diff mbox series

jfs: fix array-index-out-of-bounds in diNewExt

Message ID tencent_B86ECD2ECECC92A7ED86EF92D0064A499206@qq.com (mailing list archive)
State New, archived
Headers show
Series jfs: fix array-index-out-of-bounds in diNewExt | expand

Commit Message

Edward Adam Davis Dec. 12, 2023, 1:36 a.m. UTC 
[Syz report]
UBSAN: array-index-out-of-bounds in fs/jfs/jfs_imap.c:2360:2
index -878706688 is out of range for type 'struct iagctl[128]'
CPU: 1 PID: 5065 Comm: syz-executor282 Not tainted 6.7.0-rc4-syzkaller-00009-gbee0e7762ad2 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/10/2023
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0x1e7/0x2d0 lib/dump_stack.c:106
 ubsan_epilogue lib/ubsan.c:217 [inline]
 __ubsan_handle_out_of_bounds+0x11c/0x150 lib/ubsan.c:348
 diNewExt+0x3cf3/0x4000 fs/jfs/jfs_imap.c:2360
 diAllocExt fs/jfs/jfs_imap.c:1949 [inline]
 diAllocAG+0xbe8/0x1e50 fs/jfs/jfs_imap.c:1666
 diAlloc+0x1d3/0x1760 fs/jfs/jfs_imap.c:1587
 ialloc+0x8f/0x900 fs/jfs/jfs_inode.c:56
 jfs_mkdir+0x1c5/0xb90 fs/jfs/namei.c:225
 vfs_mkdir+0x2f1/0x4b0 fs/namei.c:4106
 do_mkdirat+0x264/0x3a0 fs/namei.c:4129
 __do_sys_mkdir fs/namei.c:4149 [inline]
 __se_sys_mkdir fs/namei.c:4147 [inline]
 __x64_sys_mkdir+0x6e/0x80 fs/namei.c:4147
 do_syscall_x64 arch/x86/entry/common.c:51 [inline]
 do_syscall_64+0x45/0x110 arch/x86/entry/common.c:82
 entry_SYSCALL_64_after_hwframe+0x63/0x6b
RIP: 0033:0x7fcb7e6a0b57
Code: ff ff 77 07 31 c0 c3 0f 1f 40 00 48 c7 c2 b8 ff ff ff f7 d8 64 89 02 b8 ff ff ff ff c3 66 0f 1f 44 00 00 b8 53 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffd83023038 EFLAGS: 00000286 ORIG_RAX: 0000000000000053
RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007fcb7e6a0b57
RDX: 00000000000a1020 RSI: 00000000000001ff RDI: 0000000020000140
RBP: 0000000020000140 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000286 R12: 00007ffd830230d0
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000

[Analysis]
When the agstart is too large, it can cause agno overflow.

[Fix]
After obtaining agno, if the value is invalid, exit the subsequent process.

Reported-and-tested-by: syzbot+553d90297e6d2f50dbc7@syzkaller.appspotmail.com
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
---
 fs/jfs/jfs_imap.c | 3 +++
 1 file changed, 3 insertions(+)

