CVE-2026-68266 (GCVE-0-2026-68266)
Vulnerability from cvelistv5
Published
2026-08-10 12:01
Modified
2026-08-17 05:01
Severity ?
VLAI Severity ?
EPSS score ?
Summary
In the Linux kernel, the following vulnerability has been resolved:
drm/xe: Hold a dma-buf reference for imported BOs
An imported dma-buf BO is created as a ttm_bo_type_sg BO whose
reservation object is the exporter's dma_buf->resv. The importer,
however, only takes a dma-buf reference after a successful
dma_buf_dynamic_attach(). Until then nothing keeps the exporter alive,
so if the exporter is freed while the BO still references its resv, a
later access to that resv is a use-after-free:
Oops: general protection fault, probably for non-canonical address
0x6b6b6b6b6b6b6b9c
Workqueue: ttm ttm_bo_delayed_delete [ttm]
RIP: 0010:mutex_can_spin_on_owner+0x3f/0xc0
This can be reached on two paths:
- dma_buf_dynamic_attach() fails, or
- ttm_bo_init_reserved() fails during BO creation.
In both cases the BO already has bo->base.resv pointing at the exporter
resv, and sg BOs are always torn down via ttm_bo_delayed_delete(), which
locks bo->base.resv asynchronously - potentially after the exporter has
been freed.
Take the dma-buf reference in xe_bo_init_locked(), before
ttm_bo_init_reserved(), so it also covers a creation failure there, and
release it in xe_ttm_bo_destroy(). The reference is held for the whole
BO lifetime, keeping the shared resv alive on every path.
v2:
- Reworked the fix to avoid creating the imported sg BO before
dma_buf_dynamic_attach() succeeds.
- Attach with importer_priv == NULL and make invalidate_mappings ignore
incomplete imports.
v3:
- Dropped the xe-side reordering approach since importer_priv must be
valid when dma_buf_dynamic_attach() publishes the attachment.
- Per Christian's suggestion on the v1 thread, keyed the check on
import_attach rather than removing the sg guard entirely.
- Fixes both xe and amdgpu in a single TTM patch.
v4:
- Moved import_attach check to after dma_resv_copy_fences() so fences
are copied before returning for successful imports (Thomas).
- Removed exporter-alive claim from commit message (Thomas).
v5:
- Add drm/xe patch to keep imported sg BOs off the LRU before attach
succeeds; the TTM fix alone is not sufficient for xe if the BO is
already LRU-visible. (Thomas)
v4 patch:
https://patchwork.freedesktop.org/patch/736663/?series=169129&rev=2
- Patch 1 (drm/ttm) carries Christian's Reviewed-by from v4.
v6:
- Reworked the fix based on Thomas' suggestion. Instead of the TTM resv
individualization (v1-v5) plus the xe off-LRU/placement handling (v5),
just hold a dma-buf reference for the imported BO lifetime so the
shared resv can never be freed while the BO still references it.
Single xe patch, no TTM change. (Thomas)
- Take the reference in xe_bo_init_locked() before ttm_bo_init_reserved()
so a TTM creation failure is covered too (Thomas).
- Dropped the v5 series (drm/ttm + drm/xe off-LRU); the off-LRU approach
also regressed in CI BAT via ttm_bo_pipeline_gutting() creating a ghost
BO that outlived the exporter.
Link to v5: https://patchwork.freedesktop.org/series/169984/
v7:
- Move changelog above --- so it stays in the commit message.
