{"api_version":"1","generated_at":"2026-06-28T04:41:34+00:00","cve":"CVE-2026-53145","urls":{"html":"https://cve.report/CVE-2026-53145","api":"https://cve.report/api/cve/CVE-2026-53145.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-53145","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-53145"},"summary":{"title":"drm/gem: Try to fix change_handle ioctl, attempt 4","description":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/gem: Try to fix change_handle ioctl, attempt 4\n\n[airlied: just added some comments on how to reenable]\nOn-list because the cat is out of the bag and we're clearly not good\nenough to figure this out in private. The story thus far:\n\n5e28b7b94408 (\"drm: Set old handle to NULL before prime swap in\nchange_handle\") tried to fix a race condition between the gem_close and\ngem_change_handle ioctls, but got a few things wrong:\n\n- There's a confusion with the local variable handle, which is actually\n  the new handle, and so the two-stage trick was actually applied to the\n  wrong idr slot. 7164d78559b0 (\"drm/gem: fix race between\n  change_handle and handle_delete\") tried to fix that by adding yet\n  another code block, but forgot to add the error handling. Which meant\n  we now have two paths, both kinda wrong.\n\n- dc366607c41c (\"drm: Replace old pointer to new idr\") tried to apply\n  another fix, but inconsistently, again because of the handle confusion\n  - this would be the right fix (kinda, somewhat, it's a mess) if we'd\n  do the two-stage approach for the new handle. Except that wasn't the\n  intent of the original fix.\n\nWe also didn't have an igt merged for the original ioctl, which is a big\nno-go. This was attempted to address off-list in the original bugfix,\nand amd QA people claimed the bug was fixed now. Very clearly that's not\nthe case. Here's my attempt to sort this out:\n\n- Rename the local variable to new_handle, the old aliasing with\n  args->handle is just too dangerously confusing.\n\n- Merge the gem obj lookup with the two-stage idr_replace so that we\n  avoid getting ourselves confused there.\n\n- This means we don't have a surplus temporary reference anymore, only\n  an inherited from the idr. A concurrent gem_close on the new_handle\n  could steal that. Fix that with the same two-stage approach\n  create_tail uses. This is a bit overkill as documented in the comment,\n  but I also don't trust my ability to understand this all correctly, so\n  go with the established pattern we have from other ioctls instead for\n  maximum paranoia.\n\n- Adjust error paths. I've tried to make the error and success paths\n  common, because they are identical except for which handle is removed\n  and on which we call idr_replace to (re)install the object again. But\n  that made things messier to read, so I've left it at the more verbose\n  version, which unfortunately hides the symmetry in the entire code\n  flow a bit.\n\n- While at it, also replace the 7 space indent with 1 tab.\n\nAnd finally, because I flat out don't trust my abilities here at all\nanymore:\n\n- Disable the ioctl until we have the igt situation and everything else\n  sorted out on-list and with full consensus.\n\nv2:\n\nSashiko noticed that I didn't handle the error path for idr_replace\ncorrectly, it must be checked with IS_ERR_OR_NULL like in\ngem_handle_delete. So yeah, definitely should just the existing paths\n1:1 because this is endless amounts of tricky.\n\nAlso add the Fixes: line for the original ioctl, I forgot that too.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-06-25 09:16:31","updated_at":"2026-06-25 09:16:31"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/1a4f03d22fb655e5f192244fb2c87d8066fcfca2","name":"https://git.kernel.org/stable/c/1a4f03d22fb655e5f192244fb2c87d8066fcfca2","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/c0639ede2f24ac224b2079cd35ecd5fd8ad4e3cd","name":"https://git.kernel.org/stable/c/c0639ede2f24ac224b2079cd35ecd5fd8ad4e3cd","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/1d9b93df7fc768228906e24220591ec1cddad391","name":"https://git.kernel.org/stable/c/1d9b93df7fc768228906e24220591ec1cddad391","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-53145","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-53145","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 672464dd53231509c9c771110798c56d4660e19e c0639ede2f24ac224b2079cd35ecd5fd8ad4e3cd git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 61bd96d3e5472c253f9c1ab77608f0c8aaa9d025 1d9b93df7fc768228906e24220591ec1cddad391 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 5e28b7b94408897e41c63477aabc9e1db439bc8c 1a4f03d22fb655e5f192244fb2c87d8066fcfca2 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 