linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] spufs raises two exceptions
@ 2012-03-06  9:26 masterzorag
  2012-03-07  3:49 ` Benjamin Herrenschmidt
  0 siblings, 1 reply; 7+ messages in thread
From: masterzorag @ 2012-03-06  9:26 UTC (permalink / raw)
  To: linuxppc-dev

I'm running my test program, it uses all available spus to compute via 
OpenCL
kernel 3.2.5 on a ps3
even on testing spu directly, it crashes


=====================================
[ BUG: bad unlock balance detected! ]
-------------------------------------
test/1067 is trying to release lock (&sb->s_type->i_mutex_key) at:
[<d0000000005828a8>] .do_spu_create+0x90/0xd8 [spufs]
but there are no more locks to release!
other info that might help us debug this:
no locks held by test/1067.
stack backtrace:
Call Trace:
[c00000000e9bfa30] [c0000000000110d0] .show_stack+0x6c/0x16c (unreliable)
[c00000000e9bfae0] [c000000000081f90] .print_unlock_inbalance_bug+0xe8/0x110
[c00000000e9bfb70] [c0000000000868cc] .lock_release+0xd8/0x200
[c00000000e9bfc10] [c0000000003efb60] .__mutex_unlock_slowpath+0x11c/0x1d8
[c00000000e9bfcb0] [d0000000005828a8] .do_spu_create+0x90/0xd8 [spufs]
[c00000000e9bfd70] [c0000000000346ac] .sys_spu_create+0x164/0x1c0
[c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
------------[ cut here ]------------
kernel BUG at fs/dcache.c:474!
Oops: Exception in kernel mode, sig: 5 [#1]
SMP NR_CPUS=2 NUMA PS3
Modules linked in: spufs dm_mod btusb bluetooth usb_storage ohci_hcd 
snd_ps3 ehci_hcd snd_pcm snd_page_alloc snd_timer sg snd usbcore 
usb_common ps3flash rtc_ps3 soundcore ps3_lpm ps3vram [last unloaded: 
scsi_wait_scan]
NIP: c000000000109f94 LR: c000000000109f84 CTR: c0000000000a029c
REGS: c00000000e9bf930 TRAP: 0700 Not tainted (3.2.5)
MSR: 8000000000028032 <EE,CE,IR,DR> CR: 22004822 XER: 00000000
TASK = c0000000062f0ec0[1067] 'test' THREAD: c00000000e9bc000 CPU: 1
GPR00: 0000000000000001 c00000000e9bfbb0 c0000000006812e8 c00000000543b798
GPR04: 0000000000000000 0000000000000000 0000000000000000 0000000000000002
GPR08: 0000000000000000 0000000000000000 c000000000109f84 c0000000062f0ec0
GPR12: 0000000082004824 c000000007ffe280 0000000000000004 00000000f7850688
GPR16: 00000000f7830734 00000000f78517a4 00000000f7852008 00000000f78517a8
GPR20: 00000000ff805dc0 000000000fd958a0 0000000000000000 000000000000000d
GPR24: 000000000fd98240 c00000000e101e10 0000000040000010 c00000000616e080
GPR28: c00000000543b738 c00000000543b798 c0000000006149e8 c00000000543b738
NIP [c000000000109f94] .dput+0x48/0x214
LR [c000000000109f84] .dput+0x38/0x214
Call Trace:
[c00000000e9bfbb0] [c000000000109f84] .dput+0x38/0x214 (unreliable)
[c00000000e9bfc50] [c0000000000f1740] .fput+0x24c/0x288
[c00000000e9bfd00] [c0000000000ed708] .filp_close+0xbc/0xe4
[c00000000e9bfd90] [c0000000000ed800] .SyS_close+0xd0/0x128
[c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
Instruction dump:
fb61ffd8 fb81ffe0 fba1ffe8 f821ff61 418201c8 3bbf0060 7fa3eb78 482e7f31
60000000 813f0058 7d200074 7800d182 <0b000000> 2b890001 409d0010 3809ffff
---[ end trace c337aad05d94532f ]---
------------[ cut here ]------------
kernel BUG at fs/dcache.c:474!
Oops: Exception in kernel mode, sig: 5 [#2]
SMP NR_CPUS=2 NUMA PS3
Modules linked in: spufs dm_mod btusb bluetooth usb_storage ohci_hcd 
snd_ps3 ehci_hcd snd_pcm snd_page_alloc snd_timer sg snd usbcore 
usb_common ps3flash rtc_ps3 soundcore ps3_lpm ps3vram [last unloaded: 
scsi_wait_scan]
NIP: c000000000109f94 LR: c000000000109f84 CTR: c0000000000a029c
REGS: c00000000e9bec20 TRAP: 0700 Tainted: G D (3.2.5)
MSR: 8000000000028032 <EE,CE,IR,DR> CR: 22004822 XER: 00000000
TASK = c0000000062f0ec0[1067] 'test' THREAD: c00000000e9bc000 CPU: 1
GPR00: 0000000000000001 c00000000e9beea0 c0000000006812e8 c0000000054361c8
GPR04: 0000000000000000 0000000000000000 0000000000000000 0000000000000002
GPR08: 0000000000000000 0000000000000000 c000000000109f84 c0000000062f0ec0
GPR12: 0000000042004824 c000000007ffe280 0000000000000004 00000000f7850688
GPR16: 00000000f7830734 00000000f78517a4 00000000f7852008 00000000f78517a8
GPR20: 00000000ff805dc0 000000000fd958a0 0000000000000000 0000000000000001
GPR24: 000000000fd98240 c00000000e9b2390 0000000000000008 c0000000062bd010
GPR28: c000000005436168 c0000000054361c8 c0000000006149e8 c000000005436168
NIP [c000000000109f94] .dput+0x48/0x214
LR [c000000000109f84] .dput+0x38/0x214
Call Trace:
[c00000000e9beea0] [c000000000109f84] .dput+0x38/0x214 (unreliable)
[c00000000e9bef40] [c0000000000f1740] .fput+0x24c/0x288
[c00000000e9beff0] [c0000000000c93a8] .remove_vma+0x68/0xcc
[c00000000e9bf080] [c0000000000c951c] .exit_mmap+0x110/0x14c
[c00000000e9bf1a0] [c00000000004b4c8] .mmput+0x5c/0x13c
[c00000000e9bf230] [d00000000058237c] .spu_forget+0x54/0x7c [spufs]
[c00000000e9bf2c0] [d00000000057c294] .spufs_dir_close+0x8c/0xc8 [spufs]
[c00000000e9bf370] [c0000000000f166c] .fput+0x178/0x288
[c00000000e9bf420] [c0000000000ed708] .filp_close+0xbc/0xe4
[c00000000e9bf4b0] [c000000000050294] .put_files_struct+0xf4/0x1b8
[c00000000e9bf560] [c0000000000520bc] .do_exit+0x23c/0x6f4
[c00000000e9bf660] [c00000000001922c] .die+0x274/0x2a4
[c00000000e9bf700] [c000000000019640] ._exception+0x88/0x17c
[c00000000e9bf8c0] [c000000000005314] program_check_common+0x114/0x180
--- Exception: 700 at .dput+0x48/0x214
LR = .dput+0x38/0x214
[c00000000e9bfc50] [c0000000000f1740] .fput+0x24c/0x288
[c00000000e9bfd00] [c0000000000ed708] .filp_close+0xbc/0xe4
[c00000000e9bfd90] [c0000000000ed800] .SyS_close+0xd0/0x128
[c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
Instruction dump:
fb61ffd8 fb81ffe0 fba1ffe8 f821ff61 418201c8 3bbf0060 7fa3eb78 482e7f31
60000000 813f0058 7d200074 7800d182 <0b000000> 2b890001 409d0010 3809ffff
---[ end trace c337aad05d945330 ]---
Fixing recursive fault but reboot is needed!

