Hi John,
A important fix to 3.10, this patch fixes an issues that was preventing the
l2cap info response command to be handled properly. Please pull. Thanks!
Gustavo
---
The following changes since commit fcb3701849957917a234a61b58ad70ed35c83eda:
brcmfmac: free primary net_device when brcmf_bus_start() fails (2013-06-13 13:24:12 -0400)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth master
for you to fetch changes up to 3f6fa3d489e127ca5a5b298eabac3ff5dbe0e112:
Bluetooth: Fix invalid length check in l2cap_information_rsp() (2013-06-23 00:24:58 +0100)
----------------------------------------------------------------
Jaganath Kanakkassery (1):
Bluetooth: Fix invalid length check in l2cap_information_rsp()
net/bluetooth/l2cap_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
* Pavel Machek <[email protected]> [2013-06-26 13:52:55 +0200]:
> On Tue 2013-06-25 22:00:39, Gustavo Padovan wrote:
> > Hi Marcel,
> >=20
> > * Marcel Holtmann <[email protected]> [2013-06-25 13:49:19 -0700]:
> >=20
> > > Hi Pavel,
> > >=20
> > > >> Jaganath Kanakkassery (1):
> > > >> Bluetooth: Fix invalid length check in l2cap_information_rsp()
> > > >=20
> > > > Could we add the "3.10 vs. thinkpad x60" patch that fixes regressio=
n?
> > > > It is from Johan Hedberg=E2=80=A6
> > >=20
> > > that one went into net-next tree. It is scheduled for 3.11 right now.
> > >=20
> > > After 3.10 is out, maybe it can go into -stable, but otherwise I thin=
k it is bit late now. Unless someone wants to cherry-pick it.
> >=20
> > That patch is actually in 3.10 already, it was included in the last pull
> > request and John took it to wireless.git.
>=20
> Is it? I tried pulling yesterday (some updates) and today (no
> changes), am on:
Dave already pulled it so it should land in mainline anytime soon.
Gustavo
On Tue 2013-06-25 22:00:39, Gustavo Padovan wrote:
> Hi Marcel,
>
> * Marcel Holtmann <[email protected]> [2013-06-25 13:49:19 -0700]:
>
> > Hi Pavel,
> >
> > >> Jaganath Kanakkassery (1):
> > >> Bluetooth: Fix invalid length check in l2cap_information_rsp()
> > >
> > > Could we add the "3.10 vs. thinkpad x60" patch that fixes regression?
> > > It is from Johan Hedberg…
> >
> > that one went into net-next tree. It is scheduled for 3.11 right now.
> >
> > After 3.10 is out, maybe it can go into -stable, but otherwise I think it is bit late now. Unless someone wants to cherry-pick it.
>
> That patch is actually in 3.10 already, it was included in the last pull
> request and John took it to wireless.git.
Is it? I tried pulling yesterday (some updates) and today (no
changes), am on:
Linux amd 3.10.0-rc7+ #267 SMP Tue Jun 25 23:32:24 CEST 2013 i686
GNU/Linux
and bluetooth is still broken :-(.
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi Marcel,
* Marcel Holtmann <[email protected]> [2013-06-25 13:49:19 -0700]:
> Hi Pavel,
>=20
> >> Jaganath Kanakkassery (1):
> >> Bluetooth: Fix invalid length check in l2cap_information_rsp()
> >=20
> > Could we add the "3.10 vs. thinkpad x60" patch that fixes regression?
> > It is from Johan Hedberg=E2=80=A6
>=20
> that one went into net-next tree. It is scheduled for 3.11 right now.
>=20
> After 3.10 is out, maybe it can go into -stable, but otherwise I think it=
is bit late now. Unless someone wants to cherry-pick it.
That patch is actually in 3.10 already, it was included in the last pull
request and John took it to wireless.git.
Gustavo
Hi Pavel,
>> Jaganath Kanakkassery (1):
>> Bluetooth: Fix invalid length check in l2cap_information_rsp()
>
> Could we add the "3.10 vs. thinkpad x60" patch that fixes regression?
> It is from Johan Hedberg?
that one went into net-next tree. It is scheduled for 3.11 right now.
After 3.10 is out, maybe it can go into -stable, but otherwise I think it is bit late now. Unless someone wants to cherry-pick it.
Regards
Marcel
Hi!
