-- Problem Description --
The following upstream patches are needed for Ubuntu to fix a hang situation reported when executing ppc64_cpu --smt=on that occurs with various disk types. We need whichever ones have not yet been pulled into the base.
The hang problem can be reproduced with the following shell script using an NVMe device executed on kernel versions 4.4.0-45 and 4.4.0-59 within 30 minutes. It was also reproduced on a 4.8.0-32-generic kernel, although it took over 3 hours to manifest.
#!/bin/bash
if [[ ${#} -eq 0 ]]; then
${0} breaker &
while true; do
dd if=/dev/nvme0n1 bs=1024k of=/dev/null
done
elif [[ ${1} == "breaker" ]]; then
while true; do
ppc64_cpu --smt=off
sleep 5
ppc64_cpu --smt=on
sleep 5
done
fi
Steve,
Can Foundations take a look at this request, please.
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.
Correction: the hang reproduced by the previous shell script is actually being fixed separately. These commits fix various other problems with NVMe drives and are required as a prerequisite .
On Tue, Feb 07, 2017 at 12:42:39PM -0800, Michael Hohnbaum wrote:
> Can Foundations take a look at this request, please.
The bug is assigned to the linux package, so the kernel team should probably
be looking at it.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>
Leann,
While the problem is in ppc64-cpu command, it appears the fix is in a
set of kernel patches. Can you have the kernel team take a look at
these. Thanks.
Michael
On 02/07/2017 01:05 PM, Steve Langasek wrote:
> On Tue, Feb 07, 2017 at 12:42:39PM -0800, Michael Hohnbaum wrote:
>> Can Foundations take a look at this request, please.
> The bug is assigned to the linux package, so the kernel team should probably
> be looking at it.
>
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.
-- Problem Description --
The following upstream patches are needed for Ubuntu to fix a hang situation reported when executing ppc64_cpu --smt=on that occurs with various disk types. We need whichever ones have not yet been pulled into the base.
https:/ /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=e57690fe009 b2ab0cee8a57f53 be634540e49c9d /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=0e87e58bf60 edb6bb28e493c7a 143f41b091a5e5 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=135e8c9250d d5c8c9aae5984fd e6f230d0cbfeaf /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=c02ebfdddba fa9a6a0f52fbd71 5e6bfa229af9d3 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=d1b1cea1e58 477dad88ff769f5 4c0d2dfa56d923 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=36e1f3d1078 67b25c616c2fd29 4f5a1c9d4e5d09 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=71f79fb3179 e69b0c1448a2101 a866d871c66e7f
https:/
https:/
https:/
https:/
https:/
https:/
The hang problem can be reproduced with the following shell script using an NVMe device executed on kernel versions 4.4.0-45 and 4.4.0-59 within 30 minutes. It was also reproduced on a 4.8.0-32-generic kernel, although it took over 3 hours to manifest.
#!/bin/bash
if [[ ${#} -eq 0 ]]; then
${0} breaker &
while true; do
dd if=/dev/nvme0n1 bs=1024k of=/dev/null
done
elif [[ ${1} == "breaker" ]]; then
while true; do
ppc64_cpu --smt=off
sleep 5
ppc64_cpu --smt=on
sleep 5
done
fi
Steve,
Can Foundations take a look at this request, please.
Michael
On 02/07/2017 12:39 PM, Launchpad Bug Tracker wrote: /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=e57690fe009 b2ab0cee8a57f53 be634540e49c9d /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=0e87e58bf60 edb6bb28e493c7a 143f41b091a5e5 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=135e8c9250d d5c8c9aae5984fd e6f230d0cbfeaf /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=c02ebfdddba fa9a6a0f52fbd71 5e6bfa229af9d3 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=d1b1cea1e58 477dad88ff769f5 4c0d2dfa56d923 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=36e1f3d1078 67b25c616c2fd29 4f5a1c9d4e5d09 /git.kernel. org/cgit/ linux/kernel/ git/torvalds/ linux.git/ commit/ ?id=71f79fb3179 e69b0c1448a2101 a866d871c66e7f ppc64le bugnameltc-146759 severity-critical targetmilestone -inin16041
> bugproxy (bugproxy) has assigned this bug to you for Ubuntu:
>
> -- Problem Description --
> The following upstream patches are needed for Ubuntu to fix a hang situation reported when executing ppc64_cpu --smt=on that occurs with various disk types. We need whichever ones have not yet been pulled into the base.
>
> https:/
> https:/
> https:/
> https:/
> https:/
> https:/
> https:/
>
> The hang problem can be reproduced with the following shell script using
> an NVMe device executed on kernel versions 4.4.0-45 and 4.4.0-59 within
> 30 minutes. It was also reproduced on a 4.8.0-32-generic kernel,
> although it took over 3 hours to manifest.
>
> #!/bin/bash
>
> if [[ ${#} -eq 0 ]]; then
> ${0} breaker &
> while true; do
> dd if=/dev/nvme0n1 bs=1024k of=/dev/null
> done
> elif [[ ${1} == "breaker" ]]; then
> while true; do
> ppc64_cpu --smt=off
> sleep 5
> ppc64_cpu --smt=on
> sleep 5
> done
> fi
>
> ** Affects: ubuntu
> Importance: Undecided
> Assignee: Taco Screen team (taco-screen-team)
> Status: New
>
>
> ** Tags: architecture-
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.
Correction: the hang reproduced by the previous shell script is actually being fixed separately. These commits fix various other problems with NVMe drives and are required as a prerequisite .
On Tue, Feb 07, 2017 at 12:42:39PM -0800, Michael Hohnbaum wrote:
> Can Foundations take a look at this request, please.
The bug is assigned to the linux package, so the kernel team should probably
be looking at it.
-- www.debian. org/
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://
<email address hidden> <email address hidden>
Leann,
While the problem is in ppc64-cpu command, it appears the fix is in a
set of kernel patches. Can you have the kernel team take a look at
these. Thanks.
Michael
On 02/07/2017 01:05 PM, Steve Langasek wrote:
> On Tue, Feb 07, 2017 at 12:42:39PM -0800, Michael Hohnbaum wrote:
>> Can Foundations take a look at this request, please.
> The bug is assigned to the linux package, so the kernel team should probably
> be looking at it.
>
--
Michael Hohnbaum
OIL Program Manager
Power (ppc64el) Development Project Manager
Canonical, Ltd.
https:/ /lists. ubuntu. com/archives/ kernel- team/2017- February/ 082383. html /lists. ubuntu. com/archives/ kernel- team/2017- February/ 082386. html
https:/