Bug #72115 [Opn]: imagexbm read out of bounds
| From: | fernando at null-life dot com | Date: | Thu, 28 Apr 2016 03:19:14 +0000 |
| Subject: | Bug #72115 [Opn]: imagexbm read out of bounds | ||
| References: | 1 | Groups: | php.bugs |
| Request: | Send a blank email to php-bugs+get-200799@lists.php.net to get a copy of this message | ||
Edit report at https://bugs.php.net/bug.php?id=72115&edit=1
ID: 72115
User updated by: fernando at null-life dot com
Reported by: fernando at null-life dot com
Summary: imagexbm read out of bounds
Status: Open
Type: Bug
Package: GD related
Operating System: Linux
PHP Version: 5.5.34
Block user comment: N
Private report: N
New Comment:
I believe this could be a security bug if the file name could be controlled by users since this
leaks environment variables and other stuff from the memory.
Previous Comments:
------------------------------------------------------------------------
[2016-04-28 03:17:32] fernando at null-life dot com
Description:
------------
Compile PHP with ASAN. Only affects PHP 5.5.34 since https://bugs.php.net/bug.php?id=66339 reported by
me too and patched on 5.6 prevents me from triggering it.
BP on https://github.com/php/php-src/blob/PHP-5.5/ext/gd/gd_ctx.c#L39
(gdb) r
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /home/user/php/php-55/sapi/cli/php -n
-dextension=/home/user/php/php-55/modules/gd.so /home/user/img.php
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/i386-linux-gnu/libthread_db.so.1".
Breakpoint 1, _php_image_output_putbuf (ctx=0xb46ddf38, buf=0xbfffa69c, l=8017) at
/home/user/php/php-55/ext/gd/gd_ctx.c:39
39 return php_write((void *)buf, l TSRMLS_CC);
(gdb) print l
$7 = 8017
This length comes from the failed vsnprintf attempt [1] to copy more than 8000 chars on a 4095
buffer, vsnprintf [2] "a return value of size or more means that the output was
truncated", however libgd returns this length and PHP prints more information from memory that
it should.
While this issue directly comes from the fact that libgd isn't checking the vsnprintf return
value, PHP should probably backport the #66339 patch to 5.5 to prevent it.
[1] https://github.com/libgd/libgd/blob/master/src/gd_xbm.c#L188
[2] http://linux.die.net/man/3/vsnprintf /
Test script:
---------------
<?php
$var1=imagecreatetruecolor ( 2 , 2);
$var2=str_repeat("ABCD", 1030);
$var3=0;
imagexbm($var1, $var2, $var3);
Expected result:
----------------
Error trying to create a long filename or something (current PHP 5.6 behavior)
Image on stdout (PHP 5.5 *broken* behavior)
Actual result:
--------------
....ABCDAB?\??G?y??!??=F?8?m??F?y?????J?P????0 ?<??4C?!?8?m??O?P?m????P??????
=F??O?P?m????P???/
N?!?l?m?8?m???m???m????P?m?P???!?8?m?????????A?N??M?0
??????0 ,?m?A`p? 5 ??
??0 ??0 0
?????\?A`p? ?>?? ??l?m?l?m?@?C??l?m?l?m?@M??C??0
????7<??@P?????m???m????@???-90 ?=
??GL?m?O?m? ? ?i?8?m???O?`??????0
???AhW? ?? ?0 ?<??? -90 ?=
??G??m???m????tu_ ?? _?q_ ?@?=
0
@?=
?l??0 ?<?? \?9?M?pl?(?m?@?=
?l?Z? ??m?????0
(?m?ll??l??Ml??l??l?0C?m?8?m?d???l??l?(?m?? 0
?l?@?=
??=
?Th 4?m?8?Ghp?=
?Sh ??=
Sh ?=
,?G?%? ?l?`?=
,?G?%? ?$? 0
@?=
?.G ??m?p???<?????D??? ?=
??=
}?G0???F??0??????A?D? ?,G ?=
??m?+d??&=l?m??4K? ??D?m?-90 ?=
??GK?m?N?m?\?7{6?? ????m????A?0
????\?<?{6?? ??J?m?8?m?????0<??8?m???????0 <??0
\?8?m??,G 0
????}~ ??????G8?m?_!W??r_!W\?{6?? ??~??0<?????????0
~??<??0
?, ??????<??0
@???`
???????????????
?hp-55/sapi/cli/p???A??? ?) ?? `???????=
???@???`??? ????????=
????"????}!W???"#???A??? ?
????????/home/user/img.php?H?c??~???????<???\ W?ptm\ W?J?m\ W?J?????\?H?c?{6??
?????#0<???$?`???#0
???<??0
?, ?$????P????????`??????#????A??? ?) /home/user/php/php-55/sapi/cli/php???H?????????H???4
9v???? ????????x??? ?L??????? ??h+????? ?L! ??????? ?? E?u?@??????~6?:b? ???=?????? ??????=??
