Skip to contentSkip to conversation

block: restore the iterator when rejecting a misaligned bio

PATCH Applied Andrew Morton 2 participants 2 messages Oct 20

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 block: restore the iterator when rejecting a misaligned bio

From: Andrew Morton <[email protected]>
Subject: block: restore the iterator when rejecting a misaligned bio
Date: Sun, 18 Oct 2026 17:30:00 -0700

bio_iov_iter_get_pages() can extract some iovecs before encountering a misaligned one. It releases the pages and rejects the entire bio, but leaves the supplied iterator advanced.

An iomap direct-I/O caller may subsequently reexpand the iterator's byte count and fall back to buffered I/O. The reexpanded count then includes bytes before the iterator's current segment, potentially causing iterate_iovec() to read beyond the imported iovec array.

In a reproducer using an initial empty iovec followed by 10-byte and 1015-byte segments, writev() returned 1015 instead of 1025 and KASAN reported an 8-byte read beyond the nine-element iovec allocation. Saving and restoring the iterator on the rejected-bio path makes writev() return 1025 and eliminates the KASAN report.

Restore the entire iterator rather than reverting a byte count:
advancing past an empty segment changes its position without consuming any bytes.

Leave successful and partially accepted bio paths unchanged.

Fixes 14b007e17881 ("block: validate user space vectors during extraction")
Cc Jens Axboe <[email protected]>
Cc Keith Busch <[email protected]>
Cc Christoph Hellwig <[email protected]>
Cc Tal Zussman <[email protected]>
Cc Daehyeon Ko <[email protected]>
Cc Al Viro <[email protected]>
Cc Hannes Reinecke <[email protected]>

---
Testing on x86-64 QEMU/KVM with generic KASAN:
- Before: the ext4 writev() reproducer returned 1015 instead of 1025 and triggered an 8-byte slab out-of-bounds read.
- After: 100 consecutive reproducer runs returned 1025 without KASAN; 60-second buffered and direct ext4 fio runs passed.
- Daehyeon's alignment patch independently avoids the reproducer.

Tal Zussman's pending atomic-write changes introduce another rejected-bio cleanup path. If both patches are combined, restore the iterator there too.

 block/bio.c |    3 +++
 1 file changed, 3 insertions(+)
block/bio.c renamed from block/bio.c~block-restore-the-iterator-when-rejecting-a-misaligned-bio +3 -0
@@ -1285,6 +1285,7 @@ int bio_iov_iter_get_pages(struct bio *b
 			   unsigned mem_align_mask, unsigned len_align_mask)
 {
 	iov_iter_extraction_t flags = 0;
+	struct iov_iter saved_iter;
 
 	if (WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED)))
 		return -EIO;
@@ -1303,6 +1304,7 @@ int bio_iov_iter_get_pages(struct bio *b
 	if (bio->bi_bdev && blk_queue_pci_p2pdma(bio->bi_bdev->bd_disk->queue))
 		flags |= ITER_ALLOW_P2PDMA;
 
+	saved_iter = *iter;
 	do {
 		ssize_t ret;
 
@@ -1320,6 +1322,7 @@ int bio_iov_iter_get_pages(struct bio *b
 				bio_release_pages(bio, false);
 				bio_clear_flag(bio, BIO_PAGE_PINNED);
 				bio->bi_vcnt = 0;
+				*iter = saved_iter;
 			}
 			if (!bio->bi_vcnt)
 				return ret;

_

maintainer Applied
#1 lore raw
Quoted 10 lines · Andrew Morton

bio_iov_iter_get_pages() can extract some iovecs before encountering a misaligned one. It releases the pages and rejects the entire bio, but leaves the supplied iterator advanced.

An iomap direct-I/O caller may subsequently reexpand the iterator's byte count and fall back to buffered I/O. The reexpanded count then includes bytes before the iterator's current segment, potentially causing iterate_iovec() to read beyond the imported iovec array.

[...]

Applied, thanks!

[1/1] block: restore the iterator when rejecting a misaligned bio

      commit: 40d070cbb1906896d6fd65c39722d2411830ed9f

Best regards,

Signature · 1 line
Jens Axboe