mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] lib/lzo: check output space before writing the RLE header
@ 2026-10-16 17:09 Sean Snider
  2026-11-06 11:28 ` Herbert Xu
  0 siblings, 1 reply; 2+ messages in thread
From: Sean Snider @ 2026-10-16 17:09 UTC (permalink / raw)
  To: herbert; +Cc: davem, linux-crypto, linux-kernel

The safe LZO-RLE compressor writes its two-byte version header before
checking the available output space. With an output capacity of zero or
one, these stores exceed the supplied limit. Advancing op past op_end
also makes the unsigned HAVE_OP() comparison accept later writes.

Check for the two header bytes before emitting them. Use the existing
output_overrun path, leaving the output length unchanged on this error.
Ordinary LZO does not emit this header. Valid output is unchanged.

Bounded userspace tests of the kernel codec under ASan/UBSan reproduced
the boundary violation and passed after the correction. Successful
full-capacity streams remained byte-identical, and the existing Crypto
LZO and LZO-RLE compression vectors passed.

Fixes: cc47f07234f7 ("crypto: lzo - Fix compression buffer overrun")
Assisted-by: LLM
Signed-off-by: Sean Snider <ssnider5@outlook.com>
---
 lib/lzo/lzo1x_compress.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/lib/lzo/lzo1x_compress.c b/lib/lzo/lzo1x_compress.c
index 01586af23..621bbb969 100644
--- a/lib/lzo/lzo1x_compress.c
+++ b/lib/lzo/lzo1x_compress.c
@@ -349,6 +349,7 @@ static int LZO_SAFE(lzogeneric1x_1_compress)(
 	// LZO v0 will never write 17 as first byte (except for zero-length
 	// input), so this is used to version the bitstream
 	if (bitstream_version > 0) {
+		NEED_OP(2);
 		*op++ = 17;
 		*op++ = bitstream_version;
 		m4_max_offset = M4_MAX_OFFSET_V1;
-- 
2.43.0


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] lib/lzo: check output space before writing the RLE header
  2026-10-16 17:09 [PATCH] lib/lzo: check output space before writing the RLE header Sean Snider
@ 2026-11-06 11:28 ` Herbert Xu
  0 siblings, 0 replies; 2+ messages in thread
From: Herbert Xu @ 2026-11-06 11:28 UTC (permalink / raw)
  To: Sean Snider; +Cc: davem, linux-crypto, linux-kernel

On Fri, Oct 16, 2026 at 05:09:03PM +0000, Sean Snider wrote:
> The safe LZO-RLE compressor writes its two-byte version header before
> checking the available output space. With an output capacity of zero or
> one, these stores exceed the supplied limit. Advancing op past op_end
> also makes the unsigned HAVE_OP() comparison accept later writes.
> 
> Check for the two header bytes before emitting them. Use the existing
> output_overrun path, leaving the output length unchanged on this error.
> Ordinary LZO does not emit this header. Valid output is unchanged.
> 
> Bounded userspace tests of the kernel codec under ASan/UBSan reproduced
> the boundary violation and passed after the correction. Successful
> full-capacity streams remained byte-identical, and the existing Crypto
> LZO and LZO-RLE compression vectors passed.
> 
> Fixes: cc47f07234f7 ("crypto: lzo - Fix compression buffer overrun")
> Assisted-by: LLM
> Signed-off-by: Sean Snider <ssnider5@outlook.com>
> ---
>  lib/lzo/lzo1x_compress.c | 1 +
>  1 file changed, 1 insertion(+)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-11-06 11:28 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-10-16 17:09 [PATCH] lib/lzo: check output space before writing the RLE header Sean Snider
2026-11-06 11:28 ` Herbert Xu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®