memory: Skip assertion in memory_region_unregister_iommu_notifier
v2 cover memory: Skip assertion in memory_region_unregister_iommu_notifier
I am able to hit this assertion when a Red Hat 7 guest virtio_net device raises an "Invalidation" of all the TLB entries. This happens in the guest's startup if 'intel_iommu=on' argument is passed to the guest kernel and right IOMMU/ATS devices are declared in qemu's command line.
Command line:/home/qemu/x86_64-softmmu/qemu-system-x86_64 -name \
guest=rhel7-test,debug-threads=on -machine \
pc-q35-5.1,accel=kvm,usb=off,dump-guest-core=off,kernel_irqchip=split \
-cpu \
Broadwell,vme=on,ss=on,vmx=on,f16c=on,rdrand=on,hypervisor=on,arat=on,tsc-adjust=on,umip=on,arch-capabilities=on,xsaveopt=on,pdpe1gb=on,abm=on,skip-l1dfl-vmentry=on,rtm=on,hle=on \
-m 8096 -realtime mlock=off -smp 2,sockets=2,cores=1,threads=1 -uuid \
d022ecbf-679e-4755-87ce-eb87fc5bbc5d -display none -no-user-config \
-nodefaults -rtc base=utc,driftfix=slew -global \
kvm-pit.lost_tick_policy=delay -no-hpet -no-shutdown -global \
ICH9-LPC.disable_s3=1 -global ICH9-LPC.disable_s4=1 -boot strict=on \
-device intel-iommu,intremap=on,device-iotlb=on -device \
pcie-root-port,port=0x8,chassis=1,id=pci.1,bus=pcie.0,multifunction=on,addr=0x1 \
-device \
pcie-root-port,port=0x9,chassis=2,id=pci.2,bus=pcie.0,addr=0x1.0x1 \
-device \
pcie-root-port,port=0xa,chassis=3,id=pci.3,bus=pcie.0,addr=0x1.0x2 \
-device \
pcie-root-port,port=0xb,chassis=4,id=pci.4,bus=pcie.0,addr=0x1.0x3 \
-device \
pcie-root-port,port=0xc,chassis=5,id=pci.5,bus=pcie.0,addr=0x1.0x4 \
-device \
pcie-root-port,port=0xd,chassis=6,id=pci.6,bus=pcie.0,addr=0x1.0x5 \
-device \
pcie-root-port,port=0xe,chassis=7,id=pci.7,bus=pcie.0,addr=0x1.0x6 \
-device qemu-xhci,p2=15,p3=15,id=usb,bus=pci.2,addr=0x0 -device \
virtio-serial-pci,id=virtio-serial0,bus=pci.3,addr=0x0 -drive \
file=/home/virtio-test2.qcow2,format=qcow2,if=none,id=drive-virtio-disk0 \
-device \
virtio-blk-pci,scsi=off,bus=pci.4,addr=0x0,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1 \
-netdev tap,id=hostnet0,vhost=on,vhostforce=on -device \
virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:0d:1d:f2,bus=pci.1,addr=0x0,iommu_platform=on,ats=on \
-device virtio-balloon-pci,id=balloon0,bus=pci.5,addr=0x0 -object \
rng-random,id=objrng0,filename=/dev/urandom -device \
virtio-rng-pci,rng=objrng0,id=rng0,bus=pci.6,addr=0x0 -s -msg \
timestamp=on
Full backtrace:
#0 0x00007ffff521370f in raise () at /lib64/libc.so.6
#1 0x00007ffff51fdb25 in abort () at /lib64/libc.so.6
#2 0x00007ffff51fd9f9 in _nl_load_domain.cold.0 () at /lib64/libc.so.6
#3 0x00007ffff520bcc6 in .annobin_assert.c_end () at /lib64/libc.so.6
#4 0x0000555555888171 in memory_region_notify_one (notifier=0x7ffde0487fa8,
entry=0x7ffde5dfe200)
at /home/qemu/memory.c:1918
#5 0x0000555555888247 in memory_region_notify_iommu (iommu_mr=0x555556f6c0b0,
iommu_idx=0, entry=...)
