git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH v2 06/12] reftable/block: fix OOB write with bogus inflated log size

From
Patrick Steinhardt <ps@pks.im>
Date
Jul 3, 2026, 10:32 UTC
Message-ID
<akePvoX0j64cRnos@pks.im>
In-Reply-To
<871pdk4eau.fsf@emacs.iotcl.com>
On Fri, Jul 03, 2026 at 11:23:37AM +0200, Toon Claes wrote:
Show 53 quoted lines
> Patrick Steinhardt <ps@pks.im> writes:
> 
> > diff --git a/t/unit-tests/u-reftable-block.c b/t/unit-tests/u-reftable-block.c
> > index f4e926ce3a..088162483e 100644
> > --- a/t/unit-tests/u-reftable-block.c
> > +++ b/t/unit-tests/u-reftable-block.c
> > @@ -465,3 +465,35 @@ void test_reftable_block__iterator(void)
> >  	reftable_block_release(&block);
> >  	reftable_buf_release(&data);
> >  }
> > +
> > +void test_reftable_block__corrupt_log_block_size(void)
> > +{
> > +	struct reftable_block_source source = { 0 };
> > +	struct reftable_record rec = {
> > +		.type = REFTABLE_BLOCK_TYPE_LOG,
> > +		.u.log = {
> > +			.refname = (char *) "refs/heads/main",
> > +			.update_index = 1,
> > +			.value_type = REFTABLE_LOG_UPDATE,
> > +		},
> > +	};
> > +	struct reftable_block block = { 0 };
> > +	struct reftable_buf data = REFTABLE_BUF_INIT;
> > +
> > +	cl_reftable_write_block(&data, REFTABLE_BLOCK_TYPE_LOG, &rec, 1);
> > +
> > +	/*
> > +	 * Log blocks store their inflated size as a big-endian 24-bit integer
> > +	 * right after the one-byte block type. Rewrite it to claim a size that
> > +	 * is smaller than the block header.
> > +	 */
> > +	reftable_put_be24((uint8_t *) data.buf + 1, 1);
> 
> Can I suggest to make this test a bit more strict:
> 
> 	/*
> 	 * Log blocks store their inflated size as a big-endian 24-bit integer
> 	 * right after the one-byte block type.
> 	 * First sanity check if the expected value is actually there,
> 	 * then rewrite it to claim a size that is smaller than the block header.
> 	 */
> 	void *p = (uint8_t *) data.buf + 1;
> 	uint32_t block_size = reftable_get_be24(p);
> 	cl_assert_equal_i(block_size, 82);
> 	reftable_put_be24(p, 5);
> 
> So first do a sanity check to see if the current block_size is what we
> expect, this helps us ensure we are actually reading the correct point
> in the buffers. Or does this depend on the architecture?
> 
> Then write a size that's just below what the expected value is. By using
> a value that's just too small, maybe we can avoid off-by-one errors?

That would cause a failure, but it wouldn't catch the out-of-bounds write that the current test catches. This is because the offset would still point into the buffer, even though it would point at the wrong location. The requirement here is that the block size must be smaller than the header size itself, so it must be between 1 and 3.

So in this case I think it doesn't quite help us to assert the written block size.