- Reorder changelog entries oldest-to-newest. (Thomas)
(cherry picked from commit 3516f3fae6be35642f8f06f8a218da6425c0306a)
References
| URL | Tags | |
|---|---|---|
Impacted products
{
"containers": {
"cna": {
"affected": [
{
"defaultStatus": "unaffected",
"product": "Linux",
"programFiles": [
"drivers/gpu/drm/xe/xe_bo.c",
"drivers/gpu/drm/xe/xe_bo.h",
"drivers/gpu/drm/xe/xe_bo_types.h",
"drivers/gpu/drm/xe/xe_dma_buf.c"
],
"repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
"vendor": "Linux",
"versions": [
{
"lessThan": "c22d65d62b3318e237c0e5b1177d90ab83d9fe06",
"status": "affected",
"version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
"versionType": "git"
},
{
"lessThan": "c1954c66662de477a8f4309335b775f7b07bd28b",
"status": "affected",
"version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
"versionType": "git"
},
{
"lessThan": "ba8c4cbb31c6f81fa5b12d6e28f1f706040aff48",
"status": "affected",
"version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
"versionType": "git"
},
{
"lessThan": "62775525a27c3b0d56382e08ba81ee2d322058b6",
"status": "affected",
"version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
"versionType": "git"
}
]
},
{
"defaultStatus": "affected",
"product": "Linux",
"programFiles": [
"drivers/gpu/drm/xe/xe_bo.c",
"drivers/gpu/drm/xe/xe_bo.h",
"drivers/gpu/drm/xe/xe_bo_types.h",
"drivers/gpu/drm/xe/xe_dma_buf.c"
],
"repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
"vendor": "Linux",
"versions": [
{
"status": "affected",
"version": "6.8"
},
{
"lessThan": "6.8",
"status": "unaffected",
"version": "0",
"versionType": "semver"
},
{
"lessThanOrEqual": "6.12.*",
"status": "unaffected",
"version": "6.12.103",
"versionType": "semver"
},
{
"lessThanOrEqual": "6.18.*",
"status": "unaffected",
"version": "6.18.42",
"versionType": "semver"
},
{
"lessThanOrEqual": "7.1.*",
"status": "unaffected",
"version": "7.1.6",
"versionType": "semver"
},
{
"lessThanOrEqual": "*",
"status": "unaffected",
"version": "7.2",
"versionType": "original_commit_for_fix"
}
]
}
],
"cpeApplicability": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
"versionEndExcluding": "6.12.103",
"versionStartIncluding": "6.8",
"vulnerable": true
},
{
"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
"versionEndExcluding": "6.18.42",
"versionStartIncluding": "6.8",
"vulnerable": true
},
{
"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
"versionEndExcluding": "7.1.6",
"versionStartIncluding": "6.8",
"vulnerable": true
},
{
"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
"versionEndExcluding": "7.2",
"versionStartIncluding": "6.8",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/xe: Hold a dma-buf reference for imported BOs\n\nAn imported dma-buf BO is created as a ttm_bo_type_sg BO whose\nreservation object is the exporter\u0027s dma_buf-\u003eresv. The importer,\nhowever, only takes a dma-buf reference after a successful\ndma_buf_dynamic_attach(). Until then nothing keeps the exporter alive,\nso if the exporter is freed while the BO still references its resv, a\nlater access to that resv is a use-after-free:\n\n Oops: general protection fault, probably for non-canonical address\n 0x6b6b6b6b6b6b6b9c\n Workqueue: ttm ttm_bo_delayed_delete [ttm]\n RIP: 0010:mutex_can_spin_on_owner+0x3f/0xc0\n\nThis can be reached on two paths:\n\n - dma_buf_dynamic_attach() fails, or\n - ttm_bo_init_reserved() fails during BO creation.\n\nIn both cases the BO already has bo-\u003ebase.resv pointing at the exporter\nresv, and sg BOs are always torn down via ttm_bo_delayed_delete(), which\nlocks bo-\u003ebase.resv asynchronously - potentially after the exporter has\nbeen freed.\n\nTake the dma-buf reference in xe_bo_init_locked(), before\nttm_bo_init_reserved(), so it also covers a creation failure there, and\nrelease it in xe_ttm_bo_destroy(). The reference is held for the whole\nBO lifetime, keeping the shared resv alive on every path.\n\nv2:\n - Reworked the fix to avoid creating the imported sg BO before\n dma_buf_dynamic_attach() succeeds.\n - Attach with importer_priv == NULL and make invalidate_mappings ignore\n incomplete imports.\n\nv3:\n - Dropped the xe-side reordering approach since importer_priv must be\n valid when dma_buf_dynamic_attach() publishes the attachment.\n - Per Christian\u0027s suggestion on the v1 thread, keyed the check on\n import_attach rather than removing the sg guard entirely.\n - Fixes both xe and amdgpu in a single TTM patch.\n\nv4:\n - Moved import_attach check to after dma_resv_copy_fences() so fences\n are copied before returning for successful imports (Thomas).\n - Removed exporter-alive claim from commit message (Thomas).\n\nv5:\n - Add drm/xe patch to keep imported sg BOs off the LRU before attach\n succeeds; the TTM fix alone is not sufficient for xe if the BO is\n already LRU-visible. (Thomas)\n v4 patch:\n https://patchwork.freedesktop.org/patch/736663/?series=169129\u0026rev=2\n - Patch 1 (drm/ttm) carries Christian\u0027s Reviewed-by from v4.