at /home/qemu/memory.c:1941
#6 0x0000555555951c8d in vtd_process_device_iotlb_desc (s=0x555557609000,
inv_desc=0x7ffde5dfe2d0)
at /home/qemu/hw/i386/intel_iommu.c:2468
#7 0x0000555555951e6a in vtd_process_inv_desc (s=0x555557609000)
at /home/qemu/hw/i386/intel_iommu.c:2531
#8 0x0000555555951fa5 in vtd_fetch_inv_desc (s=0x555557609000)
at /home/qemu/hw/i386/intel_iommu.c:2563
#9 0x00005555559520e5 in vtd_handle_iqt_write (s=0x555557609000)
at /home/qemu/hw/i386/intel_iommu.c:2590
#10 0x0000555555952b45 in vtd_mem_write (opaque=0x555557609000, addr=136,
val=2688, size=4)
at /home/qemu/hw/i386/intel_iommu.c:2837
#11 0x0000555555883e17 in memory_region_write_accessor (mr=0x555557609330,
addr=136,
value=0x7ffde5dfe478,
size=4,
shift=0,
mask=4294967295,
attrs=...)
at /home/qemu/memory.c:483
#12 0x000055555588401d in access_with_adjusted_size (addr=136,
value=0x7ffde5dfe478,
size=4,
access_size_min=4,
access_size_max=8,
access_fn=0x555555883d38 <memory_region_write_accessor>,
mr=0x555557609330,
attrs=...)
at /home/qemu/memory.c:544
#13 0x0000555555886f37 in memory_region_dispatch_write (mr=0x555557609330,
addr=136,
data=2688,
op=MO_32,
attrs=...)
at /home/qemu/memory.c:1476
#14 0x0000555555827a03 in flatview_write_continue (fv=0x7ffdd8503150,
addr=4275634312,
attrs=...,
ptr=0x7ffff7ff0028,
len=4,
addr1=136,
l=4,
mr=0x555557609330)
at /home/qemu/exec.c:3146
#15 0x0000555555827b48 in flatview_write (fv=0x7ffdd8503150,
addr=4275634312,
attrs=...,
buf=0x7ffff7ff0028,
len=4)
at /home/qemu/exec.c:3186
#16 0x0000555555827e9d in address_space_write (
as=0x5555567ca640 <address_space_memory>,
addr=4275634312,
attrs=...,
buf=0x7ffff7ff0028,
len=4)
at /home/qemu/exec.c:3277
#17 0x0000555555827f0a in address_space_rw (
as=0x5555567ca640 <address_space_memory>,
addr=4275634312,
attrs=...,
buf=0x7ffff7ff0028,
len=4,
is_write=true)
at /home/qemu/exec.c:3287
#18 0x000055555589b633 in kvm_cpu_exec (cpu=0x555556b65640)
at /home/qemu/accel/kvm/kvm-all.c:2511
#19 0x0000555555876ba8 in qemu_kvm_cpu_thread_fn (arg=0x555556b65640)
at /home/qemu/cpus.c:1284
#20 0x0000555555dafff1 in qemu_thread_start (args=0x555556b8c3b0)
at util/qemu-thread-posix.c:521
#21 0x00007ffff55a62de in start_thread () at /lib64/libpthread.so.0
#22 0x00007ffff52d7e83 in clone () at /lib64/libc.so.6(gdb) frame 4
#4 0x0000555555888171 in memory_region_notify_one
(notifier=0x7ffde0487fa8, entry=0x7ffde5dfe200)
at /home/qemu/memory.c:19181918 assert(entry->iova >= notifier->start && entry_end <=notifier->end);
(gdb) p *entry
$1 = {target_as = 0x555556f6c050, iova = 0, translated_addr = 0, addr_mask = 18446744073709551615, perm = IOMMU_NONE} --
Tested with vhost-net and qemu driver, host<->guest communication.
v2: * Delete underflow assertion
* Tested again rebased over ("3e407488349 Merge remote-tracking branch 'remotes/rth/tags/pull-mb-20201014' into staging")
v1: * IOMMU_NOTIFIER_ALL now includes IOMMU_NOTIFIER_DEVIOTLB_EVENTS
also. VFIO IOMMU notifier will register for all events (as before of the patching)
* Cosmetic changes, like:
- Expand commit messages
- Better naming and checks
- Fix alignment issues
- Avoid an already present casting from void *
at https://lists.gnu.org/archive/html/qemu-devel/2020-09/msg01505.html
RFC v8: Fix use of "tmp" notification in memory.c:memory_region_notify_iommu_one
v7: Add IOMMUTLBNotification, and move introduced "type" from
IOMMUTLBEntry to the former.
v6: Introduce "type" field for IOMMUTLBEntry. Fill in all uses.
