PT-2026-151661
CVE-2026-99450
·
تاريخ النشر
2026-11-19
·
تاريخ التحديث
2026-11-19
الوصف من
الدرجة
N/A
غير متوفرلا تتوفر حاليًا تقييمات أو مقاييس للخطورة. سنحدّث المعلومات المقابلة في الصفحة فور توفرها.
In the Linux kernel, the following vulnerability has been resolved:
vfio/nvgrace-gpu: Use a core-only open device in fallback mode
When the ACPI memory properties are missing or the C2C link is down, the
device is exposed through nvgrace gpu pci core ops. That table opens
through nvgrace gpu open device() but closes through
vfio pci core close device(), so open and close are not symmetric.
nvgrace gpu open device() registers nvdev->usemem for poison handling
unconditionally. In fallback mode usemem.memlength is zero, and
register pfn address space() inserts a node covering [0, ULONG MAX].
Nothing removes it: a second open fails with -EBUSY, and once the device
is unbound the node outlives the nvdev that contains it. The next open
after a rebind reports:
BUG: KASAN: slab-use-after-free in interval tree iter first+0x209/0x310
Read of size 8 at addr ffff88800894aa30 by task nvgrace repro/203
Call Trace:
dump stack lvl+0x5d/0x80
print report+0x153/0x4c6
kasan report+0xe4/0x1a0
interval tree iter first+0x209/0x310
register pfn address space+0x97/0x100
nvgrace gpu open device+0x21c/0x730
vfio df open+0x213/0x520
vfio df ioctl bind iommufd+0x3e8/0x6e0
vfio device fops unl ioctl+0x782/0x1570
x64 sys ioctl+0x13c/0x1c0
do syscall 64+0x106/0x630
entry SYSCALL 64 after hwframe+0x77/0x7f
Allocated by task 188:
...
vfio alloc device+0x3a/0x500
nvgrace gpu probe+0xcb/0x660
local pci probe+0xcf/0x170
...
Freed by task 188:
...
kfree+0x20d/0x600
device release+0xcf/0x250
kobject put+0x152/0x2f0
pci device remove+0xb0/0x1c0
device release driver internal+0x38d/0x540
unbind store+0xde/0x100
...
Give the fallback table its own open device that only calls
vfio pci core enable() and vfio pci core finish enable(), the
counterpart of vfio pci core close device(). Fallback mode then never
touches the nvgrace memory regions, as Alex suggested.
This should be safe because the fallback table uses the vfio-pci core
read, write and mmap handlers, so nothing there uses the state
nvgrace gpu open device() sets up. The one visible difference is that a
fallback open no longer fails when BAR0 cannot be mapped, which matches
vfio-pci.
وجدت خطأً في الوصف؟ لديك ما تضيفه؟ لا تتردد في مراسلتنا 👾