??t??? E?u?mG?????F??????????????
=================================================================
==3897==ERROR: AddressSanitizer: stack-buffer-underflow on address 0xbfffb750 at pc 0xb7aa6dbd bp
0xbfffa408 sp 0xbfff9fdc
READ of size 8017 at 0xbfffb750 thread T0
#0 0xb7aa6dbc (/usr/lib/i386-linux-gnu/libasan.so.2+0x3ddbc)
#1 0x99388cf in sapi_cli_single_write /home/user/php/php-55/sapi/cli/php_cli.c:273
#2 0x9938a01 in sapi_cli_ub_write /home/user/php/php-55/sapi/cli/php_cli.c:308
#3 0x920f82f in php_output_op /home/user/php/php-55/main/output.c:1094
#4 0x920f82f in php_output_write /home/user/php/php-55/main/output.c:270
#5 0x919debb in php_write /home/user/php/php-55/main/main.c:671
#6 0xb44b1a3b in _php_image_output_putbuf /home/user/php/php-55/ext/gd/gd_ctx.c:39
#7 0xb44638ba (/usr/lib/i386-linux-gnu/libgd.so.3+0x318ba)
#8 0xb4463ddf in gdImageXbmCtx (/usr/lib/i386-linux-gnu/libgd.so.3+0x31ddf)
#9 0xb44e0a2e in _php_image_output_ctx /home/user/php/php-55/ext/gd/gd_ctx.c:187
#10 0xb44e0a2e in zif_imagexbm /home/user/php/php-55/ext/gd/gd.c:2702
#11 0x992dc59 in zend_do_fcall_common_helper_SPEC
/home/user/php/php-55/Zend/zend_vm_execute.h:550
#12 0x96854cf in execute_ex /home/user/php/php-55/Zend/zend_vm_execute.h:363
#13 0x99225d6 in zend_execute /home/user/php/php-55/Zend/zend_vm_execute.h:388
#14 0x9472ed1 in zend_execute_scripts /home/user/php/php-55/Zend/zend.c:1327
#15 0x91a7e7c in php_execute_script /home/user/php/php-55/main/main.c:2525
#16 0x99379c4 in do_cli /home/user/php/php-55/sapi/cli/php_cli.c:994
#17 0x808a06c in main /home/user/php/php-55/sapi/cli/php_cli.c:1378
#18 0xb764c636 in __libc_start_main (/lib/i386-linux-gnu/libc.so.6+0x18636)
#19 0x808a78a (/home/user/php/php-55/sapi/cli/php+0x808a78a)
Address 0xbfffb750 is located in stack of thread T0 at offset 0 in frame
#0 0xb44dfe0f in zif_imagexbm /home/user/php/php-55/ext/gd/gd.c:2701
This frame has 6 object(s):
[32, 36) 'imgind' <== Memory access at offset 0 partially underflows this variable
[96, 100) 'file' <== Memory access at offset 0 partially underflows this variable
[160, 164) 'file_len' <== Memory access at offset 0 partially underflows this
variable
[224, 228) 'quality' <== Memory access at offset 0 partially underflows this
variable
[288, 292) 'basefilter' <== Memory access at offset 0 partially underflows this
variable
[352, 356) 'to_zval' <== Memory access at offset 0 partially underflows this
variable
HINT: this may be a false positive if your program uses some custom stack unwind mechanism or
swapcontext
(longjmp and C++ exceptions *are* supported)
SUMMARY: AddressSanitizer: stack-buffer-underflow ??:0 ??
Shadow bytes around the buggy address:
0x37fff690: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x37fff6a0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x37fff6b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x37fff6c0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x37fff6d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
=>0x37fff6e0: 00 00 00 00 00 00 00 00 00 00[f1]f1 f1 f1 04 f4
0x37fff6f0: f4 f4 f2 f2 f2 f2 04 f4 f4 f4 f2 f2 f2 f2 04 f4
0x37fff700: f4 f4 f2 f2 f2 f2 04 f4 f4 f4 f2 f2 f2 f2 04 f4
0x37fff710: f4 f4 f2 f2 f2 f2 04 f4 f4 f4 f3 f3 f3 f3 00 00
0x37fff720: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
0x37fff730: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
Shadow byte legend (one shadow byte represents 8 application bytes):
Addressable: 00
Partially addressable: 01 02 03 04 05 06 07
Heap left redzone: fa
Heap right redzone: fb
Freed heap region: fd
Stack left redzone: f1
Stack mid redzone: f2
Stack right redzone: f3
Stack partial redzone: f4
Stack after return: f5
Stack use after scope: f8
Global redzone: f9
Global init order: f6
Poisoned by user: f7
Container overflow: fc
Array cookie: ac
Intra object redzone: bb
ASan internal: fe
==3897==ABORTING
------------------------------------------------------------------------
--
Edit this bug report at https://bugs.php.net/bug.php?id=72115&edit=1