First time, the mutex gets unlocked in spufs_create_context, then the 
second time in do_spu_create.
It seems that SPU main directory dentry has invalid d_count.


This patch fixes all, OpenCL is running fine, testing spe runs without 
issues.

--- arch/powerpc/platforms/cell/spufs/syscalls.c
+++ arch/powerpc/platforms/cell/spufs/syscalls.c.new
@@ -70,8 +70,8 @@
      ret = PTR_ERR(dentry);
      if (!IS_ERR(dentry)) {
          ret = spufs_create(&path, dentry, flags, mode, neighbor);
-        mutex_unlock(&path.dentry->d_inode->i_mutex);
-        dput(dentry);
+        if (ret < 0)
+            dput(dentry);
          path_put(&path);
      }

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-06  9:26 [PATCH] spufs raises two exceptions masterzorag
@ 2012-03-07  3:49 ` Benjamin Herrenschmidt
  2012-03-07  3:51   ` Benjamin Herrenschmidt
  0 siblings, 1 reply; 7+ messages in thread
From: Benjamin Herrenschmidt @ 2012-03-07  3:49 UTC (permalink / raw)
  To: masterzorag; +Cc: linuxppc-dev, Al Viro, Arnd Bergmann

On Tue, 2012-03-06 at 10:26 +0100, masterzorag wrote:
> I'm running my test program, it uses all available spus to compute via 
> OpenCL
> kernel 3.2.5 on a ps3
> even on testing spu directly, it crashes

I think the patch is not 100% right yet. Looking at the code, we
have a real mess of who gets to clean what up here. This is an
attempt at sorting things by having the mutex and dentry dropped
in spufs_create() always. Can you give it a spin (untested):

Al, I'm not familiar with the vfs, can you take a quick look ?

Thanks !

Cheers,
Ben.