Comments

Dave Kleikamp Dec. 20, 2023, 8:02 p.m. UTC | #1 
On 12/11/23 7:36PM, Edward Adam Davis wrote:
> [Syz report]
> UBSAN: array-index-out-of-bounds in fs/jfs/jfs_imap.c:2360:2
> index -878706688 is out of range for type 'struct iagctl[128]'
> CPU: 1 PID: 5065 Comm: syz-executor282 Not tainted 6.7.0-rc4-syzkaller-00009-gbee0e7762ad2 #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/10/2023
> Call Trace:
>   <TASK>
>   __dump_stack lib/dump_stack.c:88 [inline]
>   dump_stack_lvl+0x1e7/0x2d0 lib/dump_stack.c:106
>   ubsan_epilogue lib/ubsan.c:217 [inline]
>   __ubsan_handle_out_of_bounds+0x11c/0x150 lib/ubsan.c:348
>   diNewExt+0x3cf3/0x4000 fs/jfs/jfs_imap.c:2360
>   diAllocExt fs/jfs/jfs_imap.c:1949 [inline]
>   diAllocAG+0xbe8/0x1e50 fs/jfs/jfs_imap.c:1666
>   diAlloc+0x1d3/0x1760 fs/jfs/jfs_imap.c:1587
>   ialloc+0x8f/0x900 fs/jfs/jfs_inode.c:56
>   jfs_mkdir+0x1c5/0xb90 fs/jfs/namei.c:225
>   vfs_mkdir+0x2f1/0x4b0 fs/namei.c:4106
>   do_mkdirat+0x264/0x3a0 fs/namei.c:4129
>   __do_sys_mkdir fs/namei.c:4149 [inline]
>   __se_sys_mkdir fs/namei.c:4147 [inline]
>   __x64_sys_mkdir+0x6e/0x80 fs/namei.c:4147
>   do_syscall_x64 arch/x86/entry/common.c:51 [inline]
>   do_syscall_64+0x45/0x110 arch/x86/entry/common.c:82
>   entry_SYSCALL_64_after_hwframe+0x63/0x6b
> RIP: 0033:0x7fcb7e6a0b57
> Code: ff ff 77 07 31 c0 c3 0f 1f 40 00 48 c7 c2 b8 ff ff ff f7 d8 64 89 02 b8 ff ff ff ff c3 66 0f 1f 44 00 00 b8 53 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007ffd83023038 EFLAGS: 00000286 ORIG_RAX: 0000000000000053
> RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007fcb7e6a0b57
> RDX: 00000000000a1020 RSI: 00000000000001ff RDI: 0000000020000140
> RBP: 0000000020000140 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000286 R12: 00007ffd830230d0
> R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
> 
> [Analysis]
> When the agstart is too large, it can cause agno overflow.
> 
> [Fix]
> After obtaining agno, if the value is invalid, exit the subsequent process.

Looks good. Applied.

Shaggy

> 
> Reported-and-tested-by: syzbot+553d90297e6d2f50dbc7@syzkaller.appspotmail.com
> Signed-off-by: Edward Adam Davis <eadavis@qq.com>
> ---
>   fs/jfs/jfs_imap.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
> index a037ee59e398..cc5819b3ec9a 100644
> --- a/fs/jfs/jfs_imap.c
> +++ b/fs/jfs/jfs_imap.c
> @@ -2179,6 +2179,9 @@ static int diNewExt(struct inomap * imap, struct iag * iagp, int extno)
>   	/* get the ag and iag numbers for this iag.
>   	 */
>   	agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
> +	if (agno > MAXAG || agno < 0)
> +		return -EIO;
> +
>   	iagno = le32_to_cpu(iagp->iagnum);
>   
>   	/* check if this is the last free extent within the
Dan Carpenter Jan. 2, 2024, 1:29 p.m. UTC | #2 
Hi Edward,

kernel test robot noticed the following build warnings:

https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Edward-Adam-Davis/jfs-fix-array-index-out-of-bounds-in-diNewExt/20231212-095530
base:   https://github.com/kleikamp/linux-shaggy jfs-next
patch link:    https://lore.kernel.org/r/tencent_B86ECD2ECECC92A7ED86EF92D0064A499206%40qq.com
patch subject: [PATCH] jfs: fix array-index-out-of-bounds in diNewExt
config: i386-randconfig-141-20231212 (https://download.01.org/0day-ci/archive/20231214/202312142348.6HRZtXTB-lkp@intel.com/config)
compiler: gcc-7 (Ubuntu 7.5.0-6ubuntu2) 7.5.0
reproduce: (https://download.01.org/0day-ci/archive/20231214/202312142348.6HRZtXTB-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
| Closes: https://lore.kernel.org/r/202312142348.6HRZtXTB-lkp@intel.com/

New smatch warnings:
fs/jfs/jfs_imap.c:2213 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128