Update tests reports with more fine-tuning (CPU, RPS/XPS tunning).
v5: Skip regular IOTLB notifications in dev_iotlb notifiers
v4: Rename IOMMU_NOTIFIER_IOTLB -> IOMMU_NOTIFIER_DEVIOTLB.
Make vhost-net notifier just IOMMU_NOTIFIER_DEVIOTLB, not IOMMU_NOTIFIER_UNMAP
v3: Skip the assertion in case notifier is a IOTLB one, since they can manage
arbitrary ranges. Using a flag in the notifier for now, as Peter suggested.
v2: Actually delete assertion instead of just commenting out using C99
Eugenio Pérez (5):
memory: Rename memory_region_notify_one to
memory_region_notify_iommu_one
memory: Add IOMMUTLBEvent
memory: Add IOMMU_NOTIFIER_DEVIOTLB_UNMAP IOMMUTLBNotificationType
intel_iommu: Skip page walking on device iotlb invalidations
memory: Skip bad range assertion if notifier is DEVIOTLB_UNMAP type
include/exec/memory.h | 40 ++++++++++--------- hw/arm/smmu-common.c | 13 +++--- hw/arm/smmuv3.c | 13 +++--- hw/i386/intel_iommu.c | 92 +++++++++++++++++++++++++------------------ hw/misc/tz-mpc.c | 32 ++++++++------- hw/ppc/spapr_iommu.c | 15 +++---- hw/virtio/vhost.c | 2 +- softmmu/memory.c | 30 ++++++++------ 8 files changed, 134 insertions(+), 103 deletions(-)
Signature · 1 line
2.18.1
PATCH v2 1/5 memory: Rename memory_region_notify_one to memory_region_notify_iommu_one
Previous name didn't reflect the iommu operation.
--- include/exec/memory.h | 6 +++--- hw/arm/smmu-common.c | 2 +- hw/arm/smmuv3.c | 2 +- hw/i386/intel_iommu.c | 4 ++-- softmmu/memory.c | 6 +++--- 5 files changed, 10 insertions(+), 10 deletions(-)
include/exec/memory.h +3 -3
@@ -221,7 +221,7 @@ enum IOMMUMemoryRegionAttr {* The IOMMU implementation must use the IOMMU notifier infrastructure * to report whenever mappings are changed, by calling * memory_region_notify_iommu() (or, if necessary, by calling -* memory_region_notify_one() for each registered notifier). +* memory_region_notify_iommu_one() for each registered notifier). * * Conceptually an IOMMU provides a mapping from input address * to an output TLB entry. If the IOMMU is aware of memory transaction
@@ -1300,7 +1300,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr, IOMMUTLBEntry entry);/** -* memory_region_notify_one: notify a change in an IOMMU translation +* memory_region_notify_iommu_one: notify a change in an IOMMU translation * entry to a single notifier * * This works just like memory_region_notify_iommu(), but it only
@@ -1311,7 +1311,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr,* replaces all old entries for the same virtual I/O address range. * Deleted entries have .@perm == 0. */ -void memory_region_notify_one (IOMMUNotifier *notifier, +void memory_region_notify_iommu_one (IOMMUNotifier *notifier,IOMMUTLBEntry *entry);/**
hw/arm/smmu-common.c +1 -1
@@ -472,7 +472,7 @@ static void smmu_unmap_notifier_range(IOMMUNotifier *n) entry.perm =IOMMU_NONE ; entry.addr_mask = n->end - n->start; -memory_region_notify_one (n, &entry); +memory_region_notify_iommu_one (n, &entry); }/* Unmap all notifiers attached to @mr */