> 
> =====================================
> [ BUG: bad unlock balance detected! ]
> -------------------------------------
> test/1067 is trying to release lock (&sb->s_type->i_mutex_key) at:
> [<d0000000005828a8>] .do_spu_create+0x90/0xd8 [spufs]
> but there are no more locks to release!
> other info that might help us debug this:
> no locks held by test/1067.
> stack backtrace:
> Call Trace:
> [c00000000e9bfa30] [c0000000000110d0] .show_stack+0x6c/0x16c (unreliable)
> [c00000000e9bfae0] [c000000000081f90] .print_unlock_inbalance_bug+0xe8/0x110
> [c00000000e9bfb70] [c0000000000868cc] .lock_release+0xd8/0x200
> [c00000000e9bfc10] [c0000000003efb60] .__mutex_unlock_slowpath+0x11c/0x1d8
> [c00000000e9bfcb0] [d0000000005828a8] .do_spu_create+0x90/0xd8 [spufs]
> [c00000000e9bfd70] [c0000000000346ac] .sys_spu_create+0x164/0x1c0
> [c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
> ------------[ cut here ]------------
> kernel BUG at fs/dcache.c:474!
> Oops: Exception in kernel mode, sig: 5 [#1]
> SMP NR_CPUS=2 NUMA PS3
> Modules linked in: spufs dm_mod btusb bluetooth usb_storage ohci_hcd 
> snd_ps3 ehci_hcd snd_pcm snd_page_alloc snd_timer sg snd usbcore 
> usb_common ps3flash rtc_ps3 soundcore ps3_lpm ps3vram [last unloaded: 
> scsi_wait_scan]
> NIP: c000000000109f94 LR: c000000000109f84 CTR: c0000000000a029c
> REGS: c00000000e9bf930 TRAP: 0700 Not tainted (3.2.5)
> MSR: 8000000000028032 <EE,CE,IR,DR> CR: 22004822 XER: 00000000
> TASK = c0000000062f0ec0[1067] 'test' THREAD: c00000000e9bc000 CPU: 1
> GPR00: 0000000000000001 c00000000e9bfbb0 c0000000006812e8 c00000000543b798
> GPR04: 0000000000000000 0000000000000000 0000000000000000 0000000000000002
> GPR08: 0000000000000000 0000000000000000 c000000000109f84 c0000000062f0ec0
> GPR12: 0000000082004824 c000000007ffe280 0000000000000004 00000000f7850688
> GPR16: 00000000f7830734 00000000f78517a4 00000000f7852008 00000000f78517a8
> GPR20: 00000000ff805dc0 000000000fd958a0 0000000000000000 000000000000000d
> GPR24: 000000000fd98240 c00000000e101e10 0000000040000010 c00000000616e080
> GPR28: c00000000543b738 c00000000543b798 c0000000006149e8 c00000000543b738
> NIP [c000000000109f94] .dput+0x48/0x214
> LR [c000000000109f84] .dput+0x38/0x214
> Call Trace:
> [c00000000e9bfbb0] [c000000000109f84] .dput+0x38/0x214 (unreliable)
> [c00000000e9bfc50] [c0000000000f1740] .fput+0x24c/0x288
> [c00000000e9bfd00] [c0000000000ed708] .filp_close+0xbc/0xe4
> [c00000000e9bfd90] [c0000000000ed800] .SyS_close+0xd0/0x128
> [c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
> Instruction dump:
> fb61ffd8 fb81ffe0 fba1ffe8 f821ff61 418201c8 3bbf0060 7fa3eb78 482e7f31
> 60000000 813f0058 7d200074 7800d182 <0b000000> 2b890001 409d0010 3809ffff
> ---[ end trace c337aad05d94532f ]---
> ------------[ cut here ]------------
> kernel BUG at fs/dcache.c:474!
> Oops: Exception in kernel mode, sig: 5 [#2]
> SMP NR_CPUS=2 NUMA PS3
> Modules linked in: spufs dm_mod btusb bluetooth usb_storage ohci_hcd 
> snd_ps3 ehci_hcd snd_pcm snd_page_alloc snd_timer sg snd usbcore 
> usb_common ps3flash rtc_ps3 soundcore ps3_lpm ps3vram [last unloaded: 
> scsi_wait_scan]
> NIP: c000000000109f94 LR: c000000000109f84 CTR: c0000000000a029c
> REGS: c00000000e9bec20 TRAP: 0700 Tainted: G D (3.2.5)
> MSR: 8000000000028032 <EE,CE,IR,DR> CR: 22004822 XER: 00000000
> TASK = c0000000062f0ec0[1067] 'test' THREAD: c00000000e9bc000 CPU: 1
> GPR00: 0000000000000001 c00000000e9beea0 c0000000006812e8 c0000000054361c8
> GPR04: 0000000000000000 0000000000000000 0000000000000000 0000000000000002
> GPR08: 0000000000000000 0000000000000000 c000000000109f84 c0000000062f0ec0
> GPR12: 0000000042004824 c000000007ffe280 0000000000000004 00000000f7850688
> GPR16: 00000000f7830734 00000000f78517a4 00000000f7852008 00000000f78517a8
> GPR20: 00000000ff805dc0 000000000fd958a0 0000000000000000 0000000000000001
> GPR24: 000000000fd98240 c00000000e9b2390 0000000000000008 c0000000062bd010
> GPR28: c000000005436168 c0000000054361c8 c0000000006149e8 c000000005436168
> NIP [c000000000109f94] .dput+0x48/0x214
> LR [c000000000109f84] .dput+0x38/0x214
> Call Trace:
> [c00000000e9beea0] [c000000000109f84] .dput+0x38/0x214 (unreliable)
> [c00000000e9bef40] [c0000000000f1740] .fput+0x24c/0x288
> [c00000000e9beff0] [c0000000000c93a8] .remove_vma+0x68/0xcc
> [c00000000e9bf080] [c0000000000c951c] .exit_mmap+0x110/0x14c
> [c00000000e9bf1a0] [c00000000004b4c8] .mmput+0x5c/0x13c
> [c00000000e9bf230] [d00000000058237c] .spu_forget+0x54/0x7c [spufs]
> [c00000000e9bf2c0] [d00000000057c294] .spufs_dir_close+0x8c/0xc8 [spufs]
> [c00000000e9bf370] [c0000000000f166c] .fput+0x178/0x288
> [c00000000e9bf420] [c0000000000ed708] .filp_close+0xbc/0xe4
> [c00000000e9bf4b0] [c000000000050294] .put_files_struct+0xf4/0x1b8
> [c00000000e9bf560] [c0000000000520bc] .do_exit+0x23c/0x6f4
> [c00000000e9bf660] [c00000000001922c] .die+0x274/0x2a4
> [c00000000e9bf700] [c000000000019640] ._exception+0x88/0x17c
> [c00000000e9bf8c0] [c000000000005314] program_check_common+0x114/0x180
> --- Exception: 700 at .dput+0x48/0x214
> LR = .dput+0x38/0x214
> [c00000000e9bfc50] [c0000000000f1740] .fput+0x24c/0x288
> [c00000000e9bfd00] [c0000000000ed708] .filp_close+0xbc/0xe4
> [c00000000e9bfd90] [c0000000000ed800] .SyS_close+0xd0/0x128
> [c00000000e9bfe30] [c0000000000097d8] syscall_exit+0x0/0x40
> Instruction dump:
> fb61ffd8 fb81ffe0 fba1ffe8 f821ff61 418201c8 3bbf0060 7fa3eb78 482e7f31
> 60000000 813f0058 7d200074 7800d182 <0b000000> 2b890001 409d0010 3809ffff
> ---[ end trace c337aad05d945330 ]---
> Fixing recursive fault but reboot is needed!
> 
> First time, the mutex gets unlocked in spufs_create_context, then the 
> second time in do_spu_create.
> It seems that SPU main directory dentry has invalid d_count.
> 
> 
> This patch fixes all, OpenCL is running fine, testing spe runs without 
> issues.
> 
> --- arch/powerpc/platforms/cell/spufs/syscalls.c
> +++ arch/powerpc/platforms/cell/spufs/syscalls.c.new
> @@ -70,8 +70,8 @@
>       ret = PTR_ERR(dentry);
>       if (!IS_ERR(dentry)) {
>           ret = spufs_create(&path, dentry, flags, mode, neighbor);
> -        mutex_unlock(&path.dentry->d_inode->i_mutex);
> -        dput(dentry);
> +        if (ret < 0)
> +            dput(dentry);
>           path_put(&path);
>       }
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-07  3:49 ` Benjamin Herrenschmidt
@ 2012-03-07  3:51   ` Benjamin Herrenschmidt
  2012-03-07 12:48     ` Arnd Bergmann
  0 siblings, 1 reply; 7+ messages in thread
From: Benjamin Herrenschmidt @ 2012-03-07  3:51 UTC (permalink / raw)
  To: masterzorag; +Cc: linuxppc-dev, Al Viro, Arnd Bergmann

