Feb. Fedora Kernel Patch Report

Feb. Fedora Kernel Patch Report

[Josh Boyer]

It's been a while since I sent one of these.  Mostly that's due to the
overlap between which upstream stable version we're using in Fedora
across the releases, and how fast those have been happening upstream.
We're settled on 3.13.y now, and with 3.14-rc4 out there things have
calmed down enough to take stock again.

Here's the patches we have on top of 3.13.5.

josh

fs-proc-devtree-remove_proc_entry.patch
- Upstream commit c1d867a54d426b45da017fbe8e585f8a3064ce8d

drm-radeon-Disable-writeback-by-default-on-ppc.patch
- This is superseded by upstream commit
ea31bf697d27270188a93cd78cf9de4bc968aca3 but that seems to be a bit
big for a stable backport

dm-cache-policy-mq_fix-large-scale-table-allocation-bug.patch (rhbz 993744)
- Still pending upstream (no idea why...)

sunrpc-create-a-new-dummy-pipe-for-gssd-to-hold-open.patch
sunrpc-replace-gssd_running-with-more-reliable-check.patch
nfs-check-gssd-running-before-krb5i-auth.patch
- Added to fix a 15sec mount delay
- Upstream commits 4b9a445e3eeb8bd9278b1ae51c1b3a651e370cd6
  89f842435c630f8426f414e6030bc2ffea0d6f81
  6aa23d76a7b549521a03b63b6d5b7880ea87eab7 respectively

rpc_pipe-remove-the-clntXX-dir-if-creating-the-pipe-fails.patch
sunrpc-add-an-info-file-for-the-dummy-gssd-pipe.patch
rpc_pipe-fix-cleanup-of-dummy-gssd-directory-when-notification-fails.patch
- Fixes for some of the patches above (rhbz 1037793)
- Upstream commits 3396f92f8be606ea485b0a82d4e7749a448b013b
  e2f0c83a9de331d9352185ca3642616c13127539
  23e66ba97127ff3b064d4c6c5138aa34eafc492f respectively

elantech-Properly-differentiate-between-clickpads-an.patch (rhbz 1030802)
- Upstream commit c15bdfd5b9831e4cab8cfc118243956e267dd30e

KVM-MMU-handle-invalid-root_hpa-at-__direct_map.patch (rhbz 924916)
- Upstream commit 989c6b34f6a9480e397b170cc62237e89bf4fdb9
- Probably should grab 37f6a4e237303549c8676dfe1fd1991ceab512eb too

KVM-VMX-fix-use-after-free-of-vmx-loaded_vmcs.patch (rhbz 1047892)
- Upstream commit 26a865f4aa8e66a6d94958de7656f7f1b03c6c56

0001-Input-wacom-make-sure-touch_max-is-set-for-touch-dev.patch
0002-Input-wacom-add-support-for-three-new-Intuos-devices.patch
0003-Input-wacom-add-reporting-of-SW_MUTE_DEVICE-events.patch
- rhbz 1003167 1046238
- Upstream commits 1d0d6df02750b4a6f466768cbfbf860e24f4c8d4
  b5fd2a3e92ca5c8c1f3c20d31ac5daed3ec4d604
  961794a00eab03f4344b7d5e825e8e789e55da87

Input-ALPS-add-support-for-Dolphin-devices.patch (rhbz 953211)
- Upstream commit ee65d4b36de8ddf4467f788faa5d8ddd1bfcdaa2

xhci-fix-resume-issues-on-renesas-chips-in-samsung-laptops.patch (rhbz 950630)
- Upstream commit 1aa9578c1a9450fb21501c4f549f5b1edb557e6d
- Oddly not queued for stable?

cgroup-fixes.patch (rhbz 1045755)
- Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
- There was a 3.12 version of this patch sent to stable list from
Michal Hocko.  I don't remember why it wasn't applied, or why it
wasn't applied for 3.13.y...

ipv6-introduce-IFA_F_NOPREFIXROUTE-and-IFA_F_MANAGETEMPADDR-flags.patch
ipv6-addrconf-revert-if_inet6ifa_flag-format.patch
- rhbz 1064430 1056711
- Backports of commits to add this stuff in 3.14.  Not suitable for stable.

