDMA/devices: Re-organize device.c locking") Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Link: https://lore.kernel.org/r/20190731081841.32345-2-leon@kernel.org Signed-off-by: Doug Ledford <dledford@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 621e55ff5b8e0ab5d1063f0eae0ef3960bef8f6e upstream.

lockdep reports:

   WARNING: possible circular locking dependency detected

   modprobe/302 is trying to acquire lock:
   0000000007c8919c ((wq_completion)ib_cm){+.+.}, at: flush_workqueue+0xdf/0x990

   but task is already holding lock:
   000000002d3d2ca9 (&device->client_data_rwsem){++++}, at: remove_client_context+0x79/0xd0 [ib_core]

   which lock already depends on the new lock.

   the existing dependency chain (in reverse order) is:

   -> #2 (&device->client_data_rwsem){++++}:
          down_read+0x3f/0x160
          ib_get_net_dev_by_params+0xd5/0x200 [ib_core]
          cma_ib_req_handler+0x5f6/0x2090 [rdma_cm]
          cm_process_work+0x29/0x110 [ib_cm]
          cm_req_handler+0x10f5/0x1c00 [ib_cm]
          cm_work_handler+0x54c/0x311d [ib_cm]
          process_one_work+0x4aa/0xa30
          worker_thread+0x62/0x5b0
          kthread+0x1ca/0x1f0
          ret_from_fork+0x24/0x30

   -> #1 ((work_completion)(&(&work->work)->work)){+.+.}:
          process_one_work+0x45f/0xa30
          worker_thread+0x62/0x5b0
          kthread+0x1ca/0x1f0
          ret_from_fork+0x24/0x30

   -> #0 ((wq_completion)ib_cm){+.+.}:
          lock_acquire+0xc8/0x1d0
          flush_workqueue+0x102/0x990
          cm_remove_one+0x30e/0x3c0 [ib_cm]
          remove_client_context+0x94/0xd0 [ib_core]
          disable_device+0x10a/0x1f0 [ib_core]
          __ib_unregister_device+0x5a/0xe0 [ib_core]
          ib_unregister_device+0x21/0x30 [ib_core]
          mlx5_ib_stage_ib_reg_cleanup+0x9/0x10 [mlx5_ib]
          __mlx5_ib_remove+0x3d/0x70 [mlx5_ib]
          mlx5_ib_remove+0x12e/0x140 [mlx5_ib]
          mlx5_remove_device+0x144/0x150 [mlx5_core]
          mlx5_unregister_interface+0x3f/0xf0 [mlx5_core]
          mlx5_ib_cleanup+0x10/0x3a [mlx5_ib]
          __x64_sys_delete_module+0x227/0x350
          do_syscall_64+0xc3/0x6a4
          entry_SYSCALL_64_after_hwframe+0x49/0xbe

Which is due to the read side of the client_data_rwsem being obtained
recursively through a work queue flush during cm client removal.

The lock is being held across the remove in remove_client_context() so
that the function is a fence, once it returns the client is removed. This
is required so that the two callers do not proceed with destruction until
the client completes removal.

Instead of using client_data_rwsem use the existing device unregistration
refcount and add a similar client unregistration (client->uses) refcount.

This will fence the two unregistration paths without holding any locks.

Cc: <stable@vger.kernel.org>
Fixes: 921eab1143aa ("RDMA/devices: Re-organize device.c locking")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Link: https://lore.kernel.org/r/20190731081841.32345-2-leon@kernel.org
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>