Old smatch warnings:
fs/jfs/jfs_imap.c:2229 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2304 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2318 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2330 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2332 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2363 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
fs/jfs/jfs_imap.c:2364 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128

vim +2213 fs/jfs/jfs_imap.c

^1da177e4c3f41 Linus Torvalds    2005-04-16  2152  static int diNewExt(struct inomap * imap, struct iag * iagp, int extno)
^1da177e4c3f41 Linus Torvalds    2005-04-16  2153  {
^1da177e4c3f41 Linus Torvalds    2005-04-16  2154  	int agno, iagno, fwd, back, freei = 0, sword, rc;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2155  	struct iag *aiagp = NULL, *biagp = NULL, *ciagp = NULL;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2156  	struct metapage *amp, *bmp, *cmp, *dmp;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2157  	struct inode *ipimap;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2158  	s64 blkno, hint;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2159  	int i, j;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2160  	u32 mask;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2161  	ino_t ino;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2162  	struct dinode *dp;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2163  	struct jfs_sb_info *sbi;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2164  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2165  	/* better have free extents.
^1da177e4c3f41 Linus Torvalds    2005-04-16  2166  	 */
^1da177e4c3f41 Linus Torvalds    2005-04-16  2167  	if (!iagp->nfreeexts) {
eb8630d7d2fd13 Joe Perches       2013-06-04  2168  		jfs_error(imap->im_ipimap->i_sb, "no free extents\n");
^1da177e4c3f41 Linus Torvalds    2005-04-16  2169  		return -EIO;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2170  	}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2171  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2172  	/* get the inode map inode.
^1da177e4c3f41 Linus Torvalds    2005-04-16  2173  	 */
^1da177e4c3f41 Linus Torvalds    2005-04-16  2174  	ipimap = imap->im_ipimap;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2175  	sbi = JFS_SBI(ipimap->i_sb);
^1da177e4c3f41 Linus Torvalds    2005-04-16  2176  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2177  	amp = bmp = cmp = NULL;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2178  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2179  	/* get the ag and iag numbers for this iag.
^1da177e4c3f41 Linus Torvalds    2005-04-16  2180  	 */
^1da177e4c3f41 Linus Torvalds    2005-04-16  2181  	agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
f93b91b82fcf16 Edward Adam Davis 2023-12-12  2182  	if (agno > MAXAG || agno < 0)

The commit introduces this agno > MAXAG comparison.  But Smatch says
that it should be agno >= MAXAG.

f93b91b82fcf16 Edward Adam Davis 2023-12-12  2183  		return -EIO;
f93b91b82fcf16 Edward Adam Davis 2023-12-12  2184  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2185  	iagno = le32_to_cpu(iagp->iagnum);
^1da177e4c3f41 Linus Torvalds    2005-04-16  2186  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2187  	/* check if this is the last free extent within the
^1da177e4c3f41 Linus Torvalds    2005-04-16  2188  	 * iag.  if so, the iag must be removed from the ag
25985edcedea63 Lucas De Marchi   2011-03-30  2189  	 * free extent list, so get the iags preceding and
^1da177e4c3f41 Linus Torvalds    2005-04-16  2190  	 * following the iag on this list.
^1da177e4c3f41 Linus Torvalds    2005-04-16  2191  	 */
^1da177e4c3f41 Linus Torvalds    2005-04-16  2192  	if (iagp->nfreeexts == cpu_to_le32(1)) {
^1da177e4c3f41 Linus Torvalds    2005-04-16  2193  		if ((fwd = le32_to_cpu(iagp->extfreefwd)) >= 0) {
^1da177e4c3f41 Linus Torvalds    2005-04-16  2194  			if ((rc = diIAGRead(imap, fwd, &amp)))
^1da177e4c3f41 Linus Torvalds    2005-04-16  2195  				return (rc);
^1da177e4c3f41 Linus Torvalds    2005-04-16  2196  			aiagp = (struct iag *) amp->data;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2197  		}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2198  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2199  		if ((back = le32_to_cpu(iagp->extfreeback)) >= 0) {
^1da177e4c3f41 Linus Torvalds    2005-04-16  2200  			if ((rc = diIAGRead(imap, back, &bmp)))
^1da177e4c3f41 Linus Torvalds    2005-04-16  2201  				goto error_out;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2202  			biagp = (struct iag *) bmp->data;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2203  		}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2204  	} else {
^1da177e4c3f41 Linus Torvalds    2005-04-16  2205  		/* the iag has free extents.  if all extents are free
^1da177e4c3f41 Linus Torvalds    2005-04-16  2206  		 * (as is the case for a newly allocated iag), the iag
^1da177e4c3f41 Linus Torvalds    2005-04-16  2207  		 * must be added to the ag free extent list, so get
^1da177e4c3f41 Linus Torvalds    2005-04-16  2208  		 * the iag at the head of the list in preparation for
^1da177e4c3f41 Linus Torvalds    2005-04-16  2209  		 * adding this iag to this list.
^1da177e4c3f41 Linus Torvalds    2005-04-16  2210  		 */
^1da177e4c3f41 Linus Torvalds    2005-04-16  2211  		fwd = back = -1;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2212  		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG)) {
^1da177e4c3f41 Linus Torvalds    2005-04-16 @2213  			if ((fwd = imap->im_agctl[agno].extfree) >= 0) {

