kernel/git/stable/linux-stable.git/mm, branch linux-6.0.y

mm, compaction: fix fast_isolate_around() to stay within boundaries

2023-01-04T10:26:30+00:00

commit be21b32afe470c5ae98e27e49201158a47032942 upstream. Depending on the memory configuration, isolate_freepages_block() may scan pages out of the target range and causes panic. Panic can occur on systems with multiple zones in a single pageblock. The reason it is rare is that it only happens in special configurations. Depending on how many similar systems there are, it may be a good idea to fix this problem for older kernels as well. The problem is that pfn as argument of fast_isolate_around() could be out of the target range. Therefore we should consider the case where pfn < start_pfn, and also the case where end_pfn < pfn. This problem should have been addressd by the commit 6e2b7044c199 ("mm, compaction: make fast_isolate_freepages() stay within zone") but there was an oversight. Case1: pfn < start_pfn | node X's zone | node Y's zone +-----------------+------------------------------... pageblock ^ ^ ^ +-----------+-----------+-----------+-----------+... ^ ^ ^ ^ ^ end_pfn ^ start_pfn = cc->zone->zone_start_pfn pfn <---------> scanned range by "Scan After" Case2: end_pfn < pfn | node X's zone | node Y's zone +-----------------+------------------------------... pageblock ^ ^ ^ +-----------+-----------+-----------+-----------+... ^ ^ ^ ^ ^ pfn ^ end_pfn start_pfn <---------> scanned range by "Scan Before" It seems that there is no good reason to skip nr_isolated pages just after given pfn. So let perform simple scan from start to end instead of dividing the scan into "Before" and "After". Link: https://lkml.kernel.org/r/20221026112438.236336-1-a.naribayashi@fujitsu.com Fixes: 6e2b7044c199 ("mm, compaction: make fast_isolate_freepages() stay within zone"). Signed-off-by: NARIBAYASHI Akira Cc: David Rientjes Cc: Mel Gorman Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

mm/mempolicy: fix memory leak in set_mempolicy_home_node system call

2023-01-04T10:26:28+00:00

commit 38ce7c9bdfc228c14d7621ba36d3eebedd9d4f76 upstream. When encountering any vma in the range with policy other than MPOL_BIND or MPOL_PREFERRED_MANY, an error is returned without issuing a mpol_put on the policy just allocated with mpol_dup(). This allows arbitrary users to leak kernel memory. Link: https://lkml.kernel.org/r/20221215194621.202816-1-mathieu.desnoyers@efficios.com Fixes: c6018b4b2549 ("mm/mempolicy: add set_mempolicy_home_node syscall") Signed-off-by: Mathieu Desnoyers Reviewed-by: Randy Dunlap Reviewed-by: "Huang, Ying" Reviewed-by: Aneesh Kumar K.V Acked-by: Michal Hocko Cc: Aneesh Kumar K.V Cc: Dave Hansen Cc: Feng Tang Cc: Michal Hocko Cc: Andrea Arcangeli Cc: Mel Gorman Cc: Mike Kravetz Cc: Randy Dunlap Cc: Vlastimil Babka Cc: Andi Kleen Cc: Dan Williams Cc: Huang Ying Cc: [5.17+] Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

