Hi Bram!
On Mi, 13 Aug 2014, Bram Moolenaar wrote:
>
> Dominique wrote:
>
> > Bram Moolenaar wrote:
> >
> > > Patch 7.4.402
> > > Problem: Test 72 crashes under certain conditions. (Kazunobu Kuriyama)
> > > Solution: Clear the whole bufinfo_T early.
> > > Files: src/undo.c
> >
> > Test 72 no longer segfaults after this patch. However,
> > running it under valgrind shows access to uninitialized
> > memory:
> >
> > ==19752== Memcheck, a memory error detector
> > ==19752== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
> > ==19752== Using Valgrind-3.9.0.SVN and LibVEX; rerun with -h for copyright
> > info
> > ==19752== Command: ../vim -u unix.vim -U NONE --noplugin -s dotest.in
> > test72.in
> > ==19752== Parent PID: 19751
> > ==19752==
> > ==19752== Use of uninitialised value of size 8
> > ==19752== at 0x4104E9: bf_e_block (blowfish.c:360)
> > ==19752== by 0x410D1B: bf_e_cblock (blowfish.c:396)
> > ==19752== by 0x41147A: crypt_blowfish_encode (blowfish.c:618)
> > ==19752== by 0x41517B: crypt_encode (crypt.c:448)
> > ==19752== by 0x4C127D: ml_encrypt_data (memline.c:4841)
> > ==19752== by 0x5DE3AA: mf_write_block (memfile.c:1138)
> > ==19752== by 0x5DE2AC: mf_write (memfile.c:1094)
> > ==19752== by 0x5DD8E7: mf_sync (memfile.c:592)
> > ==19752== by 0x4BC8D0: ml_sync_all (memline.c:2282)
> > ==19752== by 0x4A0CE1: updatescript (getchar.c:1581)
> > ==19752== by 0x4A04C1: gotchars (getchar.c:1284)
> > ==19752== by 0x4A1DD9: vgetorpeek (getchar.c:2424)
> > ==19752== by 0x4A0DAA: vgetc (getchar.c:1637)
> > ==19752== by 0x42573D: insertchar (edit.c:6051)
> > ==19752== by 0x42528E: insert_special (edit.c:5832)
> >
> > ... many more errors...
> >
> > ==19752== Syscall param write(buf) points to uninitialised byte(s)
> > ==19752== at 0x662C040: __write_nocancel (syscall-template.S:82)
> > ==19752== by 0x498B7D: write_eintr (fileio.c:10433)
> > ==19752== by 0x5DE3D4: mf_write_block (memfile.c:1144)
> > ==19752== by 0x5DE2AC: mf_write (memfile.c:1094)
> > ==19752== by 0x5DD8E7: mf_sync (memfile.c:592)
> > ==19752== by 0x4BC8D0: ml_sync_all (memline.c:2282)
> > ==19752== by 0x4A0CE1: updatescript (getchar.c:1581)
> > ==19752== by 0x4A04C1: gotchars (getchar.c:1284)
> > ==19752== by 0x4A1DD9: vgetorpeek (getchar.c:2424)
> > ==19752== by 0x4A0DAA: vgetc (getchar.c:1637)
> > ==19752== by 0x42573D: insertchar (edit.c:6051)
> > ==19752== by 0x42528E: insert_special (edit.c:5832)
> > ==19752== by 0x41D9F3: edit (edit.c:1494)
> > ==19752== by 0x4FEA0C: invoke_edit (normal.c:9062)
> > ==19752== by 0x4FE9A5: nv_edit (normal.c:9035)
> > ==19752== Address 0x76f0510 is 4,080 bytes inside a block of size 4,096
> > alloc'd
> > ==19752== at 0x4C2C857: malloc (vg_replace_malloc.c:291)
> > ==19752== by 0x4DFEDA: lalloc (misc2.c:921)
> > ==19752== by 0x4DFDE7: alloc (misc2.c:820)
> > ==19752== by 0x4C11C0: ml_encrypt_data (memline.c:4829)
> > ==19752== by 0x5DE3AA: mf_write_block (memfile.c:1138)
> > ==19752== by 0x5DE2AC: mf_write (memfile.c:1094)
> > ==19752== by 0x5DD8E7: mf_sync (memfile.c:592)
> > ==19752== by 0x4BC8D0: ml_sync_all (memline.c:2282)
> > ==19752== by 0x4A0CE1: updatescript (getchar.c:1581)
> > ==19752== by 0x4A04C1: gotchars (getchar.c:1284)
> > ==19752== by 0x4A1DD9: vgetorpeek (getchar.c:2424)
> > ==19752== by 0x4A0DAA: vgetc (getchar.c:1637)
> > ==19752== by 0x42573D: insertchar (edit.c:6051)
> > ==19752== by 0x42528E: insert_special (edit.c:5832)
> > ==19752== by 0x41D9F3: edit (edit.c:1494)
> >
> > I assume it is a bug. In some rare cases,
> > reading uninitialized memory can be one source
> > of randomness. If that was the case, it would be
> > good to indicate it in a comment. But more likely
> > it's a bug.
>
> Turns out these valgrind warnings are already present before 7.4.399,
> once test 72 has been fixed. It was actually testing zip crypt twice.
> After applying the fix for the text I get the same errors before patch
> 7.4.399.
>
> Still haven't figured out why, they only appear twice, even when using
> blowfish twice, thus apparently it only happens the very first time Vim
> uses blowfish.
FWIW, I don't see that error with 7.4.398
Best
Christian
--
Man braucht nicht immer einen Handwerker zu bestellen. Man kann sein
Heim auch selbst ruinieren.
-- Ephraim Kishon (alias Ferenc Hoffmann)
--
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
---
You received this message because you are subscribed to the Google Groups
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
