John,
I have two more fixes that I think would be worthwhile for 3.10 but
admittedly the scenario is somewhat unlikely, so if you want to hold
them for 3.11 I can live with that. In that case, I can put them into my
-next tree, or you can pull this into -next (but if you don't pull
wireless.git first you'd get some more fixes you already have.)
These two patches fix two issues with using rfkill randomly during
traffic, which would then cause our driver to stop working and not be
able to recover at all.
johannes
The following changes since commit 622ebe994f6866b8d46ee5d3bcc329ed65d3722d:
iwlwifi: fix rate control regression (2013-06-12 14:13:39 +0200)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-fixes.git for-john
for you to fetch changes up to 8a487b1a7432b20ff3f82387a8ce7555a964b44e:
iwlwifi: pcie: wake the queue if stopped when being unmapped (2013-06-13 16:44:04 +0200)
----------------------------------------------------------------
Emmanuel Grumbach (2):
iwlwifi: pcie: fix race in queue unmapping
iwlwifi: pcie: wake the queue if stopped when being unmapped
drivers/net/wireless/iwlwifi/pcie/tx.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
On Fri, Jun 14, 2013 at 12:42:47PM +0200, Johannes Berg wrote:
> John,
>
> I have two more fixes that I think would be worthwhile for 3.10 but
> admittedly the scenario is somewhat unlikely, so if you want to hold
> them for 3.11 I can live with that. In that case, I can put them into my
> -next tree, or you can pull this into -next (but if you don't pull
> wireless.git first you'd get some more fixes you already have.)
>
> These two patches fix two issues with using rfkill randomly during
> traffic, which would then cause our driver to stop working and not be
> able to recover at all.
>
> johannes
>
>
> The following changes since commit 622ebe994f6866b8d46ee5d3bcc329ed65d3722d:
>
> iwlwifi: fix rate control regression (2013-06-12 14:13:39 +0200)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-fixes.git for-john
>
> for you to fetch changes up to 8a487b1a7432b20ff3f82387a8ce7555a964b44e:
>
> iwlwifi: pcie: wake the queue if stopped when being unmapped (2013-06-13 16:44:04 +0200)
>
> ----------------------------------------------------------------
> Emmanuel Grumbach (2):
> iwlwifi: pcie: fix race in queue unmapping
> iwlwifi: pcie: wake the queue if stopped when being unmapped
>
> drivers/net/wireless/iwlwifi/pcie/tx.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
It isn't obvious to me that these need to be included in 3.10. I can
pull this into wireless-next once the current stuff in wireless makes
it to davem's tree.
If you decide that these really need to go, please provide some more
information on how likely users are to hit the bugs, or how serious
the effects of that would be.
John
--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.
>> I have two more fixes that I think would be worthwhile for 3.10 but
>> admittedly the scenario is somewhat unlikely, so if you want to hold
>> them for 3.11 I can live with that. In that case, I can put them into my
>> -next tree, or you can pull this into -next (but if you don't pull
>> wireless.git first you'd get some more fixes you already have.)
>>
>> These two patches fix two issues with using rfkill randomly during
>> traffic, which would then cause our driver to stop working and not be
>> able to recover at all.
>>
>> johannes
>>
>>
>
> It isn't obvious to me that these need to be included in 3.10. I can
> pull this into wireless-next once the current stuff in wireless makes
> it to davem's tree.
>
> If you decide that these really need to go, please provide some more
> information on how likely users are to hit the bugs, or how serious
> the effects of that would be.
>
John, it is really your call. I understand these concerns in -rc5.
For the first patch, I am not 100% sure that it fixes anything else
that a print. It is obvious to me that without this patch, you might
get an error in the log (I reproduced it a few times). I tend to say
that the print is harmless though.
The second patch fixes a real issue. The issue is hard to reproduce
though. You need to get in a case where the queues are full (sw queues
are stopped) and at that precise time, you need to free the Tx queues
(RFkill - because any other flow would flush the queues first I
think). If you got into that situation, then the queues would remain
stopped after you switch the RFkill button again. So yes... Unlikely.
I would still recommend distros (Stanislaw is CCed) to take both
patches since they are really safe and can avoid issues.
Now you have all the data :-)
On Fri, Jun 14, 2013 at 12:42:47PM +0200, Johannes Berg wrote:
> John,
>
> I have two more fixes that I think would be worthwhile for 3.10 but
> admittedly the scenario is somewhat unlikely, so if you want to hold
> them for 3.11 I can live with that. In that case, I can put them into my
> -next tree, or you can pull this into -next (but if you don't pull
> wireless.git first you'd get some more fixes you already have.)
>
> These two patches fix two issues with using rfkill randomly during
> traffic, which would then cause our driver to stop working and not be
> able to recover at all.
>
> johannes
>
>
> The following changes since commit 622ebe994f6866b8d46ee5d3bcc329ed65d3722d:
>
> iwlwifi: fix rate control regression (2013-06-12 14:13:39 +0200)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-fixes.git for-john
>
> for you to fetch changes up to 8a487b1a7432b20ff3f82387a8ce7555a964b44e:
>
> iwlwifi: pcie: wake the queue if stopped when being unmapped (2013-06-13 16:44:04 +0200)
Pulling now (to wireless-next)...
--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.
On Fri, 2013-06-14 at 13:34 -0400, John W. Linville wrote:
> It isn't obvious to me that these need to be included in 3.10. I can
> pull this into wireless-next once the current stuff in wireless makes
> it to davem's tree.
Please do, unless you changed your mind based on Emmanuel's reply.
johannes