[BUG]
When btrfs-progs is compiled with D=asan, it can't pass even the very
basic fsck tests due to btrfs-image has memory leak:
=== START TEST /home/adam/btrfs/btrfs-progs/tests//fsck-tests/001-bad-file-extent-bytenr
restoring image default_case.img
=================================================================
==7790==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 104 byte(s) in 1 object(s) allocated from:
#0 0x7f1d3b738389 in __interceptor_malloc /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cc:86
#1 0x560ca6b7f4ff in btrfs_add_delayed_tree_ref /home/adam/btrfs/btrfs-progs/delayed-ref.c:569
#2 0x560ca6af2d0b in btrfs_free_extent /home/adam/btrfs/btrfs-progs/extent-tree.c:2155
#3 0x560ca6ac16ca in __btrfs_cow_block /home/adam/btrfs/btrfs-progs/ctree.c:319
#4 0x560ca6ac1d8c in btrfs_cow_block /home/adam/btrfs/btrfs-progs/ctree.c:383
#5 0x560ca6ac6c8e in btrfs_search_slot /home/adam/btrfs/btrfs-progs/ctree.c:1153
#6 0x560ca6ab7e83 in fixup_device_size image/main.c:2113
#7 0x560ca6ab9279 in fixup_chunks_and_devices image/main.c:2333
#8 0x560ca6ab9ada in restore_metadump image/main.c:2455
#9 0x560ca6abaeba in main image/main.c:2723
#10 0x7f1d3b148ce2 in __libc_start_main (/usr/lib/libc.so.6+0x23ce2)
... tons of similar leakage for delayed_tree_ref ...
Direct leak of 96 byte(s) in 1 object(s) allocated from:
#0 0x7f1d3b738389 in __interceptor_malloc /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cc:86
#1 0x560ca6b7f5fb in btrfs_add_delayed_tree_ref /home/adam/btrfs/btrfs-progs/delayed-ref.c:583
#2 0x560ca6af5679 in alloc_tree_block /home/adam/btrfs/btrfs-progs/extent-tree.c:2503
#3 0x560ca6af57ac in btrfs_alloc_free_block /home/adam/btrfs/btrfs-progs/extent-tree.c:2524
#4 0x560ca6ac115b in __btrfs_cow_block /home/adam/btrfs/btrfs-progs/ctree.c:290
#5 0x560ca6ac1d8c in btrfs_cow_block /home/adam/btrfs/btrfs-progs/ctree.c:383
#6 0x560ca6b7bb15 in commit_tree_roots /home/adam/btrfs/btrfs-progs/transaction.c:98
#7 0x560ca6b7c525 in btrfs_commit_transaction /home/adam/btrfs/btrfs-progs/transaction.c:192
#8 0x560ca6ab92be in fixup_chunks_and_devices image/main.c:2337
#9 0x560ca6ab9ada in restore_metadump image/main.c:2455
#10 0x560ca6abaeba in main image/main.c:2723
#11 0x7f1d3b148ce2 in __libc_start_main (/usr/lib/libc.so.6+0x23ce2)
... tons of similar leakage for delayed_ref_head ...
SUMMARY: AddressSanitizer: 1600 byte(s) leaked in 16 allocation(s).
failed to restore image ./default_case.img
[CAUSE]
Commit c6039704c5 ("btrfs-progs: Add delayed refs infrastructure")
introduces delayed ref infrastructure for free space tree, however the
refcount_dec_and_test() from kernel code is wrongly backported.
refcount_dec_and_test() will return true if the refcount reaches 0.
So kernel code will free the allocated space as expected:
if (refcount_dec_and_test(&ref->refs)) {
kmem_cache_free();
}
However btrfs-progs backport is using the opposite condition:
if (--ref->refs) {
kfree();
}
This will not free the memory for the last user, but for refs >= 2.
Causing both use-after-free and memory leak for any offline write
operation.
[FIX]
Fix the (--ref->refs) condition to (--ref->refs == 0) to fix the
backport error.
Fixes: c6039704c5 ("btrfs-progs: Add delayed refs infrastructure")
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
[BUG]
When running fuzz-tests/003 and fuzz-tests/009, btrfs-progs will crash
due to BUG_ON().
[CAUSE]
We abused BUG_ON() in btrfs_commit_transaction(), which is one of the
most error prone function for fuzzed images.
Currently to cleanup the aborted transaction, we only need to clean up
the only per-transaction data: delayed refs.
This patch will introduce a new function, btrfs_destroy_delayed_refs()
to cleanup delayed refs when we failed to commit transaction.
With that function, we will gently destroy per-trans delayed ref, and
remove the BUG_ON()s in btrfs_commit_transaction().
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This commit pulls those portions of the kernel implementation of
delayed refs which are necessary to have them working in user-space.
I've done the following modifications:
1. Replaced all kmem_cache_alloc calls to kmalloc.
2. Removed all locking-related code, since we are single threaded in
userspace.
3. Removed code which deals with data refs - delayed refs in user space
are going to be used only for cowonly trees.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