cifs-ensure-that-uncached-writes-handle-unmapped-areas-correctly.patch
cifs-sanity-check-length-of-data-to-send-before-sending.patch
cifs-mask-off-top-byte-in-get_rfc1002_length.patch
- rhbz 1064253 CVE-2014-0069 rhbz 1068862
- Upstream commit 5d81de8e8667da7135d3a32a964087c0faf5483f and two
additional fixes

cpufreq-powernow-k8-Initialize-per-cpu-data-structures-properly.patch
(rhbz 1054408)
- Upstream commit c3274763bfc3bf1ececa269ed6e6c4d7ec1c3e5e
- CC'd to stable but not queued

e100-Fix-disabling-already-disabled-device-warning.patch (rhbz 994438)
- Upstream commit 2b6e0ca175fe4a20f21ba82b1e7ccc71029c4dd4

usb-ehci-fix-deadlock-when-threadirqs-option-is-used.patch
- Still pending upstream

Re: Feb. Fedora Kernel Patch Report

[Luís Henriques]

Hi Josh,

On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
> It's been a while since I sent one of these.  Mostly that's due to the
> overlap between which upstream stable version we're using in Fedora
> across the releases, and how fast those have been happening upstream.
> We're settled on 3.13.y now, and with 3.14-rc4 out there things have
> calmed down enough to take stock again.
> 
> Here's the patches we have on top of 3.13.5.
> 
> josh

Thanks a lot for sending this list, Josh!  I'm queuing a few of these for
the kernels I'm currently maintaining (3.5 and 3.11).

> fs-proc-devtree-remove_proc_entry.patch
> - Upstream commit c1d867a54d426b45da017fbe8e585f8a3064ce8d

I'm queuing this for the 3.5 and 3.11 kernels.

> drm-radeon-Disable-writeback-by-default-on-ppc.patch
> - This is superseded by upstream commit
> ea31bf697d27270188a93cd78cf9de4bc968aca3 but that seems to be a bit
> big for a stable backport
> 
> dm-cache-policy-mq_fix-large-scale-table-allocation-bug.patch (rhbz 993744)
> - Still pending upstream (no idea why...)
> 
> sunrpc-create-a-new-dummy-pipe-for-gssd-to-hold-open.patch
> sunrpc-replace-gssd_running-with-more-reliable-check.patch
> nfs-check-gssd-running-before-krb5i-auth.patch
> - Added to fix a 15sec mount delay
> - Upstream commits 4b9a445e3eeb8bd9278b1ae51c1b3a651e370cd6
>   89f842435c630f8426f414e6030bc2ffea0d6f81
>   6aa23d76a7b549521a03b63b6d5b7880ea87eab7 respectively
> 
> rpc_pipe-remove-the-clntXX-dir-if-creating-the-pipe-fails.patch
> sunrpc-add-an-info-file-for-the-dummy-gssd-pipe.patch
> rpc_pipe-fix-cleanup-of-dummy-gssd-directory-when-notification-fails.patch
> - Fixes for some of the patches above (rhbz 1037793)
> - Upstream commits 3396f92f8be606ea485b0a82d4e7749a448b013b
>   e2f0c83a9de331d9352185ca3642616c13127539
>   23e66ba97127ff3b064d4c6c5138aa34eafc492f respectively
> 
> elantech-Properly-differentiate-between-clickpads-an.patch (rhbz 1030802)
> - Upstream commit c15bdfd5b9831e4cab8cfc118243956e267dd30e

Again, queuing this one for 3.5 and 3.11.

> KVM-MMU-handle-invalid-root_hpa-at-__direct_map.patch (rhbz 924916)
> - Upstream commit 989c6b34f6a9480e397b170cc62237e89bf4fdb9
> - Probably should grab 37f6a4e237303549c8676dfe1fd1991ceab512eb too

Queuing only 989c6b34f6a9480e397b170cc62237e89bf4fdb9 for the 3.5 and
3.11.

> KVM-VMX-fix-use-after-free-of-vmx-loaded_vmcs.patch (rhbz 1047892)
> - Upstream commit 26a865f4aa8e66a6d94958de7656f7f1b03c6c56

3.5 and 3.11.

> 0001-Input-wacom-make-sure-touch_max-is-set-for-touch-dev.patch
> 0002-Input-wacom-add-support-for-three-new-Intuos-devices.patch
> 0003-Input-wacom-add-reporting-of-SW_MUTE_DEVICE-events.patch
> - rhbz 1003167 1046238
> - Upstream commits 1d0d6df02750b4a6f466768cbfbf860e24f4c8d4
>   b5fd2a3e92ca5c8c1f3c20d31ac5daed3ec4d604
>   961794a00eab03f4344b7d5e825e8e789e55da87

Queuing 1d0d6df02750b4a6f466768cbfbf860e24f4c8d4, only for the 3.11
kernel.

