CVE-2023-54165

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> zsmalloc: move LRU update from zs_map_object() to zs_malloc()<br /> <br /> Under memory pressure, we sometimes observe the following crash:<br /> <br /> [ 5694.832838] ------------[ cut here ]------------<br /> [ 5694.842093] list_del corruption, ffff888014b6a448-&gt;next is LIST_POISON1 (dead000000000100)<br /> [ 5694.858677] WARNING: CPU: 33 PID: 418824 at lib/list_debug.c:47 __list_del_entry_valid+0x42/0x80<br /> [ 5694.961820] CPU: 33 PID: 418824 Comm: fuse_counters.s Kdump: loaded Tainted: G S 5.19.0-0_fbk3_rc3_hoangnhatpzsdynshrv41_10870_g85a9558a25de #1<br /> [ 5694.990194] Hardware name: Wiwynn Twin Lakes MP/Twin Lakes Passive MP, BIOS YMM16 05/24/2021<br /> [ 5695.007072] RIP: 0010:__list_del_entry_valid+0x42/0x80<br /> [ 5695.017351] Code: 08 48 83 c2 22 48 39 d0 74 24 48 8b 10 48 39 f2 75 2c 48 8b 51 08 b0 01 48 39 f2 75 34 c3 48 c7 c7 55 d7 78 82 e8 4e 45 3b 00 0b eb 31 48 c7 c7 27 a8 70 82 e8 3e 45 3b 00 0f 0b eb 21 48 c7<br /> [ 5695.054919] RSP: 0018:ffffc90027aef4f0 EFLAGS: 00010246<br /> [ 5695.065366] RAX: 41fe484987275300 RBX: ffff888008988180 RCX: 0000000000000000<br /> [ 5695.079636] RDX: ffff88886006c280 RSI: ffff888860060480 RDI: ffff888860060480<br /> [ 5695.093904] RBP: 0000000000000002 R08: 0000000000000000 R09: ffffc90027aef370<br /> [ 5695.108175] R10: 0000000000000000 R11: ffffffff82fdf1c0 R12: 0000000010000002<br /> [ 5695.122447] R13: ffff888014b6a448 R14: ffff888014b6a420 R15: 00000000138dc240<br /> [ 5695.136717] FS: 00007f23a7d3f740(0000) GS:ffff888860040000(0000) knlGS:0000000000000000<br /> [ 5695.152899] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033<br /> [ 5695.164388] CR2: 0000560ceaab6ac0 CR3: 000000001c06c001 CR4: 00000000007706e0<br /> [ 5695.178659] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000<br /> [ 5695.192927] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400<br /> [ 5695.207197] PKRU: 55555554<br /> [ 5695.212602] Call Trace:<br /> [ 5695.217486] <br /> [ 5695.221674] zs_map_object+0x91/0x270<br /> [ 5695.229000] zswap_frontswap_store+0x33d/0x870<br /> [ 5695.237885] ? do_raw_spin_lock+0x5d/0xa0<br /> [ 5695.245899] __frontswap_store+0x51/0xb0<br /> [ 5695.253742] swap_writepage+0x3c/0x60<br /> [ 5695.261063] shrink_page_list+0x738/0x1230<br /> [ 5695.269255] shrink_lruvec+0x5ec/0xcd0<br /> [ 5695.276749] ? shrink_slab+0x187/0x5f0<br /> [ 5695.284240] ? mem_cgroup_iter+0x6e/0x120<br /> [ 5695.292255] shrink_node+0x293/0x7b0<br /> [ 5695.299402] do_try_to_free_pages+0xea/0x550<br /> [ 5695.307940] try_to_free_pages+0x19a/0x490<br /> [ 5695.316126] __folio_alloc+0x19ff/0x3e40<br /> [ 5695.323971] ? __filemap_get_folio+0x8a/0x4e0<br /> [ 5695.332681] ? walk_component+0x2a8/0xb50<br /> [ 5695.340697] ? generic_permission+0xda/0x2a0<br /> [ 5695.349231] ? __filemap_get_folio+0x8a/0x4e0<br /> [ 5695.357940] ? walk_component+0x2a8/0xb50<br /> [ 5695.365955] vma_alloc_folio+0x10e/0x570<br /> [ 5695.373796] ? walk_component+0x52/0xb50<br /> [ 5695.381634] wp_page_copy+0x38c/0xc10<br /> [ 5695.388953] ? filename_lookup+0x378/0xbc0<br /> [ 5695.397140] handle_mm_fault+0x87f/0x1800<br /> [ 5695.405157] do_user_addr_fault+0x1bd/0x570<br /> [ 5695.413520] exc_page_fault+0x5d/0x110<br /> [ 5695.421017] asm_exc_page_fault+0x22/0x30<br /> <br /> After some investigation, I have found the following issue: unlike other<br /> zswap backends, zsmalloc performs the LRU list update at the object<br /> mapping time, rather than when the slot for the object is allocated.<br /> This deviation was discussed and agreed upon during the review process<br /> of the zsmalloc writeback patch series:<br /> <br /> https://lore.kernel.org/lkml/Y3flcAXNxxrvy3ZH@cmpxchg.org/<br /> <br /> Unfortunately, this introduces a subtle bug that occurs when there is a<br /> concurrent store and reclaim, which interleave as follows:<br /> <br /> zswap_frontswap_store() shrink_worker()<br /> zs_malloc() zs_zpool_shrink()<br /> spin_lock(&amp;pool-&gt;lock) zs_reclaim_page()<br /> zspage = find_get_zspage()<br /> spin_unlock(&amp;pool-&gt;lock)<br /> spin_lock(&amp;pool-&gt;lock)<br /> zspage = list_first_entry(&amp;pool-&gt;lru)<br /> <br /> ---truncated---