patch 9.0.2121: [security]: use-after-free in ex_substitute
Commit:
https://github.com/vim/vim/commit/26c11c56888d01e298cd8044caf860f3c26f57bb
Author: Christian Brabandt <c...@256bit.org>
Date: Wed Nov 22 21:26:41 2023 +0100
patch 9.0.2121: [security]: use-after-free in ex_substitute
Problem: [security]: use-after-free in ex_substitute
Solution: always allocate memory
closes: #13552
A recursive :substitute command could cause a heap-use-after free in Vim
(CVE-2023-48706).
The whole reproducible test is a bit tricky, I can only reproduce this
reliably when no previous substitution command has been used yet
(which is the reason, the test needs to run as first one in the
test_substitute.vim file) and as a combination of the `:~` command
together with a :s command that contains the special substitution atom `~\=`
which will make use of a sub-replace special atom and calls a vim script
function.
There was a comment in the existing :s code, that already makes the
`sub` variable allocate memory so that a recursive :s call won't be able
to cause any issues here, so this was known as a potential problem
already. But for the current test-case that one does not work, because
the substitution does not start with `\=` but with `~\=` (and since
there does not yet exist a previous substitution atom, Vim will simply
increment the `sub` pointer (which then was not allocated dynamically)
and later one happily use a sub-replace special expression (which could
then free the `sub` var).
The following commit fixes this, by making the sub var always using
allocated memory, which also means we need to free the pointer whenever
we leave the function. Since sub is now always an allocated variable,
we also do no longer need the sub_copy variable anymore, since this one
was used to indicated when sub pointed to allocated memory (and had
therefore to be freed on exit) and when not.
Github Security Advisory:
https://github.com/vim/vim/security/advisories/GHSA-c8qm-x72m-q53q
Signed-off-by: Christian Brabandt <c...@256bit.org>
diff --git a/src/ex_cmds.c b/src/ex_cmds.c
index c5f912e7e..a08682b07 100644
--- a/src/ex_cmds.c
+++ b/src/ex_cmds.c
@@ -3737,13 +3737,13 @@ ex_substitute(exarg_T *eap)
int save_do_all; // remember user specified 'g'
flag
int save_do_ask; // remember user specified 'c'
flag
char_u *pat = NULL, *sub = NULL; // init for GCC
- char_u *sub_copy = NULL;
int delimiter;
int sublen;
int got_quit = FALSE;
int got_match = FALSE;
int which_pat;
char_u *cmd;
+ char_u *p;
int save_State;
linenr_T first_line = 0; // first changed line
linenr_T last_line= 0; // below last changed line AFTER the
@@ -3827,8 +3827,12 @@ ex_substitute(exarg_T *eap)
* Small incompatibility: vi sees '
' as end of the command, but in
* Vim we want to use '
' to find/substitute a NUL.
*/
- sub = cmd; // remember the start of the substitution
+ p = cmd; // remember the start of the substitution
cmd = skip_substitute(cmd, delimiter);
+ sub = vim_strsave(p);
+ if (sub == NULL)
+ // out of memory
+ return;
if (!eap->skip)
{
@@ -3839,14 +3843,22 @@ ex_substitute(exarg_T *eap)
if (old_sub == NULL) // there is no previous command
{
emsg(_(e_no_previous_substitute_regular_expression));
+ vim_free(sub);
return;
}
- sub = old_sub;
+ vim_free(sub);
+ sub = vim_strsave(old_sub);
+ if (sub == NULL)
+ // out of memory
+ return;
}
else
{
vim_free(old_sub);
old_sub = vim_strsave(sub);
+ if (old_sub == NULL)
+ // out of memory
+ return;
}
}
}
@@ -3858,7 +3870,7 @@ ex_substitute(exarg_T *eap)
return;
}
pat = NULL; // search_regcomp() will use previous pattern
- sub = old_sub;
+ sub = vim_strsave(old_sub);
// Vi compatibility quirk: repeating with ":s" keeps the cursor in the
// last column after using "$".