> Input-ALPS-add-support-for-Dolphin-devices.patch (rhbz 953211)
> - Upstream commit ee65d4b36de8ddf4467f788faa5d8ddd1bfcdaa2
> 
> xhci-fix-resume-issues-on-renesas-chips-in-samsung-laptops.patch (rhbz 950630)
> - Upstream commit 1aa9578c1a9450fb21501c4f549f5b1edb557e6d
> - Oddly not queued for stable?

Already on 3.5 and 3.11.

> cgroup-fixes.patch (rhbz 1045755)
> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
> - There was a 3.12 version of this patch sent to stable list from
> Michal Hocko.  I don't remember why it wasn't applied, or why it
> wasn't applied for 3.13.y...
>
> ipv6-introduce-IFA_F_NOPREFIXROUTE-and-IFA_F_MANAGETEMPADDR-flags.patch
> ipv6-addrconf-revert-if_inet6ifa_flag-format.patch
> - rhbz 1064430 1056711
> - Backports of commits to add this stuff in 3.14.  Not suitable for stable.
> 
> cifs-ensure-that-uncached-writes-handle-unmapped-areas-correctly.patch
> cifs-sanity-check-length-of-data-to-send-before-sending.patch
> cifs-mask-off-top-byte-in-get_rfc1002_length.patch
> - rhbz 1064253 CVE-2014-0069 rhbz 1068862
> - Upstream commit 5d81de8e8667da7135d3a32a964087c0faf5483f and two
> additional fixes

5d81de8e8667da7135d3a32a964087c0faf5483f is already queued for 3.5 and
3.11.

> cpufreq-powernow-k8-Initialize-per-cpu-data-structures-properly.patch
> (rhbz 1054408)
> - Upstream commit c3274763bfc3bf1ececa269ed6e6c4d7ec1c3e5e
> - CC'd to stable but not queued

Already in 3.5 and 3.11.

> e100-Fix-disabling-already-disabled-device-warning.patch (rhbz 994438)
> - Upstream commit 2b6e0ca175fe4a20f21ba82b1e7ccc71029c4dd4

Queuing for 3.5 and 3.11.

> usb-ehci-fix-deadlock-when-threadirqs-option-is-used.patch
> - Still pending upstream

Cheers,
--
Luís

Re: Feb. Fedora Kernel Patch Report

[Greg KH]

On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
> It's been a while since I sent one of these.  Mostly that's due to the
> overlap between which upstream stable version we're using in Fedora
> across the releases, and how fast those have been happening upstream.
> We're settled on 3.13.y now, and with 3.14-rc4 out there things have
> calmed down enough to take stock again.
> 
> Here's the patches we have on top of 3.13.5.
> 
> josh
> 
> fs-proc-devtree-remove_proc_entry.patch
> - Upstream commit c1d867a54d426b45da017fbe8e585f8a3064ce8d

Now added.

> 
> drm-radeon-Disable-writeback-by-default-on-ppc.patch
> - This is superseded by upstream commit
> ea31bf697d27270188a93cd78cf9de4bc968aca3 but that seems to be a bit
> big for a stable backport

Yes too big.

> 
> dm-cache-policy-mq_fix-large-scale-table-allocation-bug.patch (rhbz 993744)
> - Still pending upstream (no idea why...)
> 
> sunrpc-create-a-new-dummy-pipe-for-gssd-to-hold-open.patch
> sunrpc-replace-gssd_running-with-more-reliable-check.patch
> nfs-check-gssd-running-before-krb5i-auth.patch
> - Added to fix a 15sec mount delay
> - Upstream commits 4b9a445e3eeb8bd9278b1ae51c1b3a651e370cd6
>   89f842435c630f8426f414e6030bc2ffea0d6f81
>   6aa23d76a7b549521a03b63b6d5b7880ea87eab7 respectively
> 
> rpc_pipe-remove-the-clntXX-dir-if-creating-the-pipe-fails.patch
> sunrpc-add-an-info-file-for-the-dummy-gssd-pipe.patch
> rpc_pipe-fix-cleanup-of-dummy-gssd-directory-when-notification-fails.patch
> - Fixes for some of the patches above (rhbz 1037793)
> - Upstream commits 3396f92f8be606ea485b0a82d4e7749a448b013b
>   e2f0c83a9de331d9352185ca3642616c13127539
>   23e66ba97127ff3b064d4c6c5138aa34eafc492f respectively


This seems like a new feature, which I'll not add to -stable trees.

