kernel/linux.git/fs/pstore, branch v6.6.132

pstore: ram_core: fix incorrect success return when vmap() fails

2026-03-04T12:20:36+00:00

[ Upstream commit 05363abc7625cf18c96e67f50673cd07f11da5e9 ] In persistent_ram_vmap(), vmap() may return NULL on failure. If offset is non-zero, adding offset_in_page(start) causes the function to return a non-NULL pointer even though the mapping failed. persistent_ram_buffer_map() therefore incorrectly returns success. Subsequent access to prz->buffer may dereference an invalid address and cause crashes. Add proper NULL checking for vmap() failures. Signed-off-by: Ruipeng Qi Link: https://patch.msgid.link/20260203020358.3315299-1-ruipengqi3@gmail.com Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore/ram: fix buffer overflow in persistent_ram_save_old()

2026-03-04T12:19:34+00:00

[ Upstream commit 5669645c052f235726a85f443769b6fc02f66762 ] persistent_ram_save_old() can be called multiple times for the same persistent_ram_zone (e.g., via ramoops_pstore_read -> ramoops_get_next_prz for PSTORE_TYPE_DMESG records). Currently, the function only allocates prz->old_log when it is NULL, but it unconditionally updates prz->old_log_size to the current buffer size and then performs memcpy_fromio() using this new size. If the buffer size has grown since the first allocation (which can happen across different kernel boot cycles), this leads to: 1. A heap buffer overflow (OOB write) in the memcpy_fromio() calls 2. A subsequent OOB read when ramoops_pstore_read() accesses the buffer using the incorrect (larger) old_log_size The KASAN splat would look similar to: BUG: KASAN: slab-out-of-bounds in ramoops_pstore_read+0x... Read of size N at addr ... by task ... The conditions are likely extremely hard to hit: 0. Crash with a ramoops write of less-than-record-max-size bytes. 1. Reboot: ramoops registers, pstore_get_records(0) reads old crash, allocates old_log with size X 2. Crash handler registered, timer started (if pstore_update_ms >= 0) 3. Oops happens (non-fatal, system continues) 4. pstore_dump() writes oops via ramoops_pstore_write() size Y (>X) 5. pstore_new_entry = 1, pstore_timer_kick() called 6. System continues running (not a panic oops) 7. Timer fires after pstore_update_ms milliseconds 8. pstore_timefunc() → schedule_work() → pstore_dowork() → pstore_get_records(1) 9. ramoops_get_next_prz() → persistent_ram_save_old() 10. buffer_size() returns Y, but old_log is X bytes 11. Y > X: memcpy_fromio() overflows heap Requirements: - a prior crash record exists that did not fill the record size (almost impossible since the crash handler writes as much as it can possibly fit into the record, capped by max record size and the kmsg buffer almost always exceeds the max record size) - pstore_update_ms >= 0 (disabled by default) - Non-fatal oops (system survives) Free and reallocate the buffer when the new size differs from the previously allocated size. This ensures old_log always has sufficient space for the data being copied. Fixes: 201e4aca5aa1 ("pstore/ram: Should update old dmesg buffer before reading") Signed-off-by: Sai Ritvik Tanksalkar Link: https://patch.msgid.link/20260201132240.2948732-1-stanksal@purdue.edu Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore: Change kmsg_bytes storage size to u32

2025-06-04T12:42:11+00:00

[ Upstream commit 5674609535bafa834ab014d90d9bbe8e89223a0b ] The types around kmsg_bytes were inconsistent. The global was unsigned long, the argument to pstore_set_kmsg_bytes() was int, and the filesystem option was u32. Given other internal limits, there's not much sense in making a single pstore record larger than INT_MAX and it can't be negative, so use u32 everywhere. Additionally, use READ/WRITE_ONCE and a local variable in pstore_dump() to avoid kmsg_bytes changing during a dump. Link: https://lore.kernel.org/r/20250206191655.work.798-kees@kernel.org Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore/blk: trivial typo fixes

2025-02-08T08:51:42+00:00

[ Upstream commit 542243af7182efaeaf6d0f4643f7de437541a9af ] Fix trivial typos in comments. Fixes: 2a03ddbde1e1 ("pstore/blk: Move verify_size() macro out of function") Fixes: 17639f67c1d6 ("pstore/blk: Introduce backend for block devices") Signed-off-by: Eugen Hristev Link: https://lore.kernel.org/r/20250101111921.850406-1-eugen.hristev@linaro.org Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore/zone: Add a null pointer check to the psz_kmsg_read

