[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Possible unitialised memory in FSX

From: Stefan Fuhrmann <stefan.fuhrmann_at_wandisco.com>
Date: Sun, 2 Mar 2014 22:35:59 +0100

On Sun, Mar 2, 2014 at 7:11 PM, Philip Martin <philip.martin_at_wandisco.com>wrote:

> Stefan Fuhrmann <stefan.fuhrmann_at_wandisco.com> writes:
>
> > I've got a slightly different stack but it seems to be
> > the same problem. Took me a while to trace down
> > the root cause. r1573237 should fix it.
>
> I'm still seeing one of the warnings:
>
> $ subversion/svnadmin/svnadmin create repo3 --fs-type fsx
> $ rm -rf repo3/db/format
> $ printf "1\nlayout sharded 2\n" > repo3/db/format
> $ subversion/svn/svn mkdir -mm file://`pwd`/repo3/A
>
> $ valgrind -q --suppressions=/home/pm/vg.supp
> subversion/svnadmin/.libs/lt-svnadmin pack repo3
> ==14994== Syscall param write(buf) points to uninitialised byte(s)
> ==14994== at 0x4A97110: __write_nocancel (syscall-template.S:82)
> ==14994== by 0x4A6AAF3: apr_file_flush_locked (readwrite.c:317)
> ==14994== by 0x4A6B10F: setptr (seek.c:25)
> ==14994== by 0x4A6B2E5: apr_file_seek (seek.c:70)
> ==14994== by 0x411198F: svn_io_file_seek (io.c:3553)
> ==14994== by 0x50CCA7A: write_changes_container (pack.c:1490)
> ==14994== by 0x50CD0A8: write_changes_containers (pack.c:1600)
> ==14994== by 0x50CDC84: pack_range (pack.c:1862)
> ==14994== by 0x50CE5C6: pack_log_addressed (pack.c:2040)
> ==14994== by 0x50CEA23: pack_rev_shard (pack.c:2149)
> ==14994== by 0x50CEC12: pack_shard (pack.c:2204)
> ==14994== by 0x50CF1F8: pack_body (pack.c:2333)
> ==14994== Address 0x69a6a98 is 56 bytes inside a block of size 4,096
> alloc'd
> ==14994== at 0x402C1F0: malloc (vg_replace_malloc.c:291)
> ==14994== by 0x4A6DEEB: pool_alloc (apr_pools.c:1463)
> ==14994== by 0x4A6E067: apr_palloc_debug (apr_pools.c:1504)
> ==14994== by 0x4A69573: apr_file_open (open.c:211)
> ==14994== by 0x410BB89: file_open (io.c:351)
> ==14994== by 0x41115BD: svn_io_file_open (io.c:3428)
> ==14994== by 0x50C95A6: initialize_pack_context (pack.c:261)
> ==14994== by 0x50CE35C: pack_log_addressed (pack.c:1996)
> ==14994== by 0x50CEA23: pack_rev_shard (pack.c:2149)
> ==14994== by 0x50CEC12: pack_shard (pack.c:2204)
> ==14994== by 0x50CF1F8: pack_body (pack.c:2333)
> ==14994== by 0x50DA6E8: with_some_lock_file (transaction.c:250)
> ==14994==
> Packing revisions in shard 0...done.
>
> I'm not familiar with the code, but as far as I can tell it's the line:
>
> data->len += PADDING; /* there a few extra bytes at then of the buffer
> that we want to keep */
>

r1573350 should fix that and the horribly garbled commentary.
Thanks again, Philip.

-- Stefan^2.
Received on 2014-03-02 22:36:33 CET

This is an archived mail posted to the Subversion Dev mailing list.