If agno == MAXAG then we're out of bounds here.

^1da177e4c3f41 Linus Torvalds    2005-04-16  2214  				if ((rc = diIAGRead(imap, fwd, &amp)))
^1da177e4c3f41 Linus Torvalds    2005-04-16  2215  					goto error_out;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2216  				aiagp = (struct iag *) amp->data;
^1da177e4c3f41 Linus Torvalds    2005-04-16  2217  			}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2218  		}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2219  	}
^1da177e4c3f41 Linus Torvalds    2005-04-16  2220  
^1da177e4c3f41 Linus Torvalds    2005-04-16  2221  	/* check if the iag has no free inodes.  if so, the iag
Dave Kleikamp Jan. 2, 2024, 5:10 p.m. UTC | #3 
On 1/2/24 7:29AM, Dan Carpenter wrote:
> Hi Edward,
> 
> kernel test robot noticed the following build warnings:
> 
> https://git-scm.com/docs/git-format-patch#_base_tree_information]
> 
> url:    https://github.com/intel-lab-lkp/linux/commits/Edward-Adam-Davis/jfs-fix-array-index-out-of-bounds-in-diNewExt/20231212-095530
> base:   https://github.com/kleikamp/linux-shaggy jfs-next
> patch link:    https://lore.kernel.org/r/tencent_B86ECD2ECECC92A7ED86EF92D0064A499206%40qq.com
> patch subject: [PATCH] jfs: fix array-index-out-of-bounds in diNewExt
> config: i386-randconfig-141-20231212 (https://download.01.org/0day-ci/archive/20231214/202312142348.6HRZtXTB-lkp@intel.com/config)
> compiler: gcc-7 (Ubuntu 7.5.0-6ubuntu2) 7.5.0
> reproduce: (https://download.01.org/0day-ci/archive/20231214/202312142348.6HRZtXTB-lkp@intel.com/reproduce)
> 
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@intel.com>
> | Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
> | Closes: https://lore.kernel.org/r/202312142348.6HRZtXTB-lkp@intel.com/

I modified Edward's patch in the jfs-next branch with the corrected test.

Thanks for catching this.