318b995cffcfcaa69a234d28123a3f4ae186a9df git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 38f12d0e10d83b66fa1466400d876a3a8da31542 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 0dfa42cfe4dbe114533480503934f43e33c1e83d git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected cde2c9257cbe8463b9dcf7b1075177b72b5fd938 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6.18.32 6.18.36 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7.0.9 7.0.13 semver","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":{"cve_year":"2026","cve_id":"53145","cve":"CVE-2026-53145","epss":"0.001730000","percentile":"0.070190000","score_date":"2026-06-27","updated_at":"2026-06-28 00:08:50"},"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["drivers/gpu/drm/drm_gem.c","drivers/gpu/drm/drm_ioctl.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"c0639ede2f24ac224b2079cd35ecd5fd8ad4e3cd","status":"affected","version":"672464dd53231509c9c771110798c56d4660e19e","versionType":"git"},{"lessThan":"1d9b93df7fc768228906e24220591ec1cddad391","status":"affected","version":"61bd96d3e5472c253f9c1ab77608f0c8aaa9d025","versionType":"git"},{"lessThan":"1a4f03d22fb655e5f192244fb2c87d8066fcfca2","status":"affected","version":"5e28b7b94408897e41c63477aabc9e1db439bc8c","versionType":"git"},{"status":"affected","version":"318b995cffcfcaa69a234d28123a3f4ae186a9df","versionType":"git"},{"status":"affected","version":"38f12d0e10d83b66fa1466400d876a3a8da31542","versionType":"git"},{"status":"affected","version":"0dfa42cfe4dbe114533480503934f43e33c1e83d","versionType":"git"},{"status":"affected","version":"cde2c9257cbe8463b9dcf7b1075177b72b5fd938","versionType":"git"}]},{"defaultStatus":"unaffected","product":"Linux","programFiles":["drivers/gpu/drm/drm_gem.c","drivers/gpu/drm/drm_ioctl.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"6.18.36","status":"affected","version":"6.18.32","versionType":"semver"},{"lessThan":"7.0.13","status":"affected","version":"7.0.9","versionType":"semver"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.18.36","versionStartIncluding":"6.18.32","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.0.13","versionStartIncluding":"7.0.9","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/gem: Try to fix change_handle ioctl, attempt 4\n\n[airlied: just added some comments on how to reenable]\nOn-list because the cat is out of the bag and we're clearly not good\nenough to figure this out in private. The story thus far:\n\n5e28b7b94408 (\"drm: Set old handle to NULL before prime swap in\nchange_handle\") tried to fix a race condition between the gem_close and\ngem_change_handle ioctls, but got a few things wrong:\n\n- There's a confusion with the local variable handle, which is actually\n  the new handle, and so the two-stage trick was actually applied to the\n  wrong idr slot. 7164d78559b0 (\"drm/gem: fix race between\n  change_handle and handle_delete\") tried to fix that by adding yet\n  another code block, but forgot to add the error handling. Which meant\n  we now have two paths, both kinda wrong.\n\n- dc366607c41c (\"drm: Replace old pointer to new idr\") tried to apply\n  another fix, but inconsistently, again because of the handle confusion\n  - this would be the right fix (kinda, somewhat, it's a mess) if we'd\n  do the two-stage approach for the new handle. Except that wasn't the\n  intent of the original fix.\n\nWe also didn't have an igt merged for the original ioctl, which is a big\nno-go. This was attempted to address off-list in the original bugfix,\nand amd QA people claimed the bug was fixed now. Very clearly that's not\nthe case. Here's my attempt to sort this out:\n\n- Rename the local variable to new_handle, the old aliasing with\n  args->handle is just too dangerously confusing.\n\n- Merge the gem obj lookup with the two-stage idr_replace so that we\n  avoid getting ourselves confused there.\n\n- This means we don't have a surplus temporary reference anymore, only\n  an inherited from the idr. A concurrent gem_close on the new_handle\n  could steal that. Fix that with the same two-stage approach\n  create_tail uses. This is a bit overkill as documented in the comment,\n  but I also don't trust my ability to understand this all correctly, so\n  go with the established pattern we have from other ioctls instead for\n  maximum paranoia.\n\n- Adjust error paths. I've tried to make the error and success paths\n  common, because they are identical except for which handle is removed\n  and on which we call idr_replace to (re)install the object again. But\n  that made things messier to read, so I've left it at the more verbose\n  version, which unfortunately hides the symmetry in the entire code\n  flow a bit.