mm/gup: disallow FOLL_FORCE|FOLL_WRITE on hugetlb mappings

2022-12-31T12:26:53+00:00

commit f347454d034184b4f0a2caf6e14daf7848cea01c upstream. hugetlb does not support fake write-faults (write faults without write permissions). However, we are currently able to trigger a FAULT_FLAG_WRITE fault on a VMA without VM_WRITE. If we'd ever want to support FOLL_FORCE|FOLL_WRITE, we'd have to teach hugetlb to: (1) Leave the page mapped R/O after the fake write-fault, like maybe_mkwrite() does. (2) Allow writing to an exclusive anon page that's mapped R/O when FOLL_FORCE is set, like can_follow_write_pte(). E.g., __follow_hugetlb_must_fault() needs adjustment. For now, it's not clear if that added complexity is really required. History tolds us that FOLL_FORCE is dangerous and that we better limit its use to a bare minimum. -------------------------------------------------------------------------- #include #include #include #include #include #include #include #include int main(int argc, char **argv) { char *map; int mem_fd; map = mmap(NULL, 2 * 1024 * 1024u, PROT_READ, MAP_PRIVATE|MAP_ANON|MAP_HUGETLB|MAP_HUGE_2MB, -1, 0); if (map == MAP_FAILED) { fprintf(stderr, "mmap() failed: %d\n", errno); return 1; } mem_fd = open("/proc/self/mem", O_RDWR); if (mem_fd < 0) { fprintf(stderr, "open(/proc/self/mem) failed: %d\n", errno); return 1; } if (pwrite(mem_fd, "0", 1, (uintptr_t) map) == 1) { fprintf(stderr, "write() succeeded, which is unexpected\n"); return 1; } printf("write() failed as expected: %d\n", errno); return 0; } -------------------------------------------------------------------------- Fortunately, we have a sanity check in hugetlb_wp() in place ever since commit 1d8d14641fd9 ("mm/hugetlb: support write-faults in shared mappings"), that bails out instead of silently mapping a page writable in a !PROT_WRITE VMA. Consequently, above reproducer triggers a warning, similar to the one reported by szsbot: ------------[ cut here ]------------ WARNING: CPU: 1 PID: 3612 at mm/hugetlb.c:5313 hugetlb_wp+0x20a/0x1af0 mm/hugetlb.c:5313 Modules linked in: CPU: 1 PID: 3612 Comm: syz-executor250 Not tainted 6.1.0-rc2-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/11/2022 RIP: 0010:hugetlb_wp+0x20a/0x1af0 mm/hugetlb.c:5313 Code: ea 03 80 3c 02 00 0f 85 31 14 00 00 49 8b 5f 20 31 ff 48 89 dd 83 e5 02 48 89 ee e8 70 ab b7 ff 48 85 ed 75 5b e8 76 ae b7 ff <0f> 0b 41 bd 40 00 00 00 e8 69 ae b7 ff 48 b8 00 00 00 00 00 fc ff RSP: 0018:ffffc90003caf620 EFLAGS: 00010293 RAX: 0000000000000000 RBX: 0000000008640070 RCX: 0000000000000000 RDX: ffff88807b963a80 RSI: ffffffff81c4ed2a RDI: 0000000000000007 RBP: 0000000000000000 R08: 0000000000000007 R09: 0000000000000000 R10: 0000000000000000 R11: 000000000008c07e R12: ffff888023805800 R13: 0000000000000000 R14: ffffffff91217f38 R15: ffff88801d4b0360 FS: 0000555555bba300(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fff7a47a1b8 CR3: 000000002378d000 CR4: 00000000003506e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: hugetlb_no_page mm/hugetlb.c:5755 [inline] hugetlb_fault+0x19cc/0x2060 mm/hugetlb.c:5874 follow_hugetlb_page+0x3f3/0x1850 mm/hugetlb.c:6301 __get_user_pages+0x2cb/0xf10 mm/gup.c:1202 __get_user_pages_locked mm/gup.c:1434 [inline] __get_user_pages_remote+0x18f/0x830 mm/gup.c:2187 get_user_pages_remote+0x84/0xc0 mm/gup.c:2260 __access_remote_vm+0x287/0x6b0 mm/memory.c:5517 ptrace_access_vm+0x181/0x1d0 kernel/ptrace.c:61 generic_ptrace_pokedata kernel/ptrace.c:1323 [inline] ptrace_request+0xb46/0x10c0 kernel/ptrace.c:1046 arch_ptrace+0x36/0x510 arch/x86/kernel/ptrace.c:828 __do_sys_ptrace kernel/ptrace.c:1296 [inline] __se_sys_ptrace kernel/ptrace.c:1269 [inline] __x64_sys_ptrace+0x178/0x2a0 kernel/ptrace.c:1269 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd [...] So let's silence that warning by teaching GUP code that FOLL_FORCE -- so far -- does not apply to hugetlb. Note that FOLL_FORCE for read-access seems to be working as expected. The assumption is that this has been broken forever, only ever since above commit, we actually detect the wrong handling and WARN_ON_ONCE(). I assume this has been broken at least since 2014, when mm/gup.c came to life. I failed to come up with a suitable Fixes tag quickly. Link: https://lkml.kernel.org/r/20221031152524.173644-1-david@redhat.com Fixes: 1d8d14641fd9 ("mm/hugetlb: support write-faults in shared mappings") Signed-off-by: David Hildenbrand Reported-by: Cc: Mike Kravetz Cc: Peter Xu Cc: John Hubbard Cc: Jason Gunthorpe Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