> Jaganath Kanakkassery (1):
> Bluetooth: Fix invalid length check in l2cap_information_rsp()
Could we add the "3.10 vs. thinkpad x60" patch that fixes regression?
It is from Johan Hedberg...
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
On Sun, Jun 23, 2013 at 03:13:51AM +0100, Gustavo Padovan wrote:
> Hi John,
>
> A important fix to 3.10, this patch fixes an issues that was preventing the
> l2cap info response command to be handled properly. Please pull. Thanks!
>
> Gustavo
>
> ---
> The following changes since commit fcb3701849957917a234a61b58ad70ed35c83eda:
>
> brcmfmac: free primary net_device when brcmf_bus_start() fails (2013-06-13 13:24:12 -0400)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth master
>
> for you to fetch changes up to 3f6fa3d489e127ca5a5b298eabac3ff5dbe0e112:
>
> Bluetooth: Fix invalid length check in l2cap_information_rsp() (2013-06-23 00:24:58 +0100)
>
> ----------------------------------------------------------------
> Jaganath Kanakkassery (1):
> Bluetooth: Fix invalid length check in l2cap_information_rsp()
>
> net/bluetooth/l2cap_core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
Pulling now...
--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.
Hi Gustavo,
>> A important fix to 3.10, this patch fixes an issues that was preventing the
>> l2cap info response command to be handled properly. Please pull. Thanks!
>
> I'd like to add here (since it's not obvious) that this is not just
> about failing to parse the L2CAP info response. Once the code gives up
> parsing this PDU it also gives up essential parts of the L2CAP
> connection creation process, i.e. without this patch the stack will fail
> to establish connections properly. So it'd be very good if we can get
> this into 3.10 (as well as the stable trees where the patch that
> introduced this bug is already on its way).
I moved this to linux-bluetooth only now to not derail the process of getting this fixed upstream, but why was this information not part of the commit message.
Commit messages are there to explain the impact in plain English. I really mean plain English and not assuming everybody understands Bluetooth and can read HCI traces.
Especially commit message for patches after the merge window closed should be verbose. In addition the cover letter to John needs to explain why this is important and justifies inclusion after -rc7. Why is this a regression that needs to be fixed right away.
If I would be in John's shoes, I would not classify this as an important enough fix to go in after -rc7. The cover letter does not tell him the impact and the commit message does not tell him anything either. Neither John, Dave or Linux know enough about L2CAP to come to the conclusion that this a regression and will break Bluetooth connections.
Regards
Marcel
Hi,
On Sun, Jun 23, 2013, Gustavo Padovan wrote:
> A important fix to 3.10, this patch fixes an issues that was preventing the
> l2cap info response command to be handled properly. Please pull. Thanks!
I'd like to add here (since it's not obvious) that this is not just
about failing to parse the L2CAP info response. Once the code gives up
parsing this PDU it also gives up essential parts of the L2CAP
connection creation process, i.e. without this patch the stack will fail
to establish connections properly. So it'd be very good if we can get
this into 3.10 (as well as the stable trees where the patch that
introduced this bug is already on its way).
Johan
>On Wed 2013-06-26 13:47:36, Gustavo Padovan wrote:
> * Pavel Machek <[email protected]> [2013-06-26 13:52:55 +0200]:
>
> > On Tue 2013-06-25 22:00:39, Gustavo Padovan wrote:
> > > Hi Marcel,
> > >
> > > * Marcel Holtmann <[email protected]> [2013-06-25 13:49:19 -0700]:
> > >
> > > > Hi Pavel,
> > > >
> > > > >> Jaganath Kanakkassery (1):
> > > > >> Bluetooth: Fix invalid length check in l2cap_information_rsp()
> > > > >
> > > > > Could we add the "3.10 vs. thinkpad x60" patch that fixes regression?
> > > > > It is from Johan Hedberg…
> > > >
> > > > that one went into net-next tree. It is scheduled for 3.11 right now.
> > > >
> > > > After 3.10 is out, maybe it can go into -stable, but otherwise I think it is bit late now. Unless someone wants to cherry-pick it.
> > >
> > > That patch is actually in 3.10 already, it was included in the last pull
> > > request and John took it to wireless.git.
> >
> > Is it? I tried pulling yesterday (some updates) and today (no
> > changes), am on:
>
> Dave already pulled it so it should land in mainline anytime soon.
It seems to work in 3.10 and 3.11-rc1. Thanks!
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html