> 
> elantech-Properly-differentiate-between-clickpads-an.patch (rhbz 1030802)
> - Upstream commit c15bdfd5b9831e4cab8cfc118243956e267dd30e

added

> 
> KVM-MMU-handle-invalid-root_hpa-at-__direct_map.patch (rhbz 924916)
> - Upstream commit 989c6b34f6a9480e397b170cc62237e89bf4fdb9
> - Probably should grab 37f6a4e237303549c8676dfe1fd1991ceab512eb too

Nice ones, if only the kvm developers would mark things for stable...

> 
> KVM-VMX-fix-use-after-free-of-vmx-loaded_vmcs.patch (rhbz 1047892)
> - Upstream commit 26a865f4aa8e66a6d94958de7656f7f1b03c6c56

applied.

> 
> 0001-Input-wacom-make-sure-touch_max-is-set-for-touch-dev.patch
> 0002-Input-wacom-add-support-for-three-new-Intuos-devices.patch
> 0003-Input-wacom-add-reporting-of-SW_MUTE_DEVICE-events.patch
> - rhbz 1003167 1046238
> - Upstream commits 1d0d6df02750b4a6f466768cbfbf860e24f4c8d4
>   b5fd2a3e92ca5c8c1f3c20d31ac5daed3ec4d604
>   961794a00eab03f4344b7d5e825e8e789e55da87

Applied.

> Input-ALPS-add-support-for-Dolphin-devices.patch (rhbz 953211)
> - Upstream commit ee65d4b36de8ddf4467f788faa5d8ddd1bfcdaa2

Too big for stable :(

> xhci-fix-resume-issues-on-renesas-chips-in-samsung-laptops.patch (rhbz 950630)
> - Upstream commit 1aa9578c1a9450fb21501c4f549f5b1edb557e6d
> - Oddly not queued for stable?

Someone marked it incorrectly :(

> cgroup-fixes.patch (rhbz 1045755)
> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
> - There was a 3.12 version of this patch sent to stable list from
> Michal Hocko.  I don't remember why it wasn't applied, or why it
> wasn't applied for 3.13.y...

It didn't apply to 3.13, and I think I was waiting for a backported
version, do you have one?

> ipv6-introduce-IFA_F_NOPREFIXROUTE-and-IFA_F_MANAGETEMPADDR-flags.patch
> ipv6-addrconf-revert-if_inet6ifa_flag-format.patch
> - rhbz 1064430 1056711
> - Backports of commits to add this stuff in 3.14.  Not suitable for stable.
> 
> cifs-ensure-that-uncached-writes-handle-unmapped-areas-correctly.patch
> cifs-sanity-check-length-of-data-to-send-before-sending.patch
> cifs-mask-off-top-byte-in-get_rfc1002_length.patch
> - rhbz 1064253 CVE-2014-0069 rhbz 1068862
> - Upstream commit 5d81de8e8667da7135d3a32a964087c0faf5483f and two
> additional fixes

In the last release.

> 
> cpufreq-powernow-k8-Initialize-per-cpu-data-structures-properly.patch
> (rhbz 1054408)
> - Upstream commit c3274763bfc3bf1ececa269ed6e6c4d7ec1c3e5e
> - CC'd to stable but not queued

Had to wait a release.

> e100-Fix-disabling-already-disabled-device-warning.patch (rhbz 994438)
> - Upstream commit 2b6e0ca175fe4a20f21ba82b1e7ccc71029c4dd4

Now added.

Thanks so much for this breakdown, much appreciated.

greg k-h

Re: Feb. Fedora Kernel Patch Report

[Josh Boyer]

On Tue, Mar 25, 2014 at 5:24 PM, Greg KH <greg@kroah.com> wrote:
> On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
>> cgroup-fixes.patch (rhbz 1045755)
>> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
>> - There was a 3.12 version of this patch sent to stable list from
>> Michal Hocko.  I don't remember why it wasn't applied, or why it
>> wasn't applied for 3.13.y...
>
> It didn't apply to 3.13, and I think I was waiting for a backported
> version, do you have one?