tmpfs: fix data loss from failed fallocate

2022-12-14T10:40:52+00:00

commit 44bcabd70cf1425b4243e02251c02b01638a8287 upstream. Fix tmpfs data loss when the fallocate system call is interrupted by a signal, or fails for some other reason. The partial folio handling in shmem_undo_range() forgot to consider this unfalloc case, and was liable to erase or truncate out data which had already been committed earlier. It turns out that none of the partial folio handling there is appropriate for the unfalloc case, which just wants to proceed to removal of whole folios: which find_get_entries() provides, even when partially covered. Original patch by Rui Wang. Link: https://lore.kernel.org/linux-mm/33b85d82.7764.1842e9ab207.Coremail.chenguoqic@163.com/ Link: https://lkml.kernel.org/r/a5dac112-cf4b-7af-a33-f386e347fd38@google.com Fixes: b9a8a4195c7d ("truncate,shmem: Handle truncates that split large folios") Signed-off-by: Hugh Dickins Reported-by: Guoqi Chen Link: https://lore.kernel.org/all/20221101032248.819360-1-kernel@hev.cc/ Cc: Rui Wang Cc: Huacai Chen Cc: Matthew Wilcox Cc: Vishal Moola (Oracle) Cc: [5.17+] Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

mm/gup: fix gup_pud_range() for dax

2022-12-14T10:40:52+00:00

commit fcd0ccd836ffad73d98a66f6fea7b16f735ea920 upstream. For dax pud, pud_huge() returns true on x86. So the function works as long as hugetlb is configured. However, dax doesn't depend on hugetlb. Commit 414fd080d125 ("mm/gup: fix gup_pmd_range() for dax") fixed devmap-backed huge PMDs, but missed devmap-backed huge PUDs. Fix this as well. This fixes the below kernel panic: general protection fault, probably for non-canonical address 0x69e7c000cc478: 0000 [#1] SMP < snip > Call Trace: get_user_pages_fast+0x1f/0x40 iov_iter_get_pages+0xc6/0x3b0 ? mempool_alloc+0x5d/0x170 bio_iov_iter_get_pages+0x82/0x4e0 ? bvec_alloc+0x91/0xc0 ? bio_alloc_bioset+0x19a/0x2a0 blkdev_direct_IO+0x282/0x480 ? __io_complete_rw_common+0xc0/0xc0 ? filemap_range_has_page+0x82/0xc0 generic_file_direct_write+0x9d/0x1a0 ? inode_update_time+0x24/0x30 __generic_file_write_iter+0xbd/0x1e0 blkdev_write_iter+0xb4/0x150 ? io_import_iovec+0x8d/0x340 io_write+0xf9/0x300 io_issue_sqe+0x3c3/0x1d30 ? sysvec_reschedule_ipi+0x6c/0x80 __io_queue_sqe+0x33/0x240 ? fget+0x76/0xa0 io_submit_sqes+0xe6a/0x18d0 ? __fget_light+0xd1/0x100 __x64_sys_io_uring_enter+0x199/0x880 ? __context_tracking_enter+0x1f/0x70 ? irqentry_exit_to_user_mode+0x24/0x30 ? irqentry_exit+0x1d/0x30 ? __context_tracking_exit+0xe/0x70 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x61/0xcb RIP: 0033:0x7fc97c11a7be < snip > ---[ end trace 48b2e0e67debcaeb ]--- RIP: 0010:internal_get_user_pages_fast+0x340/0x990 < snip > Kernel panic - not syncing: Fatal exception Kernel Offset: disabled Link: https://lkml.kernel.org/r/1670392853-28252-1-git-send-email-ssengar@linux.microsoft.com Fixes: 414fd080d125 ("mm/gup: fix gup_pmd_range() for dax") Signed-off-by: John Starks Signed-off-by: Saurabh Sengar Cc: Jan Kara Cc: Yu Zhao Cc: Jason Gunthorpe Cc: John Hubbard Cc: David Hildenbrand Cc: Dan Williams Cc: Alistair Popple Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

memcg: fix possible use-after-free in memcg_write_event_control()