On Wed, 2012-03-07 at 14:49 +1100, Benjamin Herrenschmidt wrote:
> On Tue, 2012-03-06 at 10:26 +0100, masterzorag wrote:
> > I'm running my test program, it uses all available spus to compute via 
> > OpenCL
> > kernel 3.2.5 on a ps3
> > even on testing spu directly, it crashes
> 
> I think the patch is not 100% right yet. Looking at the code, we
> have a real mess of who gets to clean what up here. This is an
> attempt at sorting things by having the mutex and dentry dropped
> in spufs_create() always. Can you give it a spin (untested):
> 
> Al, I'm not familiar with the vfs, can you take a quick look ?

Better with the actual patch :-)

powerpc/cell: Fix locking in spufs_create()

The error path in spufs_create() could cause double unlocks
among other horrors. Clean it up.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reported-by: masterzorag@gmail.com

diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
index d4a094c..63b4e43 100644
--- a/arch/powerpc/platforms/cell/spufs/inode.c
+++ b/arch/powerpc/platforms/cell/spufs/inode.c
@@ -454,19 +454,16 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
 	struct spu_gang *gang;
 	struct spu_context *neighbor;
 
-	ret = -EPERM;
 	if ((flags & SPU_CREATE_NOSCHED) &&
-	    !capable(CAP_SYS_NICE))
-		goto out_unlock;
+	    !capable(CAP_SYa_NICE))
+		return -EPERM;
 
-	ret = -EINVAL;
 	if ((flags & (SPU_CREATE_NOSCHED | SPU_CREATE_ISOLATE))
 	    == SPU_CREATE_ISOLATE)
-		goto out_unlock;
+		return -EINVAL;
 
-	ret = -ENODEV;
 	if ((flags & SPU_CREATE_ISOLATE) && !isolated_loader)
-		goto out_unlock;
+		return -ENODEV;
 
 	gang = NULL;
 	neighbor = NULL;
@@ -512,10 +509,6 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
 out_aff_unlock:
 	if (affinity)
 		mutex_unlock(&gang->aff_mutex);
-out_unlock:
-	mutex_unlock(&inode->i_mutex);
-out:
-	dput(dentry);
 	return ret;
 }
 
