Release refcount after xas_set to fix UAF which may cause panic like this:
page:ffffea000491fa40 refcount:1 mapcount:0 mapping:0000000000000000 index:0x1 pfn:0x1247e9 head:ffffea000491fa00 order:3 compound_mapcount:0 compound_pincount:0 memcg:ffff888104f91091 flags: 0x2fffff80010200(slab|head|node=0|zone=2|lastcpupid=0x1fffff) ... page dumped because: VM_BUG_ON_PAGE(PageTail(page)) ------------[ cut here ]------------ kernel BUG at include/linux/page-flags.h:632! invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN CPU: 1 PID: 7642 Comm: sh Not tainted 5.15.51-dirty #26 ... Call Trace: <TASK> __invalidate_mapping_pages+0xe7/0x540 drop_pagecache_sb+0x159/0x320 iterate_supers+0x120/0x240 drop_caches_sysctl_handler+0xaa/0xe0 proc_sys_call_handler+0x2b4/0x480 new_sync_write+0x3d6/0x5c0 vfs_write+0x446/0x7a0 ksys_write+0x105/0x210 do_syscall_64+0x35/0x80 entry_SYSCALL_64_after_hwframe+0x44/0xae RIP: 0033:0x7f52b5733130 ...
This problem has been fixed on mainline by patch 6b24ca4a1a8d ("mm: Use multi-index entries in the page cache") since it deletes the related code.
Fixes: 5c211ba29deb ("mm: add and use find_lock_entries") Signed-off-by: Liu Shixin liushixin2@huawei.com --- mm/filemap.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/mm/filemap.c b/mm/filemap.c index 00e391e75880..24b5d7ebdc29 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2087,16 +2087,18 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, { XA_STATE(xas, &mapping->i_pages, start); struct page *page; + bool putpage;
rcu_read_lock(); while ((page = find_get_entry(&xas, end, XA_PRESENT))) { + putpage = true; if (!xa_is_value(page)) { if (page->index < start) - goto put; + goto next; if (page->index + thp_nr_pages(page) - 1 > end) - goto put; + goto next; if (!trylock_page(page)) - goto put; + goto next; if (page->mapping != mapping || PageWriteback(page)) goto unlock; VM_BUG_ON_PAGE(!thp_contains(page, xas.xa_index), @@ -2105,20 +2107,24 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, indices[pvec->nr] = xas.xa_index; if (!pagevec_add(pvec, page)) break; + putpage = false; goto next; unlock: unlock_page(page); -put: - put_page(page); next: if (!xa_is_value(page) && PageTransHuge(page)) { unsigned int nr_pages = thp_nr_pages(page);
/* Final THP may cross MAX_LFS_FILESIZE on 32-bit */ xas_set(&xas, page->index + nr_pages); - if (xas.xa_index < nr_pages) + if (xas.xa_index < nr_pages) { + if (putpage) + put_page(page); break; + } } + if (putpage) + put_page(page); } rcu_read_unlock();
On Wed, Jul 06, 2022 at 11:24:34AM +0800, Liu Shixin wrote:
Release refcount after xas_set to fix UAF which may cause panic like this:
I think we can do better. How about this?
diff --git a/mm/filemap.c b/mm/filemap.c index 00e391e75880..11ae38cc4fd3 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2090,7 +2090,9 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start,
rcu_read_lock(); while ((page = find_get_entry(&xas, end, XA_PRESENT))) { + unsigned long next_idx = xas.xa_index + 1; if (!xa_is_value(page)) { + next_idx = page->index + thp_nr_pages(page); if (page->index < start) goto put; if (page->index + thp_nr_pages(page) - 1 > end) @@ -2111,14 +2113,11 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, put: put_page(page); next: - if (!xa_is_value(page) && PageTransHuge(page)) { - unsigned int nr_pages = thp_nr_pages(page); - - /* Final THP may cross MAX_LFS_FILESIZE on 32-bit */ - xas_set(&xas, page->index + nr_pages); - if (xas.xa_index < nr_pages) - break; - } + /* Final THP may cross MAX_LFS_FILESIZE on 32-bit */ + if (next_idx < xas.xa_index) + break; + if (next_idx != xas.xa_index + 1) + xas_set(&xas, next_idx); } rcu_read_unlock();
On 2022/7/6 11:21, Matthew Wilcox wrote:
On Wed, Jul 06, 2022 at 11:24:34AM +0800, Liu Shixin wrote:
Release refcount after xas_set to fix UAF which may cause panic like this:
I think we can do better. How about this?