2022-12-14T10:40:52+00:00

commit 4a7ba45b1a435e7097ca0f79a847d0949d0eb088 upstream. memcg_write_event_control() accesses the dentry->d_name of the specified control fd to route the write call. As a cgroup interface file can't be renamed, it's safe to access d_name as long as the specified file is a regular cgroup file. Also, as these cgroup interface files can't be removed before the directory, it's safe to access the parent too. Prior to 347c4a874710 ("memcg: remove cgroup_event->cft"), there was a call to __file_cft() which verified that the specified file is a regular cgroupfs file before further accesses. The cftype pointer returned from __file_cft() was no longer necessary and the commit inadvertently dropped the file type check with it allowing any file to slip through. With the invarients broken, the d_name and parent accesses can now race against renames and removals of arbitrary files and cause use-after-free's. Fix the bug by resurrecting the file type check in __file_cft(). Now that cgroupfs is implemented through kernfs, checking the file operations needs to go through a layer of indirection. Instead, let's check the superblock and dentry type. Link: https://lkml.kernel.org/r/Y5FRm/cfcKPGzWwl@slm.duckdns.org Fixes: 347c4a874710 ("memcg: remove cgroup_event->cft") Signed-off-by: Tejun Heo Reported-by: Jann Horn Acked-by: Roman Gushchin Acked-by: Johannes Weiner Cc: Linus Torvalds Cc: Michal Hocko Cc: Muchun Song Cc: Shakeel Butt Cc: [3.14+] Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman

hugetlb: don't delete vma_lock in hugetlb MADV_DONTNEED processing

2022-12-14T10:40:50+00:00

commit 04ada095dcfc4ae359418053c0be94453bdf1e84 upstream. madvise(MADV_DONTNEED) ends up calling zap_page_range() to clear page tables associated with the address range. For hugetlb vmas, zap_page_range will call __unmap_hugepage_range_final. However, __unmap_hugepage_range_final assumes the passed vma is about to be removed and deletes the vma_lock to prevent pmd sharing as the vma is on the way out. In the case of madvise(MADV_DONTNEED) the vma remains, but the missing vma_lock prevents pmd sharing and could potentially lead to issues with truncation/fault races. This issue was originally reported here [1] as a BUG triggered in page_try_dup_anon_rmap. Prior to the introduction of the hugetlb vma_lock, __unmap_hugepage_range_final cleared the VM_MAYSHARE flag to prevent pmd sharing. Subsequent faults on this vma were confused as VM_MAYSHARE indicates a sharable vma, but was not set so page_mapping was not set in new pages added to the page table. This resulted in pages that appeared anonymous in a VM_SHARED vma and triggered the BUG. Address issue by adding a new zap flag ZAP_FLAG_UNMAP to indicate an unmap call from unmap_vmas(). This is used to indicate the 'final' unmapping of a hugetlb vma. When called via MADV_DONTNEED, this flag is not set and the vm_lock is not deleted. NOTE - Prior to the introduction of the huegtlb vma_lock in v6.1, this issue is addressed by not clearing the VM_MAYSHARE flag when __unmap_hugepage_range_final is called in the MADV_DONTNEED case. [1] https://lore.kernel.org/lkml/CAO4mrfdLMXsao9RF4fUE8-Wfde8xmjsKrTNMNC9wjUb6JudD0g@mail.gmail.com/ Link: https://lkml.kernel.org/r/20221114235507.294320-3-mike.kravetz@oracle.com Fixes: 90e7e7f5ef3f ("mm: enable MADV_DONTNEED for hugetlb mappings") Signed-off-by: Mike Kravetz Reported-by: Wei Chen Cc: Axel Rasmussen Cc: David Hildenbrand Cc: Matthew Wilcox Cc: Mina Almasry Cc: Nadav Amit Cc: Naoya Horiguchi Cc: Peter Xu Cc: Rik van Riel Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton Signed-off-by: Mike Kravetz Signed-off-by: Sasha Levin

