CVE-2021-47227

In the Linux kernel, the following vulnerability has been resolved:

x86/fpu: Prevent state corruption in __fpu__restore_sig()

The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer.

Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust.

Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms.

Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path.

[ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved:

x86/fpu: Prevent state corruption in __fpu__restore_sig()

The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer.

Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust.

Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms.

Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path.

[ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved:

x86/fpu: Prevent state corruption in __fpu__restore_sig()

The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer.

Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust.

Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms.

Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path.

[ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved:

x86/fpu: Prevent state corruption in __fpu__restore_sig()

The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer.

Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust.

Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms.

Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path.

[ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved:

x86/fpu: Prevent state corruption in __fpu__restore_sig()

The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer.

Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust.

Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms.

Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path.

[ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved: x86/fpu: Prevent state corruption in __fpu__restore_sig() The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer. Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust. Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms. Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path. [ Changelog polished by Dave Hansen ]

In the Linux kernel, the following vulnerability has been resolved: x86/fpu: Prevent state corruption in __fpu__restore_sig() The non-compacted slowpath uses __copy_from_user() and copies the entire user buffer into the kernel buffer, verbatim. This means that the kernel buffer may now contain entirely invalid state on which XRSTOR will #GP. validate_user_xstate_header() can detect some of that corruption, but that leaves the onus on callers to clear the buffer. Prior to XSAVES support, it was possible just to reinitialize the buffer, completely, but with supervisor states that is not longer possible as the buffer clearing code split got it backwards. Fixing that is possible but not corrupting the state in the first place is more robust. Avoid corruption of the kernel XSAVE buffer by using copy_user_to_xstate() which validates the XSAVE header contents before copying the actual states to the kernel. copy_user_to_xstate() was previously only called for compacted-format kernel buffers, but it works for both compacted and non-compacted forms. Using it for the non-compacted form is slower because of multiple __copy_from_user() operations, but that cost is less important than robust code in an already slow path. [ Changelog polished by Dave Hansen ]

The SUSE Linux Enterprise 15 SP3 RT kernel was updated to receive various security bugfixes.

The following security bugs were fixed:

• CVE-2024-26921: Preserve kabi for sk_buff (bsc#1223138).
• CVE-2021-47378: Destroy cm id before destroy qp to avoid use after free (bsc#1225201).
• CVE-2021-47496: Fix flipped sign in tls_err_abort() calls (bsc#1225354)
• CVE-2021-47402: Protect fl_walk() with rcu (bsc#1225301)
• CVE-2022-48673: kABI workarounds for struct smc_link (bsc#1223934).
• CVE-2023-52871: Handle a second device without data corruption (bsc#1225534)
• CVE-2024-26828: Fixed underflow in parse_server_interfaces() (bsc#1223084).
• CVE-2021-47497: Fixed shift-out-of-bound (UBSAN) with byte size cells (bsc#1225355).
• CVE-2021-47500: Fixed trigger reference couting (bsc#1225360).
• CVE-2024-27413: Fixed incorrect allocation size (bsc#1224438).
• CVE-2021-47383: Fiedx out-of-bound vmalloc access in imageblit (bsc#1225208).
• CVE-2021-47511: Fixed negative period/buffer sizes (bsc#1225411).
• CVE-2023-52840: Fix use after free in rmi_unregister_function() (bsc#1224928).
• CVE-2021-47261: Fix initializing CQ fragments buffer (bsc#1224954)
• CVE-2021-47254: Fix use-after-free in gfs2_glock_shrink_scan (bsc#1224888).
• CVE-2024-27398: Fixed use-after-free bugs caused by sco_sock_timeout (bsc#1224174).
• CVE-2024-26929: Fixed double free of fcport (bsc#1223715).
• CVE-2024-26930: Fixed double free of the ha->vp_map pointer (bsc#1223626).
• CVE-2023-52655: Check packet for fixup for true limit (bsc#1217169).
• CVE-2023-4244: Fixed a use-after-free in the nf_tables component, which could be exploited to achieve local privilege escalation (bsc#1215420).
• CVE-2023-1829: Fixed a use-after-free vulnerability in the control index filter (tcindex) (bsc#1210335).
• CVE-2023-52686: Fix a null pointer in opal_event_init() (bsc#1065729).

The following non-security bugs were fixed:

• af_unix: annote lockless accesses to unix_tot_inflight & gc_in_progress (bsc#1223384).
• af_unix: Do not use atomic ops for unix_sk(sk)->inflight (bsc#1223384).
• af_unix: Replace BUG_ON() with WARN_ON_ONCE() (bsc#1223384).
• btrfs: do not start relocation until in progress drops are done (bsc#1222251).
• btrfs: do not start relocation until in progress drops are done (bsc#1222251).
• cifs: add missing spinlock around tcon refcount (bsc#1213476).
• cifs: avoid dup prefix path in dfs_get_automount_devname() (bsc#1213476).
• cifs: avoid race conditions with parallel reconnects (bsc#1213476).
• cifs: avoid re-lookups in dfs_cache_find() (bsc#1213476).
• cifs: avoid use of global locks for high contention data (bsc#1213476).
• cifs: check only tcon status on tcon related functions (bsc#1213476).
• cifs: do all necessary checks for credits within or before locking (bsc#1213476).
• cifs: do not block in dfs_cache_noreq_update_tgthint() (bsc#1213476).
• cifs: do not refresh cached referrals from unactive mounts (bsc#1213476).
• cifs: do not take exclusive lock for updating target hints (bsc#1213476).
• cifs: fix confusing debug message (bsc#1213476).
• cifs: fix missing unload_nls() in smb2_reconnect() (bsc#1213476).
• cifs: fix potential deadlock in cache_refresh_path() (bsc#1213476).
• cifs: fix refresh of cached referrals (bsc#1213476).
• cifs: fix return of uninitialized rc in dfs_cache_update_tgthint() (bsc#1213476).
• cifs: fix source pathname comparison of dfs supers (bsc#1213476).
• cifs: fix status checks in cifs_tree_connect (bsc#1213476).
• cifs: fix use-after-free bug in refresh_cache_worker() (bsc#1213476).
• cifs: get rid of dns resolve worker (bsc#1213476).
• cifs: get rid of mount options string parsing (bsc#1213476).
• cifs: handle cache lookup errors different than -ENOENT (bsc#1213476).
• cifs: ignore ipc reconnect failures during dfs failover (bsc#1213476).
• cifs: match even the scope id for ipv6 addresses (bsc#1213476).
• cifs: optimize reconnect of nested links (bsc#1213476).
• cifs: prevent data race in smb2_reconnect() (bsc#1213476). -…