As far as I know we're using this version:

http://thread.gmane.org/gmane.linux.kernel.cgroups/10336/focus=1646465

The commentary on the thread seems to imply that Michal was going to
send it again for some reason.  Michal, did you send the 3.12/3.13
backport for this commit directly to stable?

josh

Re: Feb. Fedora Kernel Patch Report

[Jiri Slaby]

On 03/26/2014 02:44 PM, Josh Boyer wrote:
> On Tue, Mar 25, 2014 at 5:24 PM, Greg KH <greg@kroah.com> wrote:
>> On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
>>> cgroup-fixes.patch (rhbz 1045755)
>>> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
>>> - There was a 3.12 version of this patch sent to stable list from
>>> Michal Hocko.  I don't remember why it wasn't applied, or why it
>>> wasn't applied for 3.13.y...
>>
>> It didn't apply to 3.13, and I think I was waiting for a backported
>> version, do you have one?
> 
> As far as I know we're using this version:
> 
> http://thread.gmane.org/gmane.linux.kernel.cgroups/10336/focus=1646465
> 
> The commentary on the thread seems to imply that Michal was going to
> send it again for some reason.  Michal, did you send the 3.12/3.13
> backport for this commit directly to stable?

In the meantime, I applied to the stable 3.12 Michal's version from suse
enterprise 12. I believe that it is the same as in the link above.
Please confirm.

thanks,
-- 
js
suse labs

Re: Feb. Fedora Kernel Patch Report

[Michal Hocko]

On Wed 26-03-14 15:51:35, Jiri Slaby wrote:
> On 03/26/2014 02:44 PM, Josh Boyer wrote:
> > On Tue, Mar 25, 2014 at 5:24 PM, Greg KH <greg@kroah.com> wrote:
> >> On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
> >>> cgroup-fixes.patch (rhbz 1045755)
> >>> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
> >>> - There was a 3.12 version of this patch sent to stable list from
> >>> Michal Hocko.  I don't remember why it wasn't applied, or why it
> >>> wasn't applied for 3.13.y...
> >>
> >> It didn't apply to 3.13, and I think I was waiting for a backported
> >> version, do you have one?
> > 
> > As far as I know we're using this version:
> > 
> > http://thread.gmane.org/gmane.linux.kernel.cgroups/10336/focus=1646465
> > 
> > The commentary on the thread seems to imply that Michal was going to
> > send it again for some reason.  Michal, did you send the 3.12/3.13
> > backport for this commit directly to stable?
> 
> In the meantime, I applied to the stable 3.12 Michal's version from suse
> enterprise 12. I believe that it is the same as in the link above.
> Please confirm.

Do not remember why it didn't end up in the stable back then but yes the
fix you have queued up seems to be the right one.

Thanks
-- 
Michal Hocko
SUSE Labs

Re: Feb. Fedora Kernel Patch Report

[Greg KH]

On Wed, Mar 26, 2014 at 02:16:38PM -0700, Michal Hocko wrote:
> On Wed 26-03-14 15:51:35, Jiri Slaby wrote:
> > On 03/26/2014 02:44 PM, Josh Boyer wrote:
> > > On Tue, Mar 25, 2014 at 5:24 PM, Greg KH <greg@kroah.com> wrote:
> > >> On Fri, Feb 28, 2014 at 10:13:29AM -0500, Josh Boyer wrote:
> > >>> cgroup-fixes.patch (rhbz 1045755)
> > >>> - Upstream commits 0ab02ca8f887908152d1a96db5130fc661d36a1e
> > >>> - There was a 3.12 version of this patch sent to stable list from
> > >>> Michal Hocko.  I don't remember why it wasn't applied, or why it
> > >>> wasn't applied for 3.13.y...
> > >>
> > >> It didn't apply to 3.13, and I think I was waiting for a backported
> > >> version, do you have one?
> > > 
> > > As far as I know we're using this version:
> > > 
> > > http://thread.gmane.org/gmane.linux.kernel.cgroups/10336/focus=1646465
> > > 
> > > The commentary on the thread seems to imply that Michal was going to
> > > send it again for some reason.  Michal, did you send the 3.12/3.13
> > > backport for this commit directly to stable?
> > 
> > In the meantime, I applied to the stable 3.12 Michal's version from suse
> > enterprise 12. I believe that it is the same as in the link above.
> > Please confirm.
> 
> Do not remember why it didn't end up in the stable back then but yes the
> fix you have queued up seems to be the right one.

Thanks, I've picked that one up now.

greg k-h