Dominique Pellé wrote:

> Hi
>
> vim-7.4.1967 (and earlier) accesses invalid memory
> when running this command:
>
> $ valgrind --num-callers=20 vim -u NONE -c 'ec "\<C-">' -c q 2> log
>
> log file then contains:
>
> =5725== Memcheck, a memory error detector
> ==5725== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
> ==5725== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
> ==5725== Command: vim -u NONE -c ec\ "\\\<C-"\> -c q
> ==5725==
> ==5725== Invalid read of size 1
> ==5725==    at 0x414598: skipwhite (charset.c:1522)
> ==5725==    by 0x4346EE: eval7 (eval.c:5346)
> ==5725==    by 0x433EC2: eval6 (eval.c:4964)
> ==5725==    by 0x4339A4: eval5 (eval.c:4780)
> ==5725==    by 0x432C67: eval4 (eval.c:4479)
> ==5725==    by 0x432AA4: eval3 (eval.c:4396)
> ==5725==    by 0x43291C: eval2 (eval.c:4328)
> ==5725==    by 0x432753: eval1 (eval.c:4256)
> ==5725==    by 0x453969: ex_echo (eval.c:23765)
> ==5725==    by 0x475BE4: do_one_cmd (ex_docmd.c:2923)
> ==5725==    by 0x4728D2: do_cmdline (ex_docmd.c:1108)
> ==5725==    by 0x471F0E: do_cmdline_cmd (ex_docmd.c:713)
> ==5725==    by 0x5F67E3: exe_commands (main.c:2944)
> ==5725==    by 0x5F3D4B: main (main.c:959)
> ==5725==  Address 0x769ec8b is 0 bytes after a block of size 11 alloc'd
> ==5725==    at 0x4C2AB80: malloc (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==5725==    by 0x4ECEBF: lalloc (misc2.c:920)
> ==5725==    by 0x4ECD8C: alloc (misc2.c:818)
> ==5725==    by 0x4ED319: vim_strsave (misc2.c:1253)
> ==5725==    by 0x4726FC: do_cmdline (ex_docmd.c:1038)
> ==5725==    by 0x471F0E: do_cmdline_cmd (ex_docmd.c:713)
> ==5725==    by 0x5F67E3: exe_commands (main.c:2944)
> ==5725==    by 0x5F3D4B: main (main.c:959)
> (more errors after that...)
>
> Bug was found using afl-fuzz.
>
> I'm not sure whether \<C-"> in a string is valid syntax, but
> it should not cause invalid memory access.
>
> I'll try to debug it during next weekend but
> maybe someone else can fix it earlier.
>
> Regards
> Dominique


Hi

Attached is a patch that fixes the bug.
All tests pass, but I'm not sure that the fix is correct.
As far as I can tell spatial keys like <c-"> should not be allowed.
Please check.

Regards
Dominique

-- 
-- 
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.
diff --git a/src/misc2.c b/src/misc2.c
index 1018e4b..9ecf8e9 100644
--- a/src/misc2.c
+++ b/src/misc2.c
@@ -2741,8 +2741,8 @@ find_special_key(
 		else
 #endif
 		    l = 1;
-		if (bp[l + 1] == '>')
-		    bp += l;	/* anything accepted, like <C-?> */
+		if (bp[l] != '"' && bp[l + 1] == '>')
+		    bp += l;	/* anything accepted, like <C-?>, except <C-"> */
 	    }
 	}
 	if (bp[0] == 't' && bp[1] == '_' && bp[2] && bp[3])

Raspunde prin e-mail lui