Wind River Support Network

Description

In the Linux kernel, the following vulnerability has been resolved:[EOL][EOL]ext4: fix use-after-free read in ext4_find_extent for bigalloc + inline[EOL][EOL]Syzbot found the following issue:[EOL]loop0: detected capacity change from 0 to 2048[EOL]EXT4-fs (loop0): mounted filesystem 00000000-0000-0000-0000-000000000000 without journal. Quota mode: none.[EOL]==================================================================[EOL]BUG: KASAN: use-after-free in ext4_ext_binsearch_idx fs/ext4/extents.c:768 [inline][EOL]BUG: KASAN: use-after-free in ext4_find_extent+0x76e/0xd90 fs/ext4/extents.c:931[EOL]Read of size 4 at addr ffff888073644750 by task syz-executor420/5067[EOL][EOL]CPU: 0 PID: 5067 Comm: syz-executor420 Not tainted 6.2.0-rc1-syzkaller #0[EOL]Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022[EOL]Call Trace:[EOL] <TASK>[EOL] __dump_stack lib/dump_stack.c:88 [inline][EOL] dump_stack_lvl+0x1b1/0x290 lib/dump_stack.c:106[EOL] print_address_description+0x74/0x340 mm/kasan/report.c:306[EOL] print_report+0x107/0x1f0 mm/kasan/report.c:417[EOL] kasan_report+0xcd/0x100 mm/kasan/report.c:517[EOL] ext4_ext_binsearch_idx fs/ext4/extents.c:768 [inline][EOL] ext4_find_extent+0x76e/0xd90 fs/ext4/extents.c:931[EOL] ext4_clu_mapped+0x117/0x970 fs/ext4/extents.c:5809[EOL] ext4_insert_delayed_block fs/ext4/inode.c:1696 [inline][EOL] ext4_da_map_blocks fs/ext4/inode.c:1806 [inline][EOL] ext4_da_get_block_prep+0x9e8/0x13c0 fs/ext4/inode.c:1870[EOL] ext4_block_write_begin+0x6a8/0x2290 fs/ext4/inode.c:1098[EOL] ext4_da_write_begin+0x539/0x760 fs/ext4/inode.c:3082[EOL] generic_perform_write+0x2e4/0x5e0 mm/filemap.c:3772[EOL] ext4_buffered_write_iter+0x122/0x3a0 fs/ext4/file.c:285[EOL] ext4_file_write_iter+0x1d0/0x18f0[EOL] call_write_iter include/linux/fs.h:2186 [inline][EOL] new_sync_write fs/read_write.c:491 [inline][EOL] vfs_write+0x7dc/0xc50 fs/read_write.c:584[EOL] ksys_write+0x177/0x2a0 fs/read_write.c:637[EOL] do_syscall_x64 arch/x86/entry/common.c:50 [inline][EOL] do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80[EOL] entry_SYSCALL_64_after_hwframe+0x63/0xcd[EOL]RIP: 0033:0x7f4b7a9737b9[EOL]RSP: 002b:00007ffc5cac3668 EFLAGS: 00000246 ORIG_RAX: 0000000000000001[EOL]RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f4b7a9737b9[EOL]RDX: 00000000175d9003 RSI: 0000000020000200 RDI: 0000000000000004[EOL]RBP: 00007f4b7a933050 R08: 0000000000000000 R09: 0000000000000000[EOL]R10: 000000000000079f R11: 0000000000000246 R12: 00007f4b7a9330e0[EOL]R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000[EOL] </TASK>[EOL][EOL]Above issue is happens when enable bigalloc and inline data feature. As[EOL]commit 131294c35ed6 fixed delayed allocation bug in ext4_clu_mapped for[EOL]bigalloc + inline. But it only resolved issue when has inline data, if[EOL]inline data has been converted to extent(ext4_da_convert_inline_data_to_extent)[EOL]before writepages, there is no EXT4_STATE_MAY_INLINE_DATA flag. However[EOL]i_data is still store inline data in this scene. Then will trigger UAF[EOL]when find extent.[EOL]To resolve above issue, there is need to add judge "ext4_has_inline_data(inode)"[EOL]in ext4_clu_mapped().

CVEs

• CVE-2023-53692