In the Linux kernel, the following vulnerability has been resolved:
drm/xe: Limit num_syncs to prevent oversized allocations
The exec and vmbind ioctl allow userspace to specify an arbitrary numsyncs value. Without bounds checking, a very large num_syncs can force an excessively large allocation, leading to kernel warnings from the page allocator as below.
Introduce DRMXEMAX_SYNCS (set to 1024) and reject any request exceeding this limit.
" ------------[ cut here ]------------ WARNING: CPU: 0 PID: 1217 at mm/pagealloc.c:5124 allocfrozenpagesnoprof+0x2f8/0x2180 mm/pagealloc.c:5124 ... Call Trace: <TASK> allocpagesmpol+0xe4/0x330 mm/mempolicy.c:2416 _kmalloclargenode+0xd8/0x110 mm/slub.c:4317 _kmalloclargenodenoprof+0x18/0xe0 mm/slub.c:4348 _dokmallocnode mm/slub.c:4364 [inline] _kmallocnoprof+0x3d4/0x4b0 mm/slub.c:4388 kmallocnoprof include/linux/slab.h:909 [inline] kmallocarraynoprof include/linux/slab.h:948 [inline] xeexecioctl+0xa47/0x1e70 drivers/gpu/drm/xe/xeexec.c:158 drmioctlkernel+0x1f1/0x3e0 drivers/gpu/drm/drmioctl.c:797 drmioctl+0x5e7/0xc50 drivers/gpu/drm/drmioctl.c:894 xedrmioctl+0x10b/0x170 drivers/gpu/drm/xe/xedevice.c:224 vfsioctl fs/ioctl.c:51 [inline] _dosysioctl fs/ioctl.c:598 [inline] _sesysioctl fs/ioctl.c:584 [inline] _x64sysioctl+0x18b/0x210 fs/ioctl.c:584 dosyscallx64 arch/x86/entry/syscall64.c:63 [inline] dosyscall64+0xbb/0x380 arch/x86/entry/syscall64.c:94 entrySYSCALL64afterhwframe+0x77/0x7f ... "
v2: Add "Reported-by" and Cc stable kernels. v3: Change XEMAXSYNCS from 64 to 1024. (Matt & Ashutosh) v4: s/XEMAXSYNCS/DRMXEMAX_SYNCS/ (Matt) v5: Do the check at the top of the exec func. (Matt)
(cherry picked from commit b07bac9bd708ec468cd1b8a5fe70ae2ac9b0a11c)
{
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/68xxx/CVE-2025-68802.json",
"cna_assigner": "Linux"
}