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

Re: git fsck failure on OS X with files >= 4 GiB

From
Junio C Hamano <gitster@pobox.com>
Date
Oct 29, 2015, 17:19 UTC
Message-ID
<xmqqlhalsict.fsf@gitster.mtv.corp.google.com>
In-Reply-To
<CA+izobtdwszVrYsnKU=_ytLuNbPGyRe_7kXqyrQO7u5Lo+OdPg@mail.gmail.com>
Atousa Duprat <atousa.p@gmail.com> writes:
Show 40 quoted lines
> [PATCH] Limit the size of the data block passed to SHA1_Update()
>
> This avoids issues where OS-specific implementations use
> a 32-bit integer to specify block size.  Limit currently
> set to 1GiB.
> ---
>  cache.h | 20 +++++++++++++++++++-
>  1 file changed, 19 insertions(+), 1 deletion(-)
>
> diff --git a/cache.h b/cache.h
> index 79066e5..c305985 100644
> --- a/cache.h
> +++ b/cache.h
> @@ -14,10 +14,28 @@
>  #ifndef git_SHA_CTX
>  #define git_SHA_CTX SHA_CTX
>  #define git_SHA1_Init SHA1_Init
> -#define git_SHA1_Update SHA1_Update
>  #define git_SHA1_Final SHA1_Final
>  #endif
>
> +#define SHA1_MAX_BLOCK_SIZE (1024*1024*1024)
> +
> +static inline int git_SHA1_Update(SHA_CTX *c, const void *data, size_t len)
> +{
> + size_t nr;
> + size_t total = 0;
> + char *cdata = (char*)data;
> + while(len > 0) {
> + nr = len;
> + if(nr > SHA1_MAX_BLOCK_SIZE)
> + nr = SHA1_MAX_BLOCK_SIZE;
> + SHA1_Update(c, cdata, nr);
> + total += nr;
> + cdata += nr;
> + len -= nr;
> + }
> + return total;
> +}
> +

I think the idea illustrated above is a good start, but there are a few issues:

 * SHA1_Update() is used in fairly many places; it is unclear if it
   is a good idea to inline.
 * There is no need to punish implementations with working
   SHA1_Update by another level of wrapping.
 * What would you do when you find an implementation for which 1G is
   still too big?
Perhaps something like this in the header

#ifdef SHA1_MAX_BLOCK_SIZE extern int SHA1_Update_Chunked(SHA_CTX *, const void *, size_t); #define git_SHA1_Update SHA1_Update_Chunked #endif

with compat/sha1_chunked.c that has
#ifdef SHA1_MAX_BLOCK_SIZE
int SHA1_Update_Chunked(SHA_CTX *c, const void *data, size_t len)
{
	... your looping implementation ...
}
#endif

in it, that is only triggered via a Makefile macro, e.g. might be a good workaround.

diff --git a/Makefile b/Makefile
index 8466333..83348b8 100644
--- a/Makefile
+++ b/Makefile
@@ -139,6 +139,10 @@ all::
 # Define PPC_SHA1 environment variable when running make to make use of
 # a bundled SHA1 routine optimized for PowerPC.
 #
+# Define SHA1_MAX_BLOCK_SIZE if your SSH1_Update() implementation can
+# hash only a limited amount of data in one call (e.g. APPLE_COMMON_CRYPTO
+# may want 'SHA1_MAX_BLOCK_SIZE=1024L*1024L*1024L' defined).
+#
 # Define NEEDS_CRYPTO_WITH_SSL if you need -lcrypto when using -lssl (Darwin).
 #
 # Define NEEDS_SSL_WITH_CRYPTO if you need -lssl when using -lcrypto (Darwin).
@@ -1002,6 +1006,7 @@ ifeq ($(uname_S),Darwin)
 	ifndef NO_APPLE_COMMON_CRYPTO
 		APPLE_COMMON_CRYPTO = YesPlease
 		COMPAT_CFLAGS += -DAPPLE_COMMON_CRYPTO
+		SHA1_MAX_BLOCK_SIZE=1024L*1024L*1024L
 	endif
 	NO_REGEX = YesPlease
 	PTHREAD_LIBS =
@@ -1350,6 +1355,11 @@ endif
 endif
 endif
 
+ifdef SHA1_MAX_BLOCK_SIZE
+LIB_OBJS += compat/sha1_chunked.o
+BASIC_CFLAGS += SHA1_MAX_BLOCK_SIZE="$(SHA1_MAX_BLOCK_SIZE)"
+endif
+
 ifdef NO_PERL_MAKEMAKER
 	export NO_PERL_MAKEMAKER
 endif
Previous: Atousa DupratNext: Atousa Duprat
Message 6 of 27 in “git fsck failure on OS X with files >= 4 GiB”
  1. Rafael EspíndolaOct 28, 2015
  2. Filipe CabecinhasOct 29, 2015
  3. Rafael EspíndolaOct 29, 2015
  4. Filipe CabecinhasOct 29, 2015
  5. Atousa DupratOct 29, 2015
  6. Junio C HamanoOct 29, 2015
  7. Atousa DupratOct 30, 2015
  8. Limit the size of the data block passed to SHA1_Update()Atousa Pahlevan Duprat, Oct 30, 2015
  9. Junio C HamanoOct 30, 2015
  10. Atousa DupratNov 1, 2015
  11. Junio C HamanoNov 1, 2015
  12. Eric SunshineNov 1, 2015
  13. Limit the size of the data block passed to SHA1_Update()atousa.p@gmail.com, Nov 1, 2015
  14. Eric SunshineNov 1, 2015
  15. Junio C HamanoNov 1, 2015
  16. Atousa DupratNov 2, 2015
  17. Junio C HamanoNov 2, 2015
  18. 1/2 Limit the size of the data block passed to SHA1_Update()atousa.p@gmail.com, Nov 3, 2015
  19. Torsten BögershausenNov 3, 2015
  20. Limit the size of the data block passed to SHA1_Update()atousa.p@gmail.com, Nov 4, 2015
  21. Eric SunshineNov 4, 2015
  22. Atousa DupratNov 4, 2015
  23. Junio C HamanoNov 4, 2015
  24. Limit the size of the data block passed to SHA1_Update()Atousa Pahlevan Duprat, Oct 30, 2015
  25. Randall S. BeckerOct 30, 2015
  26. Junio C HamanoOct 31, 2015
  27. Atousa DupratNov 1, 2015

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.