@@ -585,11 +578,9 @@ static int spufs_create_gang(struct inode *inode,
 			struct dentry *dentry,
 			struct vfsmount *mnt, umode_t mode)
 {
-	int ret;
-
-	ret = spufs_mkgang(inode, dentry, mode & S_IRWXUGO);
+	int ret = spufs_mkgang(inode, dentry, mode & S_IRWXUGO);
 	if (ret)
-		goto out;
+		return ret;
 
 	/*
 	 * get references for dget and mntget, will be released
@@ -600,10 +591,6 @@ static int spufs_create_gang(struct inode *inode,
 		int err = simple_rmdir(inode, dentry);
 		WARN_ON(err);
 	}
-
-out:
-	mutex_unlock(&inode->i_mutex);
-	dput(dentry);
 	return ret;
 }
 
@@ -613,22 +600,21 @@ static struct file_system_type spufs_type;
 long spufs_create(struct path *path, struct dentry *dentry,
 		unsigned int flags, umode_t mode, struct file *filp)
 {
-	int ret;
+	int ret = -EINVAL;
 
-	ret = -EINVAL;
 	/* check if we are on spufs */
 	if (path->dentry->d_sb->s_type != &spufs_type)
-		goto out;
+		goto fail;
 
 	/* don't accept undefined flags */
 	if (flags & (~SPU_CREATE_FLAG_ALL))
-		goto out;
+		goto fail;
 
 	/* only threads can be underneath a gang */
 	if (path->dentry != path->dentry->d_sb->s_root) {
 		if ((flags & SPU_CREATE_GANG) ||
 		    !SPUFS_I(path->dentry->d_inode)->i_gang)
-			goto out;
+		goto fail;
 	}
 
 	mode &= ~current_umask();
@@ -640,12 +626,17 @@ long spufs_create(struct path *path, struct dentry *dentry,
 		ret = spufs_create_context(path->dentry->d_inode,
 					    dentry, path->mnt, flags, mode,
 					    filp);
-	if (ret >= 0)
+	if (ret >= 0) {
+		/* We drop these before fsnotify */
+		mutex_unlock(&inode->i_mutex);
+		dput(dentry);
 		fsnotify_mkdir(path->dentry->d_inode, dentry);
-	return ret;
 
-out:
-	mutex_unlock(&path->dentry->d_inode->i_mutex);
+		return ret;
+	}
+ fail:
+	mutex_unlock(&inode->i_mutex);
+	dput(dentry);
 	return ret;
 }
 
diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c
index 8591bb6..1a65ef2 100644
--- a/arch/powerpc/platforms/cell/spufs/syscalls.c
+++ b/arch/powerpc/platforms/cell/spufs/syscalls.c
@@ -70,11 +70,8 @@ static long do_spu_create(const char __user *pathname, unsigned int flags,
 	ret = PTR_ERR(dentry);
 	if (!IS_ERR(dentry)) {
 		ret = spufs_create(&path, dentry, flags, mode, neighbor);
-		mutex_unlock(&path.dentry->d_inode->i_mutex);
-		dput(dentry);
 		path_put(&path);
 	}
-
 	return ret;
 }
 

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-07  3:51   ` Benjamin Herrenschmidt
@ 2012-03-07 12:48     ` Arnd Bergmann
  2012-03-07 21:01       ` Benjamin Herrenschmidt
  0 siblings, 1 reply; 7+ messages in thread
From: Arnd Bergmann @ 2012-03-07 12:48 UTC (permalink / raw)
  To: Benjamin Herrenschmidt; +Cc: linuxppc-dev, masterzorag, Al Viro

On Wednesday 07 March 2012, Benjamin Herrenschmidt wrote:
> On Wed, 2012-03-07 at 14:49 +1100, Benjamin Herrenschmidt wrote:
> > On Tue, 2012-03-06 at 10:26 +0100, masterzorag wrote:
> > > I'm running my test program, it uses all available spus to compute via 
> > > OpenCL
> > > kernel 3.2.5 on a ps3
> > > even on testing spu directly, it crashes
> > 
> > I think the patch is not 100% right yet. Looking at the code, we
> > have a real mess of who gets to clean what up here. This is an
> > attempt at sorting things by having the mutex and dentry dropped
> > in spufs_create() always. Can you give it a spin (untested):
> > 
> > Al, I'm not familiar with the vfs, can you take a quick look ?
> 
> Better with the actual patch :-)
> 
> powerpc/cell: Fix locking in spufs_create()
> 
> The error path in spufs_create() could cause double unlocks
> among other horrors. Clean it up.
> 
> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Reported-by: masterzorag@gmail.com
> 
> diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
> index d4a094c..63b4e43 100644
> --- a/arch/powerpc/platforms/cell/spufs/inode.c
> +++ b/arch/powerpc/platforms/cell/spufs/inode.c
> @@ -454,19 +454,16 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
>  	struct spu_gang *gang;
>  	struct spu_context *neighbor;
>  
> -	ret = -EPERM;
>  	if ((flags & SPU_CREATE_NOSCHED) &&
> -	    !capable(CAP_SYS_NICE))
> -		goto out_unlock;
> +	    !capable(CAP_SYa_NICE))

                    ^typo

> +		return -EPERM;
>  
> -	ret = -EINVAL;
>  	if ((flags & (SPU_CREATE_NOSCHED | SPU_CREATE_ISOLATE))
>  	    == SPU_CREATE_ISOLATE)
> -		goto out_unlock;
> +		return -EINVAL;
>  
> -	ret = -ENODEV;
>  	if ((flags & SPU_CREATE_ISOLATE) && !isolated_loader)
> -		goto out_unlock;
> +		return -ENODEV;
>  
>  	gang = NULL;
>  	neighbor = NULL;

This mostly changes coding style, pointlessly.