Thank you for your idea, it looks more concise.
diff --git a/mm/filemap.c b/mm/filemap.c index 00e391e75880..11ae38cc4fd3 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2090,7 +2090,9 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, rcu_read_lock(); while ((page = find_get_entry(&xas, end, XA_PRESENT))) {
unsigned long next_idx = xas.xa_index + 1;next_idx = page->index + thp_nr_pages(page); if (page->index < start) goto put; if (page->index + thp_nr_pages(page) - 1 > end)@@ -2111,14 +2113,11 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, put: put_page(page); next:
if (!xa_is_value(page) && PageTransHuge(page)) {unsigned int nr_pages = thp_nr_pages(page);/* Final THP may cross MAX_LFS_FILESIZE on 32-bit */xas_set(&xas, page->index + nr_pages);if (xas.xa_index < nr_pages)break;}
/* Final THP may cross MAX_LFS_FILESIZE on 32-bit */if (next_idx < xas.xa_index)break;if (next_idx != xas.xa_index + 1)xas_set(&xas, next_idx);.
On 2022/7/6 11:21, Matthew Wilcox wrote:
On Wed, Jul 06, 2022 at 11:24:34AM +0800, Liu Shixin wrote:
Release refcount after xas_set to fix UAF which may cause panic like this:
I think we can do better. How about this?
diff --git a/mm/filemap.c b/mm/filemap.c index 00e391e75880..11ae38cc4fd3 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2090,7 +2090,9 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, rcu_read_lock(); while ((page = find_get_entry(&xas, end, XA_PRESENT))) {
unsigned long next_idx = xas.xa_index + 1;next_idx = page->index + thp_nr_pages(page);
I noticed that there was a VM_BUG_ON_PAGE before which was deleted by patch 6560e8cd869b ("mm/filemap.c: remove bogus VM_BUG_ON") It seems that page->index and xas.xa_index are not guaranteed to be equal. Therefore, I conservatively retained the PageTransHuge to keep consistent with the original logic :)
@@ -2090,7 +2090,11 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start,
rcu_read_lock(); while ((page = find_get_entry(&xas, end, XA_PRESENT))) { + unsigned long next_idx = xas.xa_index; + if (!xa_is_value(page)) { + if (PageTransHuge(page)) + next_idx = page->index + thp_nr_pages(page); if (page->index < start) goto put; if (page->index + thp_nr_pages(page) - 1 > end)
if (page->index < start) goto put; if (page->index + thp_nr_pages(page) - 1 > end)@@ -2111,14 +2113,11 @@ unsigned find_lock_entries(struct address_space *mapping, pgoff_t start, put: put_page(page); next:
if (!xa_is_value(page) && PageTransHuge(page)) {unsigned int nr_pages = thp_nr_pages(page);/* Final THP may cross MAX_LFS_FILESIZE on 32-bit */xas_set(&xas, page->index + nr_pages);if (xas.xa_index < nr_pages)break;}
/* Final THP may cross MAX_LFS_FILESIZE on 32-bit */if (next_idx < xas.xa_index)break;if (next_idx != xas.xa_index + 1)xas_set(&xas, next_idx);.
linux-stable-mirror@lists.linaro.org
-
Liu Shixin -
Matthew Wilcox