mm/khugepaged: invoke MMU notifiers in shmem/file collapse paths

2022-12-14T10:40:50+00:00

commit f268f6cf875f3220afc77bdd0bf1bb136eb54db9 upstream. Any codepath that zaps page table entries must invoke MMU notifiers to ensure that secondary MMUs (like KVM) don't keep accessing pages which aren't mapped anymore. Secondary MMUs don't hold their own references to pages that are mirrored over, so failing to notify them can lead to page use-after-free. I'm marking this as addressing an issue introduced in commit f3f0e1d2150b ("khugepaged: add support of collapse for tmpfs/shmem pages"), but most of the security impact of this only came in commit 27e1f8273113 ("khugepaged: enable collapse pmd for pte-mapped THP"), which actually omitted flushes for the removal of present PTEs, not just for the removal of empty page tables. Link: https://lkml.kernel.org/r/20221129154730.2274278-3-jannh@google.com Link: https://lkml.kernel.org/r/20221128180252.1684965-3-jannh@google.com Link: https://lkml.kernel.org/r/20221125213714.4115729-3-jannh@google.com Fixes: f3f0e1d2150b ("khugepaged: add support of collapse for tmpfs/shmem pages") Signed-off-by: Jann Horn Acked-by: David Hildenbrand Reviewed-by: Yang Shi Cc: John Hubbard Cc: Peter Xu Cc: Signed-off-by: Andrew Morton [backported, no changes necessary] Signed-off-by: Jann Horn Signed-off-by: Sasha Levin

mm/khugepaged: fix GUP-fast interaction by sending IPI

2022-12-14T10:40:50+00:00

commit 2ba99c5e08812494bc57f319fb562f527d9bacd8 upstream. Since commit 70cbc3cc78a99 ("mm: gup: fix the fast GUP race against THP collapse"), the lockless_pages_from_mm() fastpath rechecks the pmd_t to ensure that the page table was not removed by khugepaged in between. However, lockless_pages_from_mm() still requires that the page table is not concurrently freed. Fix it by sending IPIs (if the architecture uses semi-RCU-style page table freeing) before freeing/reusing page tables. Link: https://lkml.kernel.org/r/20221129154730.2274278-2-jannh@google.com Link: https://lkml.kernel.org/r/20221128180252.1684965-2-jannh@google.com Link: https://lkml.kernel.org/r/20221125213714.4115729-2-jannh@google.com Fixes: ba76149f47d8 ("thp: khugepaged") Signed-off-by: Jann Horn Reviewed-by: Yang Shi Acked-by: David Hildenbrand Cc: John Hubbard Cc: Peter Xu Cc: Signed-off-by: Andrew Morton [backported, no changes necessary] Signed-off-by: Jann Horn Signed-off-by: Sasha Levin

mm/khugepaged: take the right locks for page table retraction

2022-12-14T10:40:49+00:00

commit 8d3c106e19e8d251da31ff4cc7462e4565d65084 upstream. pagetable walks on address ranges mapped by VMAs can be done under the mmap lock, the lock of an anon_vma attached to the VMA, or the lock of the VMA's address_space. Only one of these needs to be held, and it does not need to be held in exclusive mode. Under those circumstances, the rules for concurrent access to page table entries are: - Terminal page table entries (entries that don't point to another page table) can be arbitrarily changed under the page table lock, with the exception that they always need to be consistent for hardware page table walks and lockless_pages_from_mm(). This includes that they can be changed into non-terminal entries. - Non-terminal page table entries (which point to another page table) can not be modified; readers are allowed to READ_ONCE() an entry, verify that it is non-terminal, and then assume that its value will stay as-is. Retracting a page table involves modifying a non-terminal entry, so page-table-level locks are insufficient to protect against concurrent page table traversal; it requires taking all the higher-level locks under which it is possible to start a page walk in the relevant range in exclusive mode. The collapse_huge_page() path for anonymous THP already follows this rule, but the shmem/file THP path was getting it wrong, making it possible for concurrent rmap-based operations to cause corruption. Link: https://lkml.kernel.org/r/20221129154730.2274278-1-jannh@google.com Link: https://lkml.kernel.org/r/20221128180252.1684965-1-jannh@google.com Link: https://lkml.kernel.org/r/20221125213714.4115729-1-jannh@google.com Fixes: 27e1f8273113 ("khugepaged: enable collapse pmd for pte-mapped THP") Signed-off-by: Jann Horn Reviewed-by: Yang Shi Acked-by: David Hildenbrand Cc: John Hubbard Cc: Peter Xu Cc: Signed-off-by: Andrew Morton [backport fixed up manually: collapse_pte_mapped_thp returns different type] Signed-off-by: Jann Horn Signed-off-by: Sasha Levin