Patrick
Previous: Toon ClaesNext: Patrick Steinhardt
Message 26 of 48 in “reftable: harden against corrupted tables”
  1. 00/11 reftable: harden against corrupted tablesPatrick Steinhardt, Jun 24, 2026
  2. 01/11 meson: support building fuzzers with libFuzzerPatrick Steinhardt, Jun 24, 2026
  3. 02/11 oss-fuzz: add fuzzer for parsing reftablesPatrick Steinhardt, Jun 24, 2026
  4. 03/11 reftable/basics: fix OOB read on binary search of empty rangePatrick Steinhardt, Jun 24, 2026
  5. 04/11 reftable/record: don't abort when decoding invalid ref value typePatrick Steinhardt, Jun 24, 2026
  6. 05/11 reftable/block: fix OOB write with bogus inflated log sizePatrick Steinhardt, Jun 24, 2026
  7. Christian CouderJun 26, 2026
  8. Patrick SteinhardtJun 29, 2026
  9. 06/11 reftable/block: fix OOB read with bogus block sizePatrick Steinhardt, Jun 24, 2026
  10. 07/11 reftable/block: fix OOB read with bogus restart countPatrick Steinhardt, Jun 24, 2026
  11. 08/11 reftable/block: fix use of uninitialized memory when binsearch failsPatrick Steinhardt, Jun 24, 2026
  12. 09/11 reftable/block: fix OOB read with bogus restart offsetPatrick Steinhardt, Jun 24, 2026
  13. 10/11 reftable/table: fix NULL pointer access when seeking to bogus offsetsPatrick Steinhardt, Jun 24, 2026
  14. 11/11 reftable/table: fix OOB read on truncated tablePatrick Steinhardt, Jun 24, 2026
  15. 00/12 reftable: harden against corrupted tablesPatrick Steinhardt, Jun 29, 2026
  16. 01/12 meson: support building fuzzers with libFuzzerPatrick Steinhardt, Jun 29, 2026
  17. 02/12 oss-fuzz: add fuzzer for parsing reftablesPatrick Steinhardt, Jun 29, 2026
  18. 03/12 reftable/basics: fix OOB read on binary search of empty rangePatrick Steinhardt, Jun 29, 2026
  19. 04/12 reftable/record: don't abort when decoding invalid ref value typePatrick Steinhardt, Jun 29, 2026
  20. 05/12 t/unit-tests: introduce test helper to write reftable blocksPatrick Steinhardt, Jun 29, 2026
  21. Junio C HamanoJun 30, 2026
  22. Christian CouderJul 2, 2026
  23. Junio C HamanoJul 3, 2026
  24. 06/12 reftable/block: fix OOB write with bogus inflated log sizePatrick Steinhardt, Jun 29, 2026
  25. Toon ClaesJul 3, 2026
  26. Patrick SteinhardtJul 3, 2026
  27. 07/12 reftable/block: fix OOB read with bogus block sizePatrick Steinhardt, Jun 29, 2026
  28. Toon ClaesJul 3, 2026
  29. Patrick SteinhardtJul 3, 2026
  30. 08/12 reftable/block: fix OOB read with bogus restart countPatrick Steinhardt, Jun 29, 2026
  31. 09/12 reftable/block: fix use of uninitialized memory when binsearch failsPatrick Steinhardt, Jun 29, 2026
  32. 10/12 reftable/block: fix OOB read with bogus restart offsetPatrick Steinhardt, Jun 29, 2026
  33. 11/12 reftable/table: fix NULL pointer access when seeking to bogus offsetsPatrick Steinhardt, Jun 29, 2026
  34. 12/12 reftable/table: fix OOB read on truncated tablePatrick Steinhardt, Jun 29, 2026
  35. 00/12 reftable: harden against corrupted tablesPatrick Steinhardt, Jul 3, 2026
  36. 01/12 meson: support building fuzzers with libFuzzerPatrick Steinhardt, Jul 3, 2026
  37. 02/12 oss-fuzz: add fuzzer for parsing reftablesPatrick Steinhardt, Jul 3, 2026
  38. 03/12 reftable/basics: fix OOB read on binary search of empty rangePatrick Steinhardt, Jul 3, 2026
  39. 04/12 reftable/record: don't abort when decoding invalid ref value typePatrick Steinhardt, Jul 3, 2026
  40. 05/12 t/unit-tests: introduce test helper to write reftable blocksPatrick Steinhardt, Jul 3, 2026
  41. 06/12 reftable/block: fix OOB write with bogus inflated log sizePatrick Steinhardt, Jul 3, 2026
  42. 07/12 reftable/block: fix OOB read with bogus block sizePatrick Steinhardt, Jul 3, 2026
  43. 08/12 reftable/block: fix OOB read with bogus restart countPatrick Steinhardt, Jul 3, 2026
  44. 09/12 reftable/block: fix use of uninitialized memory when binsearch failsPatrick Steinhardt, Jul 3, 2026
  45. 10/12 reftable/block: fix OOB read with bogus restart offsetPatrick Steinhardt, Jul 3, 2026
  46. 11/12 reftable/table: fix NULL pointer access when seeking to bogus offsetsPatrick Steinhardt, Jul 3, 2026
  47. 12/12 reftable/table: fix OOB read on truncated tablePatrick Steinhardt, Jul 3, 2026
  48. Toon ClaesJul 8, 2026

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.