hw/arm/smmuv3.c +1 -1
@@ -827,7 +827,7 @@ static void smmuv3_notify_iova(IOMMUMemoryRegion *mr, entry.addr_mask = num_pages * (1 << granule) -1 ; entry.perm =IOMMU_NONE ; -memory_region_notify_one (n, &entry); +memory_region_notify_iommu_one (n, &entry); }/* invalidate an asid/iova range tuple in all mr's */
hw/i386/intel_iommu.c +2 -2
@@ -3497,7 +3497,7 @@ static void vtd_address_space_unmap(VTDAddressSpace *as, IOMMUNotifier *n)/* This field is meaningless for unmap */ entry.translated_addr =0 ; -memory_region_notify_one (n, &entry); +memory_region_notify_iommu_one (n, &entry); start += mask; remain -= mask;
@@ -3535,7 +3535,7 @@ static void vtd_address_space_refresh_all(IntelIOMMUState *s)static int vtd_replay_hook (IOMMUTLBEntry *entry,void *private) { -memory_region_notify_one ((IOMMUNotifier *)private, entry); +memory_region_notify_iommu_one ((IOMMUNotifier *)private, entry);return 0 ; }
softmmu/memory.c +3 -3
@@ -1905,8 +1905,8 @@ void memory_region_unregister_iommu_notifier(MemoryRegion *mr,memory_region_update_iommu_notify_flags (iommu_mr,NULL ); } -void memory_region_notify_one (IOMMUNotifier *notifier, -IOMMUTLBEntry *entry) +void memory_region_notify_iommu_one (IOMMUNotifier *notifier, +IOMMUTLBEntry *entry) {IOMMUNotifierFlag request_flags;hwaddr entry_end = entry->iova + entry->addr_mask;
@@ -1942,7 +1942,7 @@ void memory_region_notify_iommu(IOMMUMemoryRegion *iommu_mr, IOMMU_NOTIFIER_FOREACH(iommu_notifier, iommu_mr) {if (iommu_notifier->iommu_idx == iommu_idx) { -memory_region_notify_one (iommu_notifier, &entry); +memory_region_notify_iommu_one (iommu_notifier, &entry); } } }
Signature · 1 line
2.18.1
Quoted 181 lines · Eugenio Pérez
I am able to hit this assertion when a Red Hat 7 guest
virtio_netdevice raises an "Invalidation" of all the TLB entries. This happens in the guest's startup if 'intel_iommu=on' argument is passed to the guest kernel and right IOMMU/ATS devices are declared in qemu's command line.Command line:
/home/qemu/x86_64-softmmu/qemu-system-x86_64-name \
guest=rhel7-test,debug-threads=on -machine \
pc-q35-5.1,accel=kvm,usb=off,dump-guest-core=off,kernel_irqchip=split \
-cpu \
Broadwell,vme=on,ss=on,vmx=on,f16c=on,rdrand=on,hypervisor=on,arat=on,tsc-adjust=on,umip=on,arch-capabilities=on,xsaveopt=on,pdpe1gb=on,abm=on,skip-l1dfl-vmentry=on,rtm=on,hle=on \
-m 8096 -realtime mlock=off -smp 2,sockets=2,cores=1,threads=1 -uuid \
d022ecbf-679e-4755-87ce-eb87fc5bbc5d -display none -no-user-config \
-nodefaults -rtc base=utc,driftfix=slew -global \
kvm-pit.lost_tick_policy=delay -no-hpet -no-shutdown -global \
ICH9-LPC.disable_s3=1 -global ICH9-LPC.disable_s4=1 -boot strict=on \
-device intel-iommu,intremap=on,device-iotlb=on -device \
pcie-root-port,port=0x8,chassis=1,id=pci.1,bus=pcie.0,multifunction=on,addr=0x1 \
-device \
pcie-root-port,port=0x9,chassis=2,id=pci.2,bus=pcie.0,addr=0x1.0x1 \
-device \
pcie-root-port,port=0xa,chassis=3,id=pci.3,bus=pcie.0,addr=0x1.0x2 \
-device \
pcie-root-port,port=0xb,chassis=4,id=pci.4,bus=pcie.0,addr=0x1.0x3 \
