]> www.infradead.org Git - users/jedix/linux-maple.git/commitdiff
fuse: break infinite loop in fuse_fill_write_pages()
authorRoman Gushchin <klamm@yandex-team.ru>
Mon, 12 Oct 2015 13:33:44 +0000 (16:33 +0300)
committerChuck Anderson <chuck.anderson@oracle.com>
Thu, 14 Apr 2016 00:47:17 +0000 (17:47 -0700)
I got a report about unkillable task eating CPU. Further
investigation shows, that the problem is in the fuse_fill_write_pages()
function. If iov's first segment has zero length, we get an infinite
loop, because we never reach iov_iter_advance() call.

Fix this by calling iov_iter_advance() before repeating an attempt to
copy data from userspace.

A similar problem is described in 124d3b7041f ("fix writev regression:
pan hanging unkillable and un-straceable"). If zero-length segmend
is followed by segment with invalid address,
iov_iter_fault_in_readable() checks only first segment (zero-length),
iov_iter_copy_from_user_atomic() skips it, fails at second and
returns zero -> goto again without skipping zero-length segment.

Patch calls iov_iter_advance() before goto again: we'll skip zero-length
segment at second iteraction and iov_iter_fault_in_readable() will detect
invalid address.

Special thanks to Konstantin Khlebnikov, who helped a lot with the commit
description.

Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Maxim Patlasov <mpatlasov@parallels.com>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: Roman Gushchin <klamm@yandex-team.ru>
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Fixes: ea9b9907b82a ("fuse: implement perform_write")
Cc: <stable@vger.kernel.org>
(cherry picked from commit 3ca8138f014a913f98e6ef40e939868e1e9ea876)

Orabug: 22673138
CVE: CVE-2015-8785

Signed-off-by: Manjunath Govindashetty <manjunath.govindashetty@oracle.com>
fs/fuse/file.c

index ea796dbcea2fac5d94ed6676b326ef9e1fb9fa7c..46d7242b568a3c28c98de0d774efd814231c9f33 100644 (file)
@@ -1049,6 +1049,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req,
                tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes);
                flush_dcache_page(page);
 
+               iov_iter_advance(ii, tmp);
                if (!tmp) {
                        unlock_page(page);
                        page_cache_release(page);
@@ -1061,7 +1062,6 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req,
                req->page_descs[req->num_pages].length = tmp;
                req->num_pages++;
 
-               iov_iter_advance(ii, tmp);
                count += tmp;
                pos += tmp;
                offset += tmp;