[MPTCP] Re: WARNING in warn_bad_map

[MPTCP] Re: WARNING in warn_bad_map

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1954 bytes --]

On Tue, 2020-07-07 at 17:47 +0200, Dmitry Vyukov wrote:
> On Tue, Jul 7, 2020 at 5:18 PM Paolo Abeni <pabeni(a)redhat.com> wrote:
> > [stripping some recipients, since the following should be syzbot-
> > specific]
> > 
> > On Mon, 2020-06-22 at 02:30 -0700, syzbot wrote:
> > > syzbot found the following crash on:
> > > 
> > > HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> > > git tree:       upstream
> > > console output: https://syzkaller.appspot.com/x/log.txt?x=11124521100000
> > > kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> > > dashboard link: https://syzkaller.appspot.com/bug?extid=42a07faa5923cfaeb9c9
> > > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > > 
> > > Unfortunately, I don't have any reproducer for this crash yet.
> > > 
> > > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > > Reported-by: syzbot+42a07faa5923cfaeb9c9(a)syzkaller.appspotmail.com
> > 
> > This should be fixed by the combination of commit 0e7bc442b405 ("mptcp:
> > fix DSS map generation on fin retransmission") (currently on -net) and
> > commit 13f1555c1b4c49b8ab9dd80b7353deec0526ccac (currently only on net-
> > next).
> > 
> > In not clear to me how to communicate the above to syzbot. Any
> > suggestion more than welcome!
> 
> Hi Paolo,
> 
> Yes, there is no way to attach to fixing commits via email interface.
> But if both are required to stop this WARNING from happening, we can
> mark it as fixed by the one that is supposed to propagate to all trees
> later. I guess it's the one in net-next?

Thank you for the prompt reply!

Yes, the net-next commit will land into vanilla later - in some weeks
from now.

Should I wait for that event to send a '#syz fix: ' reply? I fear
otherwise syzbot may test a [net-next] tree without the first mentioned
commit and still hit the issue.

Cheers,

Paolo

[MPTCP] Re: WARNING in warn_bad_map

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1199 bytes --]

[sorry for lagging behind]
On Wed, 2020-07-08 at 10:42 +0200, Dmitry Vyukov wrote:
> On Tue, Jul 7, 2020 at 6:39 PM Paolo Abeni <pabeni(a)redhat.com> wrote:
> > Should I wait for that event to send a '#syz fix: ' reply? I fear
> > otherwise syzbot may test a [net-next] tree without the first mentioned
> > commit and still hit the issue.
> 
> https://goo.gl/tpsmEJ#bug-status-tracking should answer your question
> (or we need to improve the docs).

Indeed it answer in a clear way, thanks for the pointer

> I see the commit you mentioned
> 13f1555c1b4c49b8ab9dd80b7353deec0526ccac is actually a merge:
>      Merge branch 'MPTCP-improve-fallback-to-TCP'
> I guess what fixes it is a particular non-merge commit, right? What is
> it? The syzbot referenced commits are considered for stable trees, and
> a merge is not very useful for backporting.

The relevant non-merge commit should be:

commit e1ff9e82e2ea53d01540692a85c16a77e1089537
Author: Davide Caratti <dcaratti(a)redhat.com>
Date:   Mon Jun 29 22:26:20 2020 +0200

    net: mptcp: improve fallback to TCP

but that is part of a code re-factor, really not suitable for stable
branch IMHO.

Thanks,

Paolo

[MPTCP] Re: WARNING in warn_bad_map

[Dmitry Vyukov]

[-- Attachment #1: Type: text/plain, Size: 2528 bytes --]