The SUSE Linux Enterprise 15 SP3 kernel was updated to receive various security bugfixes.

The following security bugs were fixed:

• CVE-2021-47378: Destroy cm id before destroy qp to avoid use after free (bsc#1225201).
• CVE-2021-47496: Fix flipped sign in tls_err_abort() calls (bsc#1225354)
• CVE-2021-47402: Protect fl_walk() with rcu (bsc#1225301)
• CVE-2022-48673: kABI workarounds for struct smc_link (bsc#1223934).
• CVE-2023-52871: Handle a second device without data corruption (bsc#1225534)
• CVE-2024-26828: Fix underflow in parse_server_interfaces() (bsc#1223084).
• CVE-2021-47497: Fixed shift-out-of-bound (UBSAN) with byte size cells (bsc#1225355).
• CVE-2021-47500: Fixed trigger reference couting (bsc#1225360).
• CVE-2024-27413: Fix incorrect allocation size (bsc#1224438).
• CVE-2021-47383: Fiedx out-of-bound vmalloc access in imageblit (bsc#1225208).
• CVE-2021-47511: Fixed negative period/buffer sizes (bsc#1225411).
• CVE-2023-52840: Fix use after free in rmi_unregister_function() (bsc#1224928).
• CVE-2021-47261: Fix initializing CQ fragments buffer (bsc#1224954)
• CVE-2021-47254: Fix use-after-free in gfs2_glock_shrink_scan (bsc#1224888).
• CVE-2024-27398: Fixed use-after-free bugs caused by sco_sock_timeout (bsc#1224174).
• CVE-2024-26921: Preserve kabi for sk_buff (bsc#1223138).
• CVE-2023-52655: Check packet for fixup for true limit (bsc#1217169).
• CVE-2023-4244: Fixed a use-after-free in the nf_tables component, which could be exploited to achieve local privilege escalation (bsc#1215420).
• CVE-2023-4244: Fixed a use-after-free in the nf_tables component, which could be exploited to achieve local privilege escalation (bsc#1215420).
• CVE-2023-1829: Fixed a use-after-free vulnerability in the control index filter (tcindex) (bsc#1210335).
• CVE-2023-52686: Fix a null pointer in opal_event_init() (bsc#1065729).

The following non-security bugs were fixed:

• af_unix: annote lockless accesses to unix_tot_inflight & gc_in_progress (bsc#1223384).
• af_unix: Do not use atomic ops for unix_sk(sk)->inflight (bsc#1223384).
• af_unix: Replace BUG_ON() with WARN_ON_ONCE() (bsc#1223384).
• btrfs: do not start relocation until in progress drops are done (bsc#1222251).
• btrfs: do not start relocation until in progress drops are done (bsc#1222251).
• cifs: add missing spinlock around tcon refcount (bsc#1213476).
• cifs: avoid dup prefix path in dfs_get_automount_devname() (bsc#1213476).
• cifs: avoid race conditions with parallel reconnects (bsc#1213476).
• cifs: avoid re-lookups in dfs_cache_find() (bsc#1213476).
• cifs: avoid use of global locks for high contention data (bsc#1213476).
• cifs: check only tcon status on tcon related functions (bsc#1213476).
• cifs: do all necessary checks for credits within or before locking (bsc#1213476).
• cifs: do not block in dfs_cache_noreq_update_tgthint() (bsc#1213476).
• cifs: do not refresh cached referrals from unactive mounts (bsc#1213476).
• cifs: do not take exclusive lock for updating target hints (bsc#1213476).
• cifs: fix confusing debug message (bsc#1213476).
• cifs: fix missing unload_nls() in smb2_reconnect() (bsc#1213476).
• cifs: fix potential deadlock in cache_refresh_path() (bsc#1213476).
• cifs: fix refresh of cached referrals (bsc#1213476).
• cifs: fix return of uninitialized rc in dfs_cache_update_tgthint() (bsc#1213476).
• cifs: fix source pathname comparison of dfs supers (bsc#1213476).
• cifs: fix status checks in cifs_tree_connect (bsc#1213476).
• cifs: fix use-after-free bug in refresh_cache_worker() (bsc#1213476).
• cifs: get rid of dns resolve worker (bsc#1213476).
• cifs: get rid of mount options string parsing (bsc#1213476).
• cifs: handle cache lookup errors different than -ENOENT (bsc#1213476).
• cifs: ignore ipc reconnect failures during dfs failover (bsc#1213476).
• cifs: match even the scope id for ipv6 addresses (bsc#1213476).
• cifs: optimize reconnect of nested links (bsc#1213476).
• cifs: prevent data race in smb2_reconnect() (bsc#1213476). -…