\n\nv6:\n - Reworked the fix based on Thomas\u0027 suggestion. Instead of the TTM resv\n individualization (v1-v5) plus the xe off-LRU/placement handling (v5),\n just hold a dma-buf reference for the imported BO lifetime so the\n shared resv can never be freed while the BO still references it.\n Single xe patch, no TTM change. (Thomas)\n - Take the reference in xe_bo_init_locked() before ttm_bo_init_reserved()\n so a TTM creation failure is covered too (Thomas).\n - Dropped the v5 series (drm/ttm + drm/xe off-LRU); the off-LRU approach\n also regressed in CI BAT via ttm_bo_pipeline_gutting() creating a ghost\n BO that outlived the exporter.\n Link to v5: https://patchwork.freedesktop.org/series/169984/\n\nv7:\n - Move changelog above --- so it stays in the commit message.\n - Reorder changelog entries oldest-to-newest. (Thomas)\n\n(cherry picked from commit 3516f3fae6be35642f8f06f8a218da6425c0306a)"
}
],
"metrics": [
{
"cvssV3_1": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
"version": "3.1"
},
"scenarios": [
{
"lang": "en",
"value": "AV:L - The bug is reached through DRM_IOCTL_PRIME_FD_TO_HANDLE on an Intel xe DRM render node (/dev/dri/renderD*), which requires local access to the device file; there is no remote or network-facing path into xe_gem_prime_import().\nAC:L - The attacker fully controls both halves: they pick an exporter whose attach fails deterministically (drm_gem_map_attach() returns -ENOSYS, xe_dma_buf_attach() returns -EOPNOTSUPP for non-p2p cross-device imports) or induce the -ENOMEM paths, then simply close the dma-buf fd while the sg BO teardown is still queued on the TTM workqueue. No condition outside their control is needed.\nPR:L - DRM_IOCTL_PRIME_FD_TO_HANDLE is marked DRM_RENDER_ALLOW, so any unprivileged local user holding the normally world/group-accessible render node (any GUI user on a distro, any app on Android/ChromeOS) can perform the import; no capabilities are checked.\nUI:N - The entire sequence \u2014 export a dma-buf, import it into xe, close the fd \u2014 is performed by the attacker\u0027s own process; no victim action or cooperation is required.\nS:U - The use-after-free occurs on kernel heap memory within the same kernel security authority; there is no crossing of a VM, IOMMU, or sandbox boundary.\nC:H - The freed dma_buf/dma_resv is a slab object the attacker can reclaim with sprayed content before ttm_bo_delayed_delete() dereferences it, so the stale bo-\u003ebase.resv can be aimed at attacker-chosen kernel addresses and the resulting ww_mutex/fence walks leaked back, giving arbitrary kernel memory disclosure.\nI:H - dma_resv_lock()/dma_resv_unlock() and the subsequent ttm_bo_cleanup_memtype_use() perform writes and list manipulation through the dangling pointer into reclaimed heap memory, so the UAF yields a controllable write primitive usable for control-flow hijacking and privilege escalation.\nA:H - Even unexploited, the flaw reliably produces a kernel oops \u2014 the reported general protection fault on poisoned memory in mutex_can_spin_on_owner() from the ttm workqueue \u2014 which is a fatal fault in kernel worker context and takes down the GPU stack or the whole machine."
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-08-17T05:01:59.066Z",
"orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
"shortName": "Linux"
},
"references": [
{
"url": "https://git.kernel.org/stable/c/c22d65d62b3318e237c0e5b1177d90ab83d9fe06"
},
{
"url": "https://git.kernel.org/stable/c/c1954c66662de477a8f4309335b775f7b07bd28b"
},
{
"url": "https://git.kernel.org/stable/c/ba8c4cbb31c6f81fa5b12d6e28f1f706040aff48"
},
{
"url": "https://git.kernel.org/stable/c/62775525a27c3b0d56382e08ba81ee2d322058b6"
}
],
"title": "drm/xe: Hold a dma-buf reference for imported BOs",
"x_generator": {
"engine": "bippy-1.2.0"
}
}
},
"cveMetadata": {
"assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
"assignerShortName": "Linux",
"cveId": "CVE-2026-68266",
"datePublished": "2026-08-10T12:01:41.280Z",
"dateReserved": "2026-07-30T09:28:09.378Z",
"dateUpdated": "2026-08-17T05:01:59.066Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
Loading…
Loading…
Sightings
| Author | Source | Type | Date |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or seen somewhere by the user.
- Confirmed: The vulnerability is confirmed from an analyst perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: This vulnerability was exploited and seen by the user reporting the sighting.
- Patched: This vulnerability was successfully patched by the user reporting the sighting.
- Not exploited: This vulnerability was not exploited or seen by the user reporting the sighting.
- Not confirmed: The user expresses doubt about the veracity of the vulnerability.
- Not patched: This vulnerability was not successfully patched by the user reporting the sighting.
Loading…
Loading…