The pool-resize path frees each stream's old decoder before allocating
its replacement. If an allocation fails after some streams have already
been replaced, the failed stream is put back on the list with state ==
NULL. z_erofs_lzma_max_dictsize is still advanced as if the whole
pool had been resized.

An existing LZMA mount can select the broken stream and pass
NULL to xz_dec_microlzma_reset(). A retry at the same size also
skip another resize attempt. Since the global maximum was advanced,
thus, the invalid state is left unrepaired.

Allocate each replacement before freeing the old decoder, temporarily
retaining one old decoder during allocation. Stop at the first failure
and advance z_erofs_lzma_max_dictsize only after all streams satisfy
the request.

Record each stream's dictionary capacity so retries can skip streams
already enlarged before a partial failure.

Fixes: 622ceaddb764 ("erofs: lzma compression support")
Cc: stable@vger.kernel.org
Signed-off-by: Nikhil Gurudasani <nikhilgurudasani314@gmail.com>
---
 fs/erofs/decompressor_lzma.c | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/fs/erofs/decompressor_lzma.c b/fs/erofs/decompressor_lzma.c
index 6b0cdb446c6a..9d15f94cbee1 100644
--- a/fs/erofs/decompressor_lzma.c
+++ b/fs/erofs/decompressor_lzma.c
@@ -5,6 +5,7 @@
 struct z_erofs_lzma {
 	struct z_erofs_lzma *next;
 	struct xz_dec_microlzma *state;
+	unsigned int dict_size;
 	u8 bounce[PAGE_SIZE];
 };
 
@@ -128,11 +129,19 @@ static int z_erofs_load_lzma_config(struct super_block *sb,
 	err = 0;
 	/* 2. walk each isolated stream and grow max dict_size if needed */
 	for (strm = head; strm; strm = strm->next) {
+		struct xz_dec_microlzma *state;
+
+		if (strm->dict_size >= dict_size)
+			continue;
+		state = xz_dec_microlzma_alloc(XZ_PREALLOC, dict_size);
+		if (!state) {
+			err = -ENOMEM;
+			break;
+		}
 		if (strm->state)
 			xz_dec_microlzma_end(strm->state);
-		strm->state = xz_dec_microlzma_alloc(XZ_PREALLOC, dict_size);
-		if (!strm->state)
-			err = -ENOMEM;
+		strm->state = state;
+		strm->dict_size = dict_size;
 	}
 
 	/* 3. push back all to the global list and update max dict_size */
@@ -142,7 +151,8 @@ static int z_erofs_load_lzma_config(struct super_block *sb,
 	spin_unlock(&z_erofs_lzma_lock);
 	wake_up_all(&z_erofs_lzma_wq);
 
-	z_erofs_lzma_max_dictsize = dict_size;
+	if (!err)
+		z_erofs_lzma_max_dictsize = dict_size;
 	mutex_unlock(&lzma_resize_mutex);
 	return err;
 }

base-commit: a7d28aa0e9b2c983b915d091f9596f0e3b253355
-- 
2.43.0

On Sun, Aug 30, 2026 at 04:11:09PM +0530, Nikhil Gurudasani wrote:
> The pool-resize path frees each stream's old decoder before allocating
> its replacement. If an allocation fails after some streams have already
> been replaced, the failed stream is put back on the list with state ==
> NULL. z_erofs_lzma_max_dictsize is still advanced as if the whole
> pool had been resized.
> 
> An existing LZMA mount can select the broken stream and pass
> NULL to xz_dec_microlzma_reset(). A retry at the same size also
> skip another resize attempt. Since the global maximum was advanced,
> thus, the invalid state is left unrepaired.
> 
> Allocate each replacement before freeing the old decoder, temporarily
> retaining one old decoder during allocation. Stop at the first failure
> and advance z_erofs_lzma_max_dictsize only after all streams satisfy
> the request.
> 
> Record each stream's dictionary capacity so retries can skip streams
> already enlarged before a partial failure.
> 
> Fixes: 622ceaddb764 ("erofs: lzma compression support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Nikhil Gurudasani <nikhilgurudasani314@gmail.com>

Reviewed-by: Gao Xiang <xiang@kernel.org>

Thanks,
Gao Xiang