> @@ -512,10 +509,6 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
>  out_aff_unlock:
>  	if (affinity)
>  		mutex_unlock(&gang->aff_mutex);
> -out_unlock:
> -	mutex_unlock(&inode->i_mutex);
> -out:
> -	dput(dentry);
>  	return ret;
>  }

The original intention of this was to always unlock in the error case. It
seems that Al changed this in 1ba10681 "switch do_spufs_create() to
user_path_create(), fix double-unlock" to never unlock early but always
unlock in do_spu_create, fixing a different bug, but it looks like
he forgot this one in the process.

The reason why we originally had the unlock in the leaf functions is to
avoid a problem with spu_forget(), which had to be called without
the i_mutex held to avoid deadlocks.

> @@ -600,10 +591,6 @@ static int spufs_create_gang(struct inode *inode,
>  		int err = simple_rmdir(inode, dentry);
>  		WARN_ON(err);
>  	}
> -
> -out:
> -	mutex_unlock(&inode->i_mutex);
> -	dput(dentry);
>  	return ret;
>  }

Right, this obviously goes together with the one above,

> @@ -613,22 +600,21 @@ static struct file_system_type spufs_type;
>  long spufs_create(struct path *path, struct dentry *dentry,
>  		unsigned int flags, umode_t mode, struct file *filp)
>  {
> -	int ret;
> +	int ret = -EINVAL;
>  
> -	ret = -EINVAL;
>  	/* check if we are on spufs */
>  	if (path->dentry->d_sb->s_type != &spufs_type)
> -		goto out;
> +		goto fail;
>  
>  	/* don't accept undefined flags */
>  	if (flags & (~SPU_CREATE_FLAG_ALL))
> -		goto out;
> +		goto fail;
>  
>  	/* only threads can be underneath a gang */
>  	if (path->dentry != path->dentry->d_sb->s_root) {
>  		if ((flags & SPU_CREATE_GANG) ||
>  		    !SPUFS_I(path->dentry->d_inode)->i_gang)
> -			goto out;
> +		goto fail;
>  	}
>  
>  	mode &= ~current_umask();

These just change coding style, and not in a helpful way.

> @@ -640,12 +626,17 @@ long spufs_create(struct path *path, struct dentry *dentry,
>  		ret = spufs_create_context(path->dentry->d_inode,
>  					    dentry, path->mnt, flags, mode,
>  					    filp);
> -	if (ret >= 0)
> +	if (ret >= 0) {
> +		/* We drop these before fsnotify */
> +		mutex_unlock(&inode->i_mutex);
> +		dput(dentry);
>  		fsnotify_mkdir(path->dentry->d_inode, dentry);
> -	return ret;
>  
> -out:
> -	mutex_unlock(&path->dentry->d_inode->i_mutex);
> +		return ret;
> +	}
> + fail:
> +	mutex_unlock(&inode->i_mutex);
> +	dput(dentry);
>  	return ret;
>  }
>  
> diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c
> index 8591bb6..1a65ef2 100644
> --- a/arch/powerpc/platforms/cell/spufs/syscalls.c
> +++ b/arch/powerpc/platforms/cell/spufs/syscalls.c
> @@ -70,11 +70,8 @@ static long do_spu_create(const char __user *pathname, unsigned int flags,
>  	ret = PTR_ERR(dentry);
>  	if (!IS_ERR(dentry)) {
>  		ret = spufs_create(&path, dentry, flags, mode, neighbor);
> -		mutex_unlock(&path.dentry->d_inode->i_mutex);
> -		dput(dentry);
>  		path_put(&path);
>  	}
> -
>  	return ret;
>  }

This moves the unlock in front of the fsnotify_mkdir, where it was before Al's
change. This seems independent of the other change. 

	Arnd

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-07 12:48     ` Arnd Bergmann
@ 2012-03-07 21:01       ` Benjamin Herrenschmidt
  2012-03-07 21:23         ` Al Viro
  0 siblings, 1 reply; 7+ messages in thread
From: Benjamin Herrenschmidt @ 2012-03-07 21:01 UTC (permalink / raw)
  To: Arnd Bergmann; +Cc: linuxppc-dev, masterzorag, Al Viro


> > diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
> > index d4a094c..63b4e43 100644
> > --- a/arch/powerpc/platforms/cell/spufs/inode.c
> > +++ b/arch/powerpc/platforms/cell/spufs/inode.c
> > @@ -454,19 +454,16 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
> >  	struct spu_gang *gang;
> >  	struct spu_context *neighbor;
> >  
> > -	ret = -EPERM;
> >  	if ((flags & SPU_CREATE_NOSCHED) &&
> > -	    !capable(CAP_SYS_NICE))
> > -		goto out_unlock;
> > +	    !capable(CAP_SYa_NICE))
> 
>                     ^typo