-device \
pcie-root-port,port=0xc,chassis=5,id=pci.5,bus=pcie.0,addr=0x1.0x4 \
-device \
pcie-root-port,port=0xd,chassis=6,id=pci.6,bus=pcie.0,addr=0x1.0x5 \
-device \
pcie-root-port,port=0xe,chassis=7,id=pci.7,bus=pcie.0,addr=0x1.0x6 \
-device qemu-xhci,p2=15,p3=15,id=usb,bus=pci.2,addr=0x0 -device \
virtio-serial-pci,id=virtio-serial0,bus=pci.3,addr=0x0 -drive \
file=/home/virtio-test2.qcow2,format=qcow2,if=none,id=drive-virtio-disk0 \
-device \
virtio-blk-pci,scsi=off,bus=pci.4,addr=0x0,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1 \
-netdev tap,id=hostnet0,vhost=on,vhostforce=on -device \
virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:0d:1d:f2,bus=pci.1,addr=0x0,iommu_platform=on,ats=on \
-device virtio-balloon-pci,id=balloon0,bus=pci.5,addr=0x0 -object \
rng-random,id=objrng0,filename=/dev/urandom -device \
virtio-rng-pci,rng=objrng0,id=rng0,bus=pci.6,addr=0x0 -s -msg \
timestamp=onFull backtrace:
#0 0x00007ffff521370f in raise () at /lib64/libc.so.6 #1 0x00007ffff51fdb25 in abort () at /lib64/libc.so.6 #2 0x00007ffff51fd9f9 in _nl_load_domain.cold.0 () at /lib64/libc.so.6 #3 0x00007ffff520bcc6 in .annobin_assert.c_end () at /lib64/libc.so.6 #4 0x0000555555888171 in memory_region_notify_one (notifier=0x7ffde0487fa8, entry=0x7ffde5dfe200) at /home/qemu/memory.c:1918 #5 0x0000555555888247 in memory_region_notify_iommu (iommu_mr=0x555556f6c0b0, iommu_idx=0, entry=...) at /home/qemu/memory.c:1941 #6 0x0000555555951c8d in vtd_process_device_iotlb_desc (s=0x555557609000, inv_desc=0x7ffde5dfe2d0) at /home/qemu/hw/i386/intel_iommu.c:2468 #7 0x0000555555951e6a in vtd_process_inv_desc (s=0x555557609000) at /home/qemu/hw/i386/intel_iommu.c:2531 #8 0x0000555555951fa5 in vtd_fetch_inv_desc (s=0x555557609000) at /home/qemu/hw/i386/intel_iommu.c:2563 #9 0x00005555559520e5 in vtd_handle_iqt_write (s=0x555557609000) at /home/qemu/hw/i386/intel_iommu.c:2590 #10 0x0000555555952b45 in vtd_mem_write (opaque=0x555557609000, addr=136, val=2688, size=4) at /home/qemu/hw/i386/intel_iommu.c:2837 #11 0x0000555555883e17 in memory_region_write_accessor (mr=0x555557609330, addr=136, value=0x7ffde5dfe478, size=4, shift=0, mask=4294967295, attrs=...) at /home/qemu/memory.c:483 #12 0x000055555588401d in access_with_adjusted_size (addr=136, value=0x7ffde5dfe478, size=4, access_size_min=4, access_size_max=8, access_fn=0x555555883d38 <memory_region_write_accessor>, mr=0x555557609330, attrs=...) at /home/qemu/memory.c:544 #13 0x0000555555886f37 in memory_region_dispatch_write (mr=0x555557609330, addr=136, data=2688, op=MO_32, attrs=...) at /home/qemu/memory.c:1476 #14 0x0000555555827a03 in flatview_write_continue (fv=0x7ffdd8503150, addr=4275634312, attrs=..., ptr=0x7ffff7ff0028, len=4, addr1=136, l=4, mr=0x555557609330) at /home/qemu/exec.c:3146 #15 0x0000555555827b48 in flatview_write (fv=0x7ffdd8503150, addr=4275634312, attrs=..., buf=0x7ffff7ff0028, len=4) at /home/qemu/exec.c:3186 #16 0x0000555555827e9d in address_space_write ( as=0x5555567ca640 <address_space_memory>, addr=4275634312, attrs=..., buf=0x7ffff7ff0028, len=4) at /home/qemu/exec.c:3277 #17 0x0000555555827f0a in address_space_rw ( as=0x5555567ca640 <address_space_memory>, addr=4275634312, attrs=..., buf=0x7ffff7ff0028, len=4, is_write=true) at /home/qemu/exec.c:3287 #18 0x000055555589b633 in kvm_cpu_exec (cpu=0x555556b65640) at /home/qemu/accel/kvm/kvm-all.c:2511 #19 0x0000555555876ba8 in qemu_kvm_cpu_thread_fn (arg=0x555556b65640) at /home/qemu/cpus.c:1284 #20 0x0000555555dafff1 in qemu_thread_start (args=0x555556b8c3b0) at util/qemu-thread-posix.c:521 #21 0x00007ffff55a62de in start_thread () at /lib64/libpthread.so.0 #22 0x00007ffff52d7e83 in clone () at /lib64/libc.so.6(gdb) frame 4
#40x0000555555888171inmemory_region_notify_one(notifier=0x7ffde0487fa8, entry=0x7ffde5dfe200) at /home/qemu/memory.c:19181918 assert(entry->iova >=
notifier->start&&entry_end<=notifier->end);
(gdb) p *entry
$1 = {target_as=0x555556f6c050, iova = 0,translated_addr= 0,addr_mask= 18446744073709551615, perm =IOMMU_NONE} --Tested with vhost-net and qemu driver, host<->guest communication.
v2: * Delete underflow assertion
* Tested again rebased over ("3e407488349Merge remote-tracking branch 'remotes/rth/tags/pull-mb-20201014' into staging")v1: *
IOMMU_NOTIFIER_ALLnow includesIOMMU_NOTIFIER_DEVIOTLB_EVENTS
also. VFIO IOMMU notifier will register for all events (as before of the patching)
* Cosmetic changes, like:
- Expand commit messages
- Better naming and checks
- Fix alignment issues
- Avoid an already present casting fromvoid *
at https://lists.gnu.org/archive/html/qemu-devel/2020-09/msg01505.htmlRFC v8: Fix use of "tmp" notification in memory.c:memory_region_notify_iommu_one
v7: Add IOMMUTLBNotification, and move introduced "type" from
IOMMUTLBEntry to the former.v6: Introduce "type" field for IOMMUTLBEntry. Fill in all uses.
Update tests reports with more fine-tuning (CPU, RPS/XPS tunning).v5: Skip regular IOTLB notifications in
dev_iotlbnotifiersv4: Rename
IOMMU_NOTIFIER_IOTLB->IOMMU_NOTIFIER_DEVIOTLB.
Make vhost-net notifier justIOMMU_NOTIFIER_DEVIOTLB, notIOMMU_NOTIFIER_UNMAPv3: Skip the assertion in case notifier is a IOTLB one, since they can manage
arbitrary ranges. Using a flag in the notifier for now, as Peter suggested.v2: Actually delete assertion instead of just commenting out using C99
Eugenio Pérez (5):
memory: Renamememory_region_notify_oneto
memory_region_notify_iommu_one
memory: Add IOMMUTLBEvent
memory: AddIOMMU_NOTIFIER_DEVIOTLB_UNMAPIOMMUTLBNotificationType
intel_iommu: Skip page walking on device iotlb invalidations
memory: Skip bad range assertion if notifier isDEVIOTLB_UNMAPtype
Quoted 9 lines
include/exec/memory.h | 40 ++++++++++--------- hw/arm/smmu-common.c | 13 +++--- hw/arm/smmuv3.c | 13 +++--- hw/i386/intel_iommu.c | 92 +++++++++++++++++++++++++------------------ hw/misc/tz-mpc.c | 32 ++++++++------- hw/ppc/spapr_iommu.c | 15 +++---- hw/virtio/vhost.c | 2 +- softmmu/memory.c | 30 ++++++++------ 8 files changed, 134 insertions(+), 103 deletions(-)
Thanks