Skip to contentSkip to conversation

f2fs: fix deadlock allocating bio_post_read_ctx from mempool

PATCH In review Eric Biggers 2 participants 2 messages Dec 31, 2019 – Jan 2, 2020

Subsystems (2)
lore
Share & tools
lore

Eric Biggers

author maintainer
#0 lore raw

PATCH 1/1 f2fs: fix deadlock allocating bio_post_read_ctx from mempool

From: Eric Biggers <[email protected]>

Without any form of coordination, any case where multiple allocations from the same mempool are needed at a time to make forward progress can deadlock under memory pressure.

This is the case for struct bio_post_read_ctx, as one can be allocated to decrypt a Merkle tree page during fsverity_verify_bio(), which itself is running from a post-read callback for a data bio which has its own struct bio_post_read_ctx.

Fix this by freeing first bio_post_read_ctx before calling fsverity_verify_bio(). This works because verity (if enabled) is always the last post-read step.

This deadlock can be reproduced by trying to read from an encrypted verity file after reducing NUM_PREALLOC_POST_READ_CTXS to 1 and patching mempool_alloc() to pretend that pool->alloc() always fails.

Note that since NUM_PREALLOC_POST_READ_CTXS is actually 128, to actually hit this bug in practice would require reading from lots of encrypted verity files at the same time. But it's theoretically possible, as N available objects doesn't guarantee forward progress when > N/2 threads each need 2 objects at a time.

Fixes 95ae251fe828 ("f2fs: add fs-verity support")
---
 fs/f2fs/data.c | 25 +++++++++++++++++++------
 1 file changed, 19 insertions(+), 6 deletions(-)
fs/f2fs/data.c +19 -6
index 19cd03450066..618a05bf356e 100644
@@ -199,19 +199,32 @@ static void f2fs_verity_work(struct work_struct *work)
 {
 	struct bio_post_read_ctx *ctx =
 		container_of(work, struct bio_post_read_ctx, work);
+	struct bio *bio = ctx->bio;
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+	unsigned int enabled_steps = ctx->enabled_steps;
+#endif
+
+	/*
+	 * fsverity_verify_bio() may call readpages() again, and while verity
+	 * will be disabled for this, decryption may still be needed, resulting
+	 * in another bio_post_read_ctx being allocated.  So to prevent
+	 * deadlocks we need to release the current ctx to the mempool first.
+	 * This assumes that verity is the last post-read step.
+	 */
+	mempool_free(ctx, bio_post_read_ctx_pool);
+	bio->bi_private = NULL;
 
 #ifdef CONFIG_F2FS_FS_COMPRESSION
 	/* previous step is decompression */
-	if (ctx->enabled_steps & (1 << STEP_DECOMPRESS)) {
-
-		f2fs_verify_bio(ctx->bio);
-		f2fs_release_read_bio(ctx->bio);
+	if (enabled_steps & (1 << STEP_DECOMPRESS)) {
+		f2fs_verify_bio(bio);
+		f2fs_release_read_bio(bio);
 		return;
 	}
 #endif
 
-	fsverity_verify_bio(ctx->bio);
-	__f2fs_read_end_io(ctx->bio, false, false);
+	fsverity_verify_bio(bio);
+	__f2fs_read_end_io(bio, false, false);
 }
 
 static void f2fs_post_read_work(struct work_struct *work)
Signature · 1 line
2.24.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 19 lines · Eric Biggers

From: Eric Biggers <[email protected]>

Without any form of coordination, any case where multiple allocations from the same mempool are needed at a time to make forward progress can deadlock under memory pressure.

This is the case for struct bio_post_read_ctx, as one can be allocated to decrypt a Merkle tree page during fsverity_verify_bio(), which itself is running from a post-read callback for a data bio which has its own struct bio_post_read_ctx.

Fix this by freeing first bio_post_read_ctx before calling fsverity_verify_bio(). This works because verity (if enabled) is always the last post-read step.

This deadlock can be reproduced by trying to read from an encrypted verity file after reducing NUM_PREALLOC_POST_READ_CTXS to 1 and patching mempool_alloc() to pretend that pool->alloc() always fails.

Quoted 8 lines

Note that since NUM_PREALLOC_POST_READ_CTXS is actually 128, to actually hit this bug in practice would require reading from lots of encrypted verity files at the same time. But it's theoretically possible, as N available objects doesn't guarantee forward progress when > N/2 threads each need 2 objects at a time.

Fixes 95ae251fe828 ("f2fs: add fs-verity support")

Thanks,