linux-kernel - Re: [PATCH] autofs: don't stuck in a loop if vfs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CANaxB-y93uqDzt2+VPSTBtw7bohSgM-OT_XYU1qC+N=Zu=ECDw@mail.gmail.com>
Date:	Thu, 9 Jun 2016 10:27:48 -0700
From:	Andrei Vagin <avagin@...il.com>
To:	Ian Kent <raven@...maw.net>
Cc:	autofs@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH] autofs: don't stuck in a loop if vfs_write returns an error

On Wed, Jun 8, 2016 at 6:23 PM, Ian Kent <raven@...maw.net> wrote:
> On Mon, 2016-05-30 at 13:52 +0800, Ian Kent wrote:
>> On Tue, 2016-05-24 at 09:34 +0800, Ian Kent wrote:
>> > On Mon, 2016-05-23 at 14:50 -0700, Andrei Vagin wrote:
>> > > Hi Ian,
>> > >
>> > > When are you going to apply this patch? We can't test linux-next without
>> > > it.
>> >
>> > I though I sent this with the last series but I can't see that I have.
>> >
>> > I have the rest of that series to send over to Andrew which I was planning
>> > to
>> > do
>> > after the current merge window closes (which is about now I guess).
>> >
>> > I'll include it in that series.
>> > Sorry for the hold up, ;)
>>
>> Some bad news I'm afraid.
>>
>> I was getting ready to send these over to Andrew and found that opendir(3) is
>> failing on a number of tests (51 of 230, 9 fails are expected) with 4.6.0.
>>
>> It's not the patches, yours or mine and it doesn't happen with 4.4.x kernels.
>>
>> Looks like I'm going to have to bisect to work out what's going on and that
>> will
>> take a while.
>
> The regression has been fixed now.
>
> Al Viro sent a patch for it to Linus yesterday, it's commit e6ec03a25f1 in the
> Linux tree.

It's good!

>
> I can send my patches to Andrew (after re-testing) but any autofs related
> testing of linux.next will need the above commit.
>
> Andrew, surely this isn't the first time this type of problem has happened, how
> is it usually handled, what do I need to do to make this go smoothly?

In linux-next we catch two sorts of bugs.

1. If bugs is triggered very often, we report it, when it isn't in
Linus' tree. And such bugs are fixed very fast.
2. If bugs is triggered rarely. In this case we may detect this bug
too late, when it's in Linus' tree. In this case we may need to
workaraound it.

Here the problem belongs to the second type. It is triggered only when
one or more tests failed and we try to kill all test processes.
Actually it doesn't affect regular runs of tests, but it's annoying
when we are investigating something.

Ian, I don't think that you need to do anything special. Thank you for
handling this patch!

Thanks,
Andrew

>
>>
>> >
>> > Ian
>> >
>> > >
>> > > Thanks,
>> > > Andrew
>> > >
>> > >
>> > > On Fri, Apr 1, 2016 at 12:37 AM, Ian Kent <raven@...maw.net> wrote:
>> > > > On Thu, 2016-03-31 at 22:12 -0700, Andrey Vagin wrote:
>> > > > > From: Andrey Vagin <avagin@...nvz.org>
>> > > > >
>> > > > > __vfs_write() returns a negative value in a error case.
>> > > >
>> > > > Ha, right, I'll send this along to Andrew with my next series which
>> > > > should be soon.
>> > > >
>> > > > >
>> > > > > Cc: Ian Kent <raven@...maw.net>
>> > > > > Signed-off-by: Andrey Vagin <avagin@...nvz.org>
>> > > > > ---
>> > > > >  fs/autofs4/waitq.c | 7 ++++---
>> > > > >  1 file changed, 4 insertions(+), 3 deletions(-)
>> > > > >
>> > > > > diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c
>> > > > > index 0146d91..631f155 100644
>> > > > > --- a/fs/autofs4/waitq.c
>> > > > > +++ b/fs/autofs4/waitq.c
>> > > > > @@ -66,11 +66,12 @@ static int autofs4_write(struct autofs_sb_info
>> > > > > *sbi,
>> > > > >       set_fs(KERNEL_DS);
>> > > > >
>> > > > >       mutex_lock(&sbi->pipe_mutex);
>> > > > > -     wr = __vfs_write(file, data, bytes, &file->f_pos);
>> > > > > -     while (bytes && wr) {
>> > > > > +     while (bytes) {
>> > > > > +             wr = __vfs_write(file, data, bytes, &file->f_pos);
>> > > > > +             if (wr <= 0)
>> > > > > +                     break;
>> > > > >               data += wr;
>> > > > >               bytes -= wr;
>> > > > > -             wr = __vfs_write(file, data, bytes, &file->f_pos);
>> > > > >       }
>> > > > >       mutex_unlock(&sbi->pipe_mutex);
>> --
>> To unsubscribe from this list: send the line "unsubscribe autofs" in