2024-04-13T11:07:31+00:00

[ Upstream commit 98bc7e26e14fbb26a6abf97603d59532475e97f8 ] kasprintf() returns a pointer to dynamically allocated memory which can be NULL upon failure. Ensure the allocation was successful by checking the pointer validity. Signed-off-by: Kunwu Chan Link: https://lore.kernel.org/r/20240118100206.213928-1-chentao@kylinos.cn Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore: inode: Only d_invalidate() is needed

2024-03-26T22:19:34+00:00

[ Upstream commit a43e0fc5e9134a46515de2f2f8d4100b74e50de3 ] Unloading a modular pstore backend with records in pstorefs would trigger the dput() double-drop warning: WARNING: CPU: 0 PID: 2569 at fs/dcache.c:762 dput.part.0+0x3f3/0x410 Using the combo of d_drop()/dput() (as mentioned in Documentation/filesystems/vfs.rst) isn't the right approach here, and leads to the reference counting problem seen above. Use d_invalidate() and update the code to not bother checking for error codes that can never happen. Suggested-by: Alexander Viro Fixes: 609e28bb139e ("pstore: Remove filesystem records when backend is unregistered") Signed-off-by: Kees Cook

pstore: inode: Convert mutex usage to guard(mutex)

2024-03-26T22:19:34+00:00

[ Upstream commit e2eeddefb046dbc771a6fa426f7f98fb25adfe68 ] Replace open-coded mutex handling with cleanup.h guard(mutex) and scoped_guard(mutex, ...). Cc: Guilherme G. Piccoli Cc: Tony Luck Cc: Link: https://lore.kernel.org/r/20231205182622.1329923-2-keescook@chromium.org Signed-off-by: Kees Cook Stable-dep-of: a43e0fc5e913 ("pstore: inode: Only d_invalidate() is needed") Signed-off-by: Sasha Levin

pstore/ram: Fix crash when setting number of cpus to an odd number

2024-02-05T20:14:16+00:00

[ Upstream commit d49270a04623ce3c0afddbf3e984cb245aa48e9c ] When the number of cpu cores is adjusted to 7 or other odd numbers, the zone size will become an odd number. The address of the zone will become: addr of zone0 = BASE addr of zone1 = BASE + zone_size addr of zone2 = BASE + zone_size*2 ... The address of zone1/3/5/7 will be mapped to non-alignment va. Eventually crashes will occur when accessing these va. So, use ALIGN_DOWN() to make sure the zone size is even to avoid this bug. Signed-off-by: Weichen Chen Reviewed-by: Matthias Brugger Tested-by: "Guilherme G. Piccoli" Link: https://lore.kernel.org/r/20230224023632.6840-1-weichen.chen@mediatek.com Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore: ram_core: fix possible overflow in persistent_ram_init_ecc()

2024-01-25T23:35:17+00:00

[ Upstream commit 86222a8fc16ec517de8da2604d904c9df3a08e5d ] In persistent_ram_init_ecc(), on 64-bit arches DIV_ROUND_UP() will return 64-bit value since persistent_ram_zone::buffer_size has type size_t which is derived from the 64-bit *unsigned long*, while the ecc_blocks variable this value gets assigned to has (always 32-bit) *int* type. Even if that value fits into *int* type, an overflow is still possible when calculating the size_t typed ecc_total variable further below since there's no cast to any 64-bit type before multiplication. Declaring the ecc_blocks variable as *size_t* should fix this mess... Found by Linux Verification Center (linuxtesting.org) with the SVACE static analysis tool. Fixes: 9cc05ad97c57 ("staging: android: persistent_ram: refactor ecc support") Signed-off-by: Sergey Shtylyov Link: https://lore.kernel.org/r/20231105202936.25694-1-s.shtylyov@omp.ru Signed-off-by: Kees Cook Signed-off-by: Sasha Levin

pstore/platform: Add check for kstrdup

2023-11-20T10:58:53+00:00

[ Upstream commit a19d48f7c5d57c0f0405a7d4334d1d38fe9d3c1c ] Add check for the return value of kstrdup() and return the error if it fails in order to avoid NULL pointer dereference. Fixes: 563ca40ddf40 ("pstore/platform: Switch pstore_info::name to const") Signed-off-by: Jiasheng Jiang Link: https://lore.kernel.org/r/20230623022706.32125-1-jiasheng@iscas.ac.cn Signed-off-by: Kees Cook Signed-off-by: Sasha Levin