Odd, probably an emacs fart

> > +		return -EPERM;
> >  
> > -	ret = -EINVAL;
> >  	if ((flags & (SPU_CREATE_NOSCHED | SPU_CREATE_ISOLATE))
> >  	    == SPU_CREATE_ISOLATE)
> > -		goto out_unlock;
> > +		return -EINVAL;
> >  
> > -	ret = -ENODEV;
> >  	if ((flags & SPU_CREATE_ISOLATE) && !isolated_loader)
> > -		goto out_unlock;
> > +		return -ENODEV;
> >  
> >  	gang = NULL;
> >  	neighbor = NULL;
> 
> This mostly changes coding style, pointlessly.

How so ? it's no longer necessary to store ret and goto out since there
is no cleanup at all, which makes things smaller/simpler. I wouldn't
call that 'pointlessly'.

I tend to dislike the use of "goto xxxx" when the xxx: label does no
cleanup whatsoever.

> > @@ -512,10 +509,6 @@ spufs_create_context(struct inode *inode, struct dentry *dentry,
> >  out_aff_unlock:
> >  	if (affinity)
> >  		mutex_unlock(&gang->aff_mutex);
> > -out_unlock:
> > -	mutex_unlock(&inode->i_mutex);
> > -out:
> > -	dput(dentry);
> >  	return ret;
> >  }
> 
> The original intention of this was to always unlock in the error case. It
> seems that Al changed this in 1ba10681 "switch do_spufs_create() to
> user_path_create(), fix double-unlock" to never unlock early but always
> unlock in do_spu_create, fixing a different bug, but it looks like
> he forgot this one in the process.
> 
> The reason why we originally had the unlock in the leaf functions is to
> avoid a problem with spu_forget(), which had to be called without
> the i_mutex held to avoid deadlocks.

Ok. I assumed it was fnotify that was the problem...

> > @@ -600,10 +591,6 @@ static int spufs_create_gang(struct inode *inode,
> >  		int err = simple_rmdir(inode, dentry);
> >  		WARN_ON(err);
> >  	}
> > -
> > -out:
> > -	mutex_unlock(&inode->i_mutex);
> > -	dput(dentry);
> >  	return ret;
> >  }
> 
> Right, this obviously goes together with the one above,

Yes, whatever they do they should do the same thing.

> > @@ -613,22 +600,21 @@ static struct file_system_type spufs_type;
> >  long spufs_create(struct path *path, struct dentry *dentry
> >  		unsigned int flags, umode_t mode, struct file *filp)
> >  {
> > -	int ret;
> > +	int ret = -EINVAL;
> >  
> > -	ret = -EINVAL;
> >  	/* check if we are on spufs */
> >  	if (path->dentry->d_sb->s_type != &spufs_type)
> > -		goto out;
> > +		goto fail;
> >  
> >  	/* don't accept undefined flags */
> >  	if (flags & (~SPU_CREATE_FLAG_ALL))
> > -		goto out;
> > +		goto fail;
> >  
> >  	/* only threads can be underneath a gang */
> >  	if (path->dentry != path->dentry->d_sb->s_root) {
> >  		if ((flags & SPU_CREATE_GANG) ||
> >  		    !SPUFS_I(path->dentry->d_inode)->i_gang)
> > -			goto out;
> > +		goto fail;
> >  	}
> >  
> >  	mode &= ~current_umask();
> 
> These just change coding style, and not in a helpful way.

Oh I just renamed the label, it's helpful because the label is
specifically only for the fail case, not the general exit path, hence
"fail" is a better naming than "out". It's about code clarity.

> > @@ -640,12 +626,17 @@ long spufs_create(struct path *path, struct dentry *dentry,
> >  		ret = spufs_create_context(path->dentry->d_inode,
> >  					    dentry, path->mnt, flags, mode,
> >  					    filp);
> > -	if (ret >= 0)
> > +	if (ret >= 0) {
> > +		/* We drop these before fsnotify */
> > +		mutex_unlock(&inode->i_mutex);
> > +		dput(dentry);
> >  		fsnotify_mkdir(path->dentry->d_inode, dentry);
> > -	return ret;
> >  
> > -out:
> > -	mutex_unlock(&path->dentry->d_inode->i_mutex);
> > +		return ret;
> > +	}
> > + fail:
> > +	mutex_unlock(&inode->i_mutex);
> > +	dput(dentry);
> >  	return ret;
> >  }
> >  
> > diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c
> > index 8591bb6..1a65ef2 100644
> > --- a/arch/powerpc/platforms/cell/spufs/syscalls.c
> > +++ b/arch/powerpc/platforms/cell/spufs/syscalls.c
> > @@ -70,11 +70,8 @@ static long do_spu_create(const char __user *pathname, unsigned int flags,
> >  	ret = PTR_ERR(dentry);
> >  	if (!IS_ERR(dentry)) {
> >  		ret = spufs_create(&path, dentry, flags, mode, neighbor);
> > -		mutex_unlock(&path.dentry->d_inode->i_mutex);
> > -		dput(dentry);
> >  		path_put(&path);
> >  	}
> > -
> >  	return ret;
> >  }
> 
> This moves the unlock in front of the fsnotify_mkdir, where it was before Al's
> change. This seems independent of the other change. 