On Tue, Jul 7, 2020 at 6:39 PM Paolo Abeni <pabeni(a)redhat.com> wrote:
>
> On Tue, 2020-07-07 at 17:47 +0200, Dmitry Vyukov wrote:
> > On Tue, Jul 7, 2020 at 5:18 PM Paolo Abeni <pabeni(a)redhat.com> wrote:
> > > [stripping some recipients, since the following should be syzbot-
> > > specific]
> > >
> > > On Mon, 2020-06-22 at 02:30 -0700, syzbot wrote:
> > > > syzbot found the following crash on:
> > > >
> > > > HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> > > > git tree:       upstream
> > > > console output: https://syzkaller.appspot.com/x/log.txt?x=11124521100000
> > > > kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> > > > dashboard link: https://syzkaller.appspot.com/bug?extid=42a07faa5923cfaeb9c9
> > > > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > > >
> > > > Unfortunately, I don't have any reproducer for this crash yet.
> > > >
> > > > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > > > Reported-by: syzbot+42a07faa5923cfaeb9c9(a)syzkaller.appspotmail.com
> > >
> > > This should be fixed by the combination of commit 0e7bc442b405 ("mptcp:
> > > fix DSS map generation on fin retransmission") (currently on -net) and
> > > commit 13f1555c1b4c49b8ab9dd80b7353deec0526ccac (currently only on net-
> > > next).
> > >
> > > In not clear to me how to communicate the above to syzbot. Any
> > > suggestion more than welcome!
> >
> > Hi Paolo,
> >
> > Yes, there is no way to attach to fixing commits via email interface.
> > But if both are required to stop this WARNING from happening, we can
> > mark it as fixed by the one that is supposed to propagate to all trees
> > later. I guess it's the one in net-next?
>
> Thank you for the prompt reply!
>
> Yes, the net-next commit will land into vanilla later - in some weeks
> from now.
>
> Should I wait for that event to send a '#syz fix: ' reply? I fear
> otherwise syzbot may test a [net-next] tree without the first mentioned
> commit and still hit the issue.

https://goo.gl/tpsmEJ#bug-status-tracking should answer your question
(or we need to improve the docs).

I see the commit you mentioned
13f1555c1b4c49b8ab9dd80b7353deec0526ccac is actually a merge:
     Merge branch 'MPTCP-improve-fallback-to-TCP'
I guess what fixes it is a particular non-merge commit, right? What is
it? The syzbot referenced commits are considered for stable trees, and
a merge is not very useful for backporting.

[MPTCP] Re: WARNING in warn_bad_map

[Dmitry Vyukov]

[-- Attachment #1: Type: text/plain, Size: 1512 bytes --]

On Tue, Jul 7, 2020 at 5:18 PM Paolo Abeni <pabeni(a)redhat.com> wrote:
>
> [stripping some recipients, since the following should be syzbot-
> specific]
>
> On Mon, 2020-06-22 at 02:30 -0700, syzbot wrote:
> > syzbot found the following crash on:
> >
> > HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=11124521100000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> > dashboard link: https://syzkaller.appspot.com/bug?extid=42a07faa5923cfaeb9c9
> > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> >
> > Unfortunately, I don't have any reproducer for this crash yet.
> >
> > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > Reported-by: syzbot+42a07faa5923cfaeb9c9(a)syzkaller.appspotmail.com
>
> This should be fixed by the combination of commit 0e7bc442b405 ("mptcp:
> fix DSS map generation on fin retransmission") (currently on -net) and
> commit 13f1555c1b4c49b8ab9dd80b7353deec0526ccac (currently only on net-
> next).
>
> In not clear to me how to communicate the above to syzbot. Any
> suggestion more than welcome!

Hi Paolo,

Yes, there is no way to attach to fixing commits via email interface.
But if both are required to stop this WARNING from happening, we can
mark it as fixed by the one that is supposed to propagate to all trees
later. I guess it's the one in net-next?

[MPTCP] Re: WARNING in warn_bad_map

[Paolo Abeni]

[-- Attachment #1: Type: text/plain, Size: 1144 bytes --]

[stripping some recipients, since the following should be syzbot-
specific]

On Mon, 2020-06-22 at 02:30 -0700, syzbot wrote:
> syzbot found the following crash on:
> 
> HEAD commit:    7ae77150 Merge tag 'powerpc-5.8-1' of git://git.kernel.org..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=11124521100000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=d195fe572fb15312
> dashboard link: https://syzkaller.appspot.com/bug?extid=42a07faa5923cfaeb9c9
> compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> 
> Unfortunately, I don't have any reproducer for this crash yet.
> 
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+42a07faa5923cfaeb9c9(a)syzkaller.appspotmail.com

This should be fixed by the combination of commit 0e7bc442b405 ("mptcp:
fix DSS map generation on fin retransmission") (currently on -net) and
commit 13f1555c1b4c49b8ab9dd80b7353deec0526ccac (currently only on net-
next).

In not clear to me how to communicate the above to syzbot. Any
suggestion more than welcome!

Thanks,

Paolo