Shaggy

> 
> New smatch warnings:
> fs/jfs/jfs_imap.c:2213 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> 
> Old smatch warnings:
> fs/jfs/jfs_imap.c:2229 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2304 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2318 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2330 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2332 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2363 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> fs/jfs/jfs_imap.c:2364 diNewExt() error: buffer overflow 'imap->im_imap.in_agctl' 128 <= 128
> 
> vim +2213 fs/jfs/jfs_imap.c
> 
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2152  static int diNewExt(struct inomap * imap, struct iag * iagp, int extno)
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2153  {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2154  	int agno, iagno, fwd, back, freei = 0, sword, rc;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2155  	struct iag *aiagp = NULL, *biagp = NULL, *ciagp = NULL;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2156  	struct metapage *amp, *bmp, *cmp, *dmp;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2157  	struct inode *ipimap;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2158  	s64 blkno, hint;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2159  	int i, j;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2160  	u32 mask;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2161  	ino_t ino;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2162  	struct dinode *dp;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2163  	struct jfs_sb_info *sbi;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2164
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2165  	/* better have free extents.
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2166  	 */
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2167  	if (!iagp->nfreeexts) {
> eb8630d7d2fd13 Joe Perches       2013-06-04  2168  		jfs_error(imap->im_ipimap->i_sb, "no free extents\n");
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2169  		return -EIO;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2170  	}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2171
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2172  	/* get the inode map inode.
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2173  	 */
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2174  	ipimap = imap->im_ipimap;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2175  	sbi = JFS_SBI(ipimap->i_sb);
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2176
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2177  	amp = bmp = cmp = NULL;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2178
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2179  	/* get the ag and iag numbers for this iag.
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2180  	 */
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2181  	agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
> f93b91b82fcf16 Edward Adam Davis 2023-12-12  2182  	if (agno > MAXAG || agno < 0)
> 
> The commit introduces this agno > MAXAG comparison.  But Smatch says
> that it should be agno >= MAXAG.
> 
> f93b91b82fcf16 Edward Adam Davis 2023-12-12  2183  		return -EIO;
> f93b91b82fcf16 Edward Adam Davis 2023-12-12  2184
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2185  	iagno = le32_to_cpu(iagp->iagnum);
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2186
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2187  	/* check if this is the last free extent within the
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2188  	 * iag.  if so, the iag must be removed from the ag
> 25985edcedea63 Lucas De Marchi   2011-03-30  2189  	 * free extent list, so get the iags preceding and
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2190  	 * following the iag on this list.
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2191  	 */
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2192  	if (iagp->nfreeexts == cpu_to_le32(1)) {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2193  		if ((fwd = le32_to_cpu(iagp->extfreefwd)) >= 0) {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2194  			if ((rc = diIAGRead(imap, fwd, &amp)))
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2195  				return (rc);
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2196  			aiagp = (struct iag *) amp->data;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2197  		}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2198
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2199  		if ((back = le32_to_cpu(iagp->extfreeback)) >= 0) {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2200  			if ((rc = diIAGRead(imap, back, &bmp)))
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2201  				goto error_out;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2202  			biagp = (struct iag *) bmp->data;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2203  		}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2204  	} else {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2205  		/* the iag has free extents.  if all extents are free
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2206  		 * (as is the case for a newly allocated iag), the iag
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2207  		 * must be added to the ag free extent list, so get
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2208  		 * the iag at the head of the list in preparation for
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2209  		 * adding this iag to this list.
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2210  		 */
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2211  		fwd = back = -1;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2212  		if (iagp->nfreeexts == cpu_to_le32(EXTSPERIAG)) {
> ^1da177e4c3f41 Linus Torvalds    2005-04-16 @2213  			if ((fwd = imap->im_agctl[agno].extfree) >= 0) {
> 
> If agno == MAXAG then we're out of bounds here.
> 
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2214  				if ((rc = diIAGRead(imap, fwd, &amp)))
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2215  					goto error_out;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2216  				aiagp = (struct iag *) amp->data;
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2217  			}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2218  		}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2219  	}
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2220
> ^1da177e4c3f41 Linus Torvalds    2005-04-16  2221  	/* check if the iag has no free inodes.  if so, the iag
>
diff mbox series

Patch

diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
index a037ee59e398..cc5819b3ec9a 100644
--- a/fs/jfs/jfs_imap.c
+++ b/fs/jfs/jfs_imap.c
@@ -2179,6 +2179,9 @@  static int diNewExt(struct inomap * imap, struct iag * iagp, int extno)
 	/* get the ag and iag numbers for this iag.
 	 */
 	agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi);
+	if (agno > MAXAG || agno < 0)
+		return -EIO;
+
 	iagno = le32_to_cpu(iagp->iagnum);
 
 	/* check if this is the last free extent within the