No it's not, it all goes together. spufs_create_context() always
unlocked & dropped the dentry before returning, so I assumed the
lock had to be dropped before fsnotify.

Note that if the problem is that the lock has to be dropped before
spu_forget(), then we should indeed move it back into the leaf functions
and just remove all the unlock path from the top ones. It's a bit nasty
how we drop the mutex first, then do spu_forget, then drop the dentry
but we could go back to doing that.

What I want is consistent semantics. It's just silly to have 3 different
stacking levels which all 3 may or may not be responsible to dropping
the lock & dentry depending on circumstances.

In any case, what about this instead then:

diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
index d4a094c..114ab14 100644
--- a/arch/powerpc/platforms/cell/spufs/inode.c
+++ b/arch/powerpc/platforms/cell/spufs/inode.c
@@ -646,6 +646,7 @@ long spufs_create(struct path *path, struct dentry *dentry,
 
 out:
 	mutex_unlock(&path->dentry->d_inode->i_mutex);
+	dput(dentry);
 	return ret;
 }
 
diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c
index 8591bb6..5665dcc 100644
--- a/arch/powerpc/platforms/cell/spufs/syscalls.c
+++ b/arch/powerpc/platforms/cell/spufs/syscalls.c
@@ -70,8 +70,6 @@ static long do_spu_create(const char __user *pathname, unsigned int flags,
 	ret = PTR_ERR(dentry);
 	if (!IS_ERR(dentry)) {
 		ret = spufs_create(&path, dentry, flags, mode, neighbor);
-		mutex_unlock(&path.dentry->d_inode->i_mutex);
-		dput(dentry);
 		path_put(&path);
 	}
 
Cheers,
Ben.

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-07 21:01       ` Benjamin Herrenschmidt
@ 2012-03-07 21:23         ` Al Viro
  2012-03-07 22:32           ` Arnd Bergmann
  0 siblings, 1 reply; 7+ messages in thread
From: Al Viro @ 2012-03-07 21:23 UTC (permalink / raw)
  To: Benjamin Herrenschmidt; +Cc: linuxppc-dev, masterzorag, Arnd Bergmann


> No it's not, it all goes together. spufs_create_context() always
> unlocked & dropped the dentry before returning, so I assumed the
> lock had to be dropped before fsnotify.
> 
> Note that if the problem is that the lock has to be dropped before
> spu_forget(), then we should indeed move it back into the leaf functions
> and just remove all the unlock path from the top ones. It's a bit nasty
> how we drop the mutex first, then do spu_forget, then drop the dentry
> but we could go back to doing that.
> 
> What I want is consistent semantics. It's just silly to have 3 different
> stacking levels which all 3 may or may not be responsible to dropping
> the lock & dentry depending on circumstances.

Why not leave unlock/dput to the caller?  Details of deadlocks caused
by that approach, please...

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] spufs raises two exceptions
  2012-03-07 21:23         ` Al Viro
@ 2012-03-07 22:32           ` Arnd Bergmann
  0 siblings, 0 replies; 7+ messages in thread
From: Arnd Bergmann @ 2012-03-07 22:32 UTC (permalink / raw)
  To: Al Viro; +Cc: linuxppc-dev, masterzorag

On Wednesday 07 March 2012, Al Viro wrote:
> > No it's not, it all goes together. spufs_create_context() always
> > unlocked & dropped the dentry before returning, so I assumed the
> > lock had to be dropped before fsnotify.
> > 
> > Note that if the problem is that the lock has to be dropped before
> > spu_forget(), then we should indeed move it back into the leaf functions
> > and just remove all the unlock path from the top ones. It's a bit nasty
> > how we drop the mutex first, then do spu_forget, then drop the dentry
> > but we could go back to doing that.
> > 
> > What I want is consistent semantics. It's just silly to have 3 different
> > stacking levels which all 3 may or may not be responsible to dropping
> > the lock & dentry depending on circumstances.
> 
> Why not leave unlock/dput to the caller?  Details of deadlocks caused
> by that approach, please...

If only I could remember that part exactly. I think I was remembering
0309f02d8 "spufs: fix deadlock in spu_create error path", which had a
double lock problem in this path. Please bear with me here because it's
been six years since I debugged that particular problem.

A lot of the nastyness of spu_forget() is about the problem of having
to give up a reference on current->mm that is held by an spu context,
while the mm contains a vma that maps this context and holds a refence
on the inode, in particular in case of calling exit().

However, I don't see now how either of these two issues requires that
we call spu_forget without the i_mutex held during a context creation
failure.

	Arnd

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2012-03-07 22:38 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-03-06  9:26 [PATCH] spufs raises two exceptions masterzorag
2012-03-07  3:49 ` Benjamin Herrenschmidt
2012-03-07  3:51   ` Benjamin Herrenschmidt
2012-03-07 12:48     ` Arnd Bergmann
2012-03-07 21:01       ` Benjamin Herrenschmidt
2012-03-07 21:23         ` Al Viro
2012-03-07 22:32           ` Arnd Bergmann

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).