CVE-2022-50517

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> mm/huge_memory: do not clobber swp_entry_t during THP split<br /> <br /> The following has been observed when running stressng mmap since commit<br /> b653db77350c ("mm: Clear page-&gt;private when splitting or migrating a page")<br /> <br /> watchdog: BUG: soft lockup - CPU#75 stuck for 26s! [stress-ng:9546]<br /> CPU: 75 PID: 9546 Comm: stress-ng Tainted: G E 6.0.0-revert-b653db77-fix+ #29 0357d79b60fb09775f678e4f3f64ef0579ad1374<br /> Hardware name: SGI.COM C2112-4GP3/X10DRT-P-Series, BIOS 2.0a 05/09/2016<br /> RIP: 0010:xas_descend+0x28/0x80<br /> Code: cc cc 0f b6 0e 48 8b 57 08 48 d3 ea 83 e2 3f 89 d0 48 83 c0 04 48 8b 44 c6 08 48 89 77 18 48 89 c1 83 e1 03 48 83 f9 02 75 08 3d fd 00 00 00 76 08 88 57 12 c3 cc cc cc cc 48 c1 e8 02 89 c2<br /> RSP: 0018:ffffbbf02a2236a8 EFLAGS: 00000246<br /> RAX: ffff9cab7d6a0002 RBX: ffffe04b0af88040 RCX: 0000000000000002<br /> RDX: 0000000000000030 RSI: ffff9cab60509b60 RDI: ffffbbf02a2236c0<br /> RBP: 0000000000000000 R08: ffff9cab60509b60 R09: ffffbbf02a2236c0<br /> R10: 0000000000000001 R11: ffffbbf02a223698 R12: 0000000000000000<br /> R13: ffff9cab4e28da80 R14: 0000000000039c01 R15: ffff9cab4e28da88<br /> FS: 00007fab89b85e40(0000) GS:ffff9cea3fcc0000(0000) knlGS:0000000000000000<br /> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033<br /> CR2: 00007fab84e00000 CR3: 00000040b73a4003 CR4: 00000000003706e0<br /> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000<br /> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400<br /> Call Trace:<br /> <br /> xas_load+0x3a/0x50<br /> __filemap_get_folio+0x80/0x370<br /> ? put_swap_page+0x163/0x360<br /> pagecache_get_page+0x13/0x90<br /> __try_to_reclaim_swap+0x50/0x190<br /> scan_swap_map_slots+0x31e/0x670<br /> get_swap_pages+0x226/0x3c0<br /> folio_alloc_swap+0x1cc/0x240<br /> add_to_swap+0x14/0x70<br /> shrink_page_list+0x968/0xbc0<br /> reclaim_page_list+0x70/0xf0<br /> reclaim_pages+0xdd/0x120<br /> madvise_cold_or_pageout_pte_range+0x814/0xf30<br /> walk_pgd_range+0x637/0xa30<br /> __walk_page_range+0x142/0x170<br /> walk_page_range+0x146/0x170<br /> madvise_pageout+0xb7/0x280<br /> ? asm_common_interrupt+0x22/0x40<br /> madvise_vma_behavior+0x3b7/0xac0<br /> ? find_vma+0x4a/0x70<br /> ? find_vma+0x64/0x70<br /> ? madvise_vma_anon_name+0x40/0x40<br /> madvise_walk_vmas+0xa6/0x130<br /> do_madvise+0x2f4/0x360<br /> __x64_sys_madvise+0x26/0x30<br /> do_syscall_64+0x5b/0x80<br /> ? do_syscall_64+0x67/0x80<br /> ? syscall_exit_to_user_mode+0x17/0x40<br /> ? do_syscall_64+0x67/0x80<br /> ? syscall_exit_to_user_mode+0x17/0x40<br /> ? do_syscall_64+0x67/0x80<br /> ? do_syscall_64+0x67/0x80<br /> ? common_interrupt+0x8b/0xa0<br /> entry_SYSCALL_64_after_hwframe+0x63/0xcd<br /> <br /> The problem can be reproduced with the mmtests config<br /> config-workload-stressng-mmap. It does not always happen and when it<br /> triggers is variable but it has happened on multiple machines.<br /> <br /> The intent of commit b653db77350c patch was to avoid the case where<br /> PG_private is clear but folio-&gt;private is not-NULL. However, THP tail<br /> pages uses page-&gt;private for "swp_entry_t if folio_test_swapcache()" as<br /> stated in the documentation for struct folio. This patch only clobbers<br /> page-&gt;private for tail pages if the head page was not in swapcache and<br /> warns once if page-&gt;private had an unexpected value.