Re: ffsb job does not exit on xfs 4.14-rc1+

[Xiong Zhou]

On Thu, Oct 12, 2017 at 08:10:10AM +1100, Dave Chinner wrote:
> On Wed, Oct 11, 2017 at 09:54:15PM +0800, Xiong Zhou wrote:
> > On Mon, Sep 25, 2017 at 10:49:03AM +0200, Carlos Maiolino wrote:
> > > On Mon, Sep 25, 2017 at 01:40:06AM +, Xiong Zhou wrote:
> > > > Hi,
> > > > 
> > > > ffsb test won't exit like this on Linus tree 4.14-rc1+.
> > > > Latest commit cd4175b11685

Hi all,

This is a false alarm because I've messed up the kernel config.
After making .config right, test does not hang.

My fault. Very Sorry for the noise.

> > > 
> > > Can you provide more information? Do you have any kernel log from this 
> > > issue?
> > > dmesg, Oopses, traces, etc.
> > > Storage configuration might also be required here.
> > 
> > Turns out this only repreduces on nvdimm devices, xfs without dax
> > mount option. More logs are attached.
> 
> It's a hang, so what's the output of sysrq-w once it's hung?
> 
> > > have you also tried to reproduce it with another filesystem? If so, is 
> > > the same
> > > problem reproducible with another filesystem or only with XFS?
> > 
> > Only xfs. Test on ext4 ends shortly.
> > 
> > > 
> > > P.S. please avoid sending it to all lists (mainly LKML).
> > 
> > Why?  I thought LKML was better archived.
> 
> The lists are all archived, but that's irrelevant. The list you
> should report problems to is based on the scope of the problem, not
> whether the lists are archived or not.
> 
> Because the scope of the problem at this point is XFS, it's
> inappropriate to report it to lists that are for general kernel or
> VFS issues. There's enough noise on those lists without everyone
> bombarding them with subsystem specific issues - that's why we have
> subsystem specific lists in the first place.
> 
> If it turns out to be a problem in some other subsystem, we'll add
> cc's to other subsystem or general lists as appropriate.

Fair enough. Thanks very much for your time!

Xiong

> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> da...@fromorbit.com
___
Linux-nvdimm mailing list
Linux-nvdimm@lists.01.org
https://lists.01.org/mailman/listinfo/linux-nvdimm

Re: ffsb job does not exit on xfs 4.14-rc1+

[Dave Chinner]

On Wed, Oct 11, 2017 at 09:54:15PM +0800, Xiong Zhou wrote:
> On Mon, Sep 25, 2017 at 10:49:03AM +0200, Carlos Maiolino wrote:
> > On Mon, Sep 25, 2017 at 01:40:06AM +, Xiong Zhou wrote:
> > > Hi,
> > > 
> > > ffsb test won't exit like this on Linus tree 4.14-rc1+.
> > > Latest commit cd4175b11685
> > 
> > Can you provide more information? Do you have any kernel log from this 
> > issue?
> > dmesg, Oopses, traces, etc.
> > Storage configuration might also be required here.
> 
> Turns out this only repreduces on nvdimm devices, xfs without dax
> mount option. More logs are attached.

It's a hang, so what's the output of sysrq-w once it's hung?

> > have you also tried to reproduce it with another filesystem? If so, is the 
> > same
> > problem reproducible with another filesystem or only with XFS?
> 
> Only xfs. Test on ext4 ends shortly.
> 
> > 
> > P.S. please avoid sending it to all lists (mainly LKML).
> 
> Why?  I thought LKML was better archived.

The lists are all archived, but that's irrelevant. The list you
should report problems to is based on the scope of the problem, not
whether the lists are archived or not.

Because the scope of the problem at this point is XFS, it's
inappropriate to report it to lists that are for general kernel or
VFS issues. There's enough noise on those lists without everyone
bombarding them with subsystem specific issues - that's why we have
subsystem specific lists in the first place.

If it turns out to be a problem in some other subsystem, we'll add
cc's to other subsystem or general lists as appropriate.

Cheers,

Dave.
-- 
Dave Chinner
da...@fromorbit.com
___
Linux-nvdimm mailing list
Linux-nvdimm@lists.01.org
https://lists.01.org/mailman/listinfo/linux-nvdimm

Re: ffsb job does not exit on xfs 4.14-rc1+

[Dan Williams]

On Wed, Oct 11, 2017 at 6:54 AM, Xiong Zhou  wrote:
> On Mon, Sep 25, 2017 at 10:49:03AM +0200, Carlos Maiolino wrote:
>> On Mon, Sep 25, 2017 at 01:40:06AM +, Xiong Zhou wrote:
>> > Hi,
>> >
>> > ffsb test won't exit like this on Linus tree 4.14-rc1+.
>> > Latest commit cd4175b11685
>>
>> Can you provide more information? Do you have any kernel log from this issue?
>> dmesg, Oopses, traces, etc.
>> Storage configuration might also be required here.
>
> Turns out this only repreduces on nvdimm devices, xfs without dax
> mount option. More logs are attached.
>
>>
>> have you also tried to reproduce it with another filesystem? If so, is the 
>> same
>> problem reproducible with another filesystem or only with XFS?
>
> Only xfs. Test on ext4 ends shortly.
>
>>
>> P.S. please avoid sending it to all lists (mainly LKML).
>
> Why?  I thought LKML was better archived.
>
>> Unless it's a more generic kernel problem, keep it in fsdevel and/or the
>> respective filesystem list if it's related to a single filesystem only.
>
> 4.14-rc1+ won't survive this script, while 4.13 can.

Can you try a git bisect?
___
Linux-nvdimm mailing list
Linux-nvdimm@lists.01.org
https://lists.01.org/mailman/listinfo/linux-nvdimm