@@ -3877,7 +3889,10 @@ ex_substitute(exarg_T *eap)
linenr_T joined_lines_count;
if (eap->skip)
+ {
+ vim_free(sub);
return;
+ }
curwin->w_cursor.lnum = eap->line1;
if (*cmd == 'l')
eap->flags = EXFLAG_LIST;
@@ -3904,6 +3919,7 @@ ex_substitute(exarg_T *eap)
save_re_pat(RE_SUBST, pat, magic_isset());
// put pattern in history
add_to_history(HIST_SEARCH, pat, TRUE, NUL);
+ vim_free(sub);
return;
}
@@ -3991,6 +4007,7 @@ ex_substitute(exarg_T *eap)
if (i <= 0 && !eap->skip && subflags.do_error)
{
emsg(_(e_positive_count_required));
+ vim_free(sub);
return;
}
else if (i >= INT_MAX)
@@ -3998,6 +4015,7 @@ ex_substitute(exarg_T *eap)
char buf[20];
vim_snprintf(buf, sizeof(buf), "%ld", i);
semsg(_(e_val_too_large), buf);
+ vim_free(sub);
return;
}
eap->line1 = eap->line2;
@@ -4016,17 +4034,22 @@ ex_substitute(exarg_T *eap)
if (eap->nextcmd == NULL)
{
semsg(_(e_trailing_characters_str), cmd);
+ vim_free(sub);
return;
}
}
if (eap->skip) // not executing commands, only parsing
+ {
+ vim_free(sub);
return;
+ }
if (!subflags.do_count && !curbuf->b_p_ma)
{
// Substitution is not allowed in non-'modifiable' buffer
emsg(_(e_cannot_make_changes_modifiable_is_off));
+ vim_free(sub);
return;
}
@@ -4034,6 +4057,7 @@ ex_substitute(exarg_T *eap)
{
if (subflags.do_error)
emsg(_(e_invalid_command));
+ vim_free(sub);
return;
}
@@ -4054,20 +4078,20 @@ ex_substitute(exarg_T *eap)
*/
if (sub[0] == '\' && sub[1] == '=')
{
- sub = vim_strsave(sub);
- if (sub == NULL)
+ p = vim_strsave(sub);
+ vim_free(sub);
+ if (p == NULL)
return;
- sub_copy = sub;
+ sub = p;
}
else
{
- char_u *newsub = regtilde(sub, magic_isset());
+ p = regtilde(sub, magic_isset());
- if (newsub != sub)
+ if (p != sub)
{
- // newsub was allocated, free it later.
- sub_copy = newsub;
- sub = newsub;
+ vim_free(sub);
+ sub = p;
}
}
@@ -4965,7 +4989,7 @@ outofmem:
#endif
vim_regfree(regmatch.regprog);
- vim_free(sub_copy);
+ vim_free(sub);
// Restore the flag values, they can be used for ":&&".
subflags.do_all = save_do_all;
diff --git a/src/testdir/test_substitute.vim b/src/testdir/test_substitute.vim
index 3ed159799..7c2bbb476 100644
--- a/src/testdir/test_substitute.vim
+++ b/src/testdir/test_substitute.vim
@@ -4,6 +4,32 @@ source shared.vim
source check.vim
source screendump.vim
+" NOTE: This needs to be the first test to be
+" run in the file, since it depends on
+" that the previous substitution atom
+" was not yet set.
+"
+" recursive call of :s and sub-replace special
+" (did cause heap-use-after free in < v9.0.2121)
+func Test_aaaa_substitute_expr_recursive_special()
+ func R()
+ " FIXME: leaving out the 'n' flag leaks memory, why?
+ %s/./\='.'/gn
+ endfunc
+ new Xfoobar_UAF
+ put ='abcdef'
+ let bufnr = bufnr('%')
+ try
+ silent! :s/./~\=R()/0
+ "call assert_fails(':s/./~\=R()/0', 'E939:')
+ let @/='.'
+ ~g
+ catch /^Vim\%(( \+)\)\=:E565:/
+ endtry
+ delfunc R
+ exe bufnr .. "bw!"
+endfunc
+
func Test_multiline_subst()
enew!
call append(0, ["1 aa",
@@ -147,7 +173,6 @@ func Test_substitute_repeat()
call feedkeys("Qsc\<CR>y", 'tx')
bwipe!
endfunc
-
" Test %s/
// which is implemented as a special case to use a
" more efficient join rather than doing a regular substitution.
func Test_substitute_join()
@@ -1447,11 +1472,30 @@ func Test_substitute_expr_switch_win()
endfunc
new Xfoobar
let bufnr = bufnr('%')
- put ="abcdef"
+ put ='abcdef'
silent! s/\%')/\=R()
call assert_fails(':%s/./\=R()/g', 'E565:')
delfunc R
exe bufnr .. "bw!"
endfunc
+" recursive call of :s using test-replace special
+func Test_substitute_expr_recursive()
+ func Q()
+ %s/./\='foobar'/gn
+ return "foobar"
+ endfunc
+ func R()
+ %s/./\=Q()/g
+ endfunc
+ new Xfoobar_UAF
+ let bufnr = bufnr('%')
+ put ='abcdef'
+ silent! s/./\=R()/g
+ call assert_fails(':%s/./\=R()/g', 'E565:')
+ delfunc R
+ delfunc Q
+ exe bufnr .. "bw!"
+endfunc
+
" vim: shiftwidth=2 sts=2 expandtab
diff --git a/src/version.c b/src/version.c
index edf6b576f..642da01d8 100644
--- a/src/version.c
+++ b/src/version.c
@@ -704,6 +704,8 @@ static char *(features[]) =
static int included_patches[] =
{ /* Add new patch number below this line */
+/**/
+ 2121,
/**/
2120,
/**/
--
--
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 vim_dev+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/vim_dev/E1r5uYr-009D8B-9y%40256bit.org.