\n\n- While at it, also replace the 7 space indent with 1 tab.\n\nAnd finally, because I flat out don't trust my abilities here at all\nanymore:\n\n- Disable the ioctl until we have the igt situation and everything else\n  sorted out on-list and with full consensus.\n\nv2:\n\nSashiko noticed that I didn't handle the error path for idr_replace\ncorrectly, it must be checked with IS_ERR_OR_NULL like in\ngem_handle_delete. So yeah, definitely should just the existing paths\n1:1 because this is endless amounts of tricky.\n\nAlso add the Fixes: line for the original ioctl, I forgot that too."}],"providerMetadata":{"dateUpdated":"2026-06-25T08:38:32.228Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/c0639ede2f24ac224b2079cd35ecd5fd8ad4e3cd"},{"url":"https://git.kernel.org/stable/c/1d9b93df7fc768228906e24220591ec1cddad391"},{"url":"https://git.kernel.org/stable/c/1a4f03d22fb655e5f192244fb2c87d8066fcfca2"}],"title":"drm/gem: Try to fix change_handle ioctl, attempt 4","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-53145","datePublished":"2026-06-25T08:38:32.228Z","dateReserved":"2026-06-09T07:44:35.387Z","dateUpdated":"2026-06-25T08:38:32.228Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-06-25 09:16:31","lastModifiedDate":"2026-06-25 09:16:31","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"53145","Ordinal":"1","Title":"drm/gem: Try to fix change_handle ioctl, attempt 4","CVE":"CVE-2026-53145","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"53145","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/gem: Try to fix change_handle ioctl, attempt 4\n\n[airlied: just added some comments on how to reenable]\nOn-list because the cat is out of the bag and we're clearly not good\nenough to figure this out in private. The story thus far:\n\n5e28b7b94408 (\"drm: Set old handle to NULL before prime swap in\nchange_handle\") tried to fix a race condition between the gem_close and\ngem_change_handle ioctls, but got a few things wrong:\n\n- There's a confusion with the local variable handle, which is actually\n  the new handle, and so the two-stage trick was actually applied to the\n  wrong idr slot. 7164d78559b0 (\"drm/gem: fix race between\n  change_handle and handle_delete\") tried to fix that by adding yet\n  another code block, but forgot to add the error handling. Which meant\n  we now have two paths, both kinda wrong.\n\n- dc366607c41c (\"drm: Replace old pointer to new idr\") tried to apply\n  another fix, but inconsistently, again because of the handle confusion\n  - this would be the right fix (kinda, somewhat, it's a mess) if we'd\n  do the two-stage approach for the new handle. Except that wasn't the\n  intent of the original fix.\n\nWe also didn't have an igt merged for the original ioctl, which is a big\nno-go. This was attempted to address off-list in the original bugfix,\nand amd QA people claimed the bug was fixed now. Very clearly that's not\nthe case. Here's my attempt to sort this out:\n\n- Rename the local variable to new_handle, the old aliasing with\n  args->handle is just too dangerously confusing.\n\n- Merge the gem obj lookup with the two-stage idr_replace so that we\n  avoid getting ourselves confused there.\n\n- This means we don't have a surplus temporary reference anymore, only\n  an inherited from the idr. A concurrent gem_close on the new_handle\n  could steal that. Fix that with the same two-stage approach\n  create_tail uses. This is a bit overkill as documented in the comment,\n  but I also don't trust my ability to understand this all correctly, so\n  go with the established pattern we have from other ioctls instead for\n  maximum paranoia.\n\n- Adjust error paths. I've tried to make the error and success paths\n  common, because they are identical except for which handle is removed\n  and on which we call idr_replace to (re)install the object again. But\n  that made things messier to read, so I've left it at the more verbose\n  version, which unfortunately hides the symmetry in the entire code\n  flow a bit.\n\n- While at it, also replace the 7 space indent with 1 tab.\n\nAnd finally, because I flat out don't trust my abilities here at all\nanymore:\n\n- Disable the ioctl until we have the igt situation and everything else\n  sorted out on-list and with full consensus.\n\nv2:\n\nSashiko noticed that I didn't handle the error path for idr_replace\ncorrectly, it must be checked with IS_ERR_OR_NULL like in\ngem_handle_delete. So yeah, definitely should just the existing paths\n1:1 because this is endless amounts of tricky.\n\nAlso add the Fixes: line for the original ioctl, I forgot that too.","Type":"Description","Title":"drm/gem: Try to fix change_handle ioctl, attempt 4"}]}}}