Re: 3.9.0: WARNING: at drivers/base/core.c:575

[Felipe Balbi]

Hi,

On Fri, May 10, 2013 at 09:51:00AM +0800, Dave Young wrote:
> On Thu, May 09, 2013 at 05:53:08PM +0300, Felipe Balbi wrote:
> > On Thu, May 09, 2013 at 06:29:51PM +0800, Dave Young wrote:
> > > On Tue, May 07, 2013 at 03:25:58AM -0400, CAI Qian wrote:
> > > > Never saw any of those messages were floating in any of the RC testing, but
> > > > now happened in 3.9 GA on Power 7 systems.
> > > > 
> > > > [    0.329753] EEH: devices created
> > > > [    0.340203] atomic64 test passed
> > > > [    0.340407] NET: Registered protocol family 16
> > > > [    0.340457] EEH: No capable adapters found
> > > > [    0.340609] IBM eBus Device Driver
> > > > [    0.358825] Write permission without 'store'
> > > > [    0.358852] ------------[ cut here ]------------
> > > > [    0.358859] WARNING: at drivers/base/core.c:575
> > > > [    0.358866] Modules linked in:
> > > > [    0.358877] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.9.0+ #1
> > > > [    0.358887] task: c0000003f8bc0000 ti: c0000003f8c00000 task.ti: c0000003f8c00000
> > > > [    0.358897] NIP: c0000000004a7f44 LR: c0000000004a7f40 CTR: 0000000001766760
> > > > [    0.358906] REGS: c0000003f8c03780 TRAP: 0700   Not tainted  (3.9.0+)
> > > > [    0.358914] MSR: 8000000000029032 <SF,EE,ME,IR,DR,RI>  CR: 28139d24  XER: 00000001
> > > > [    0.358939] SOFTE: 1
> > > > [    0.358944] CFAR: c00000000074427c
> > > > [    0.358950] 
> > > > GPR00: c0000000004a7f40 c0000003f8c03a00 c000000001117e50 0000000000000020 
> > > > GPR04: 0000000000000000 0000000000000000 00000000160c9615 0000000000000000 
> > > > GPR08: c000000001064af0 0000000000000000 0000000000000000 0000000000003fef 
> > > > GPR12: 0000000028139d22 c000000007f00000 c00000000000bb30 0000000000000000 
> > > > GPR16: 0000000000000000 0000000000000000 c000000000ff1c78 0000000000000001 
> > > > GPR20: 0000000000000000 0000000000000000 0000000000000000 c00000000117eb08 
> > > > GPR24: 00000000000000a6 c000000000af0028 0000000000000000 c000000000ff21a0 
> > > > GPR28: 0000000000000002 c000000001500030 c0000000011b8200 c000000001500030 
> > > > [    0.359070] NIP [c0000000004a7f44] .device_create_file+0xa4/0xe0
> > > > [    0.359080] LR [c0000000004a7f40] .device_create_file+0xa0/0xe0
> > > > [    0.359088] Call Trace:
> > > > [    0.359095] [c0000003f8c03a00] [c0000000004a7f40] .device_create_file+0xa0/0xe0 (unreliable)
> > > > [    0.359112] [c0000003f8c03a90] [c00000000074daf8] .register_cpu_online+0x21c/0x284
> > > > [    0.359126] [c0000003f8c03b30] [c000000000a78308] .topology_init+0x130/0x200
> > > > [    0.359138] [c0000003f8c03c10] [c00000000000b4d4] .do_one_initcall+0x144/0x1f0
> > > > [    0.359150] [c0000003f8c03cd0] [c000000000a746a8] .kernel_init_freeable+0x23c/0x32c
> > > > [    0.359163] [c0000003f8c03db0] [c00000000000bb4c] .kernel_init+0x1c/0x130
> > > > [    0.359175] [c0000003f8c03e30] [c00000000000a164] .ret_from_kernel_thread+0x64/0x80
> > > > [    0.359184] Instruction dump:
> > > > [    0.359190] 3863e610 4829c2f1 60000000 0fe00000 e8810070 4bffffb0 60000000 3c62ff8c 
> > > > [    0.359211] f8810070 3863e5e8 4829c2cd 60000000 <0fe00000> e8810070 a1240008 4bffff80 
> > > > [    0.359241] ---[ end trace c949de754c984735 ]---
> > > > 
> > > Hi, Caiqian
> > > 
> > > Could you try below debug patch to print the attr name?
> > > 
> > > ---
> > > print the sysfs attribute file name when we warn about bogus permissions 
> > > 
> > > Make it obvious to see what attribute is using bogus permissions.
> > > 
> > > Signed-off-by: Dave Young <dyoung@xxxxxxxxxx>
> > > ---
> > >  drivers/base/core.c |    6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > > 
> > > --- linux-2.6.orig/drivers/base/core.c
> > > +++ linux-2.6/drivers/base/core.c
> > > @@ -572,9 +572,11 @@ int device_create_file(struct device *de
> > >  
> > >  	if (dev) {
> > >  		WARN(((attr->attr.mode & S_IWUGO) && !attr->store),
> > > -				"Write permission without 'store'\n");
> > > +			"Attribute "%s": write permission without 'store'\n",
> > 
> > sure this even builds ?
> > 
> 
> Hi, Felipe
> 
> Oops, here is the fix:
> ---
> 
> print the sysfs attribute file name when we warn about bogus permissions 
> 
> Make it obvious to see what attribute is using bogus permissions.
> 
> [v1->v2] build fix
> 
> Signed-off-by: Dave Young <dyoung@xxxxxxxxxx>

looks alright to me:

Acked-by: Felipe Balbi <balbi@xxxxxx>

But you need to resend as a proper patch so Greg can apply. Also, the
v1->v2 note shoudln't be in the commit log area.

cheers

> ---
>  drivers/base/core.c |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> --- linux-2.6.orig/drivers/base/core.c
> +++ linux-2.6/drivers/base/core.c
> @@ -572,9 +572,11 @@ int device_create_file(struct device *de
>  
>  	if (dev) {
>  		WARN(((attr->attr.mode & S_IWUGO) && !attr->store),
> -				"Write permission without 'store'\n");
> +			"Attribute %s: write permission without 'store'\n",
> +			attr->attr.name);
>  		WARN(((attr->attr.mode & S_IRUGO) && !attr->show),
> -				"Read permission without 'show'\n");
> +			"Attribute %s: Read permission without 'show'\n",
> +			attr->attr.name);
>  		error = sysfs_create_file(&dev->kobj, &attr->attr);
>  	}
>  
> 

-- 
balbi

Attachment: signature.asc
Description: Digital signature