Agreed, thanks. I'll also add the POWER9 support to the GCC side
of the patch now that the glibc code is upstream.
I do not see these bits being added in kernel side yet and GLIBC usual
only sync these kind of bits *after* they are included in kernel side.
So I would advise to either get these pieces (kernel support and hwcap
advertise) in kernel before 2.23 release, otherwise revert the patches.
Agreed, thanks. I'll also add the POWER9 support to the GCC side
of the patch now that the glibc code is upstream.
I do not see these bits being added in kernel side yet and GLIBC usual
only sync these kind of bits *after* they are included in kernel side.
So I would advise to either get these pieces (kernel support and hwcap
advertise) in kernel before 2.23 release, otherwise revert the patches.
Please revert the changes from glibc until you checkin support to linux
kernel mainline.
Leaving these bits in increases the risk that someone uses to deploy a glibc
that then may have the wrong value.
Cheers,
Carlos.
Agreed, thanks. I'll also add the POWER9 support to the GCC side
of the patch now that the glibc code is upstream.
I do not see these bits being added in kernel side yet and GLIBC usual
only sync these kind of bits *after* they are included in kernel side.
So I would advise to either get these pieces (kernel support and hwcap
advertise) in kernel before 2.23 release, otherwise revert the patches.
Please revert the changes from glibc until you checkin support to linux
kernel mainline.
Leaving these bits in increases the risk that someone uses to deploy a glibc
that then may have the wrong value.
Could you clarify this statement, please?
I fail to see how they could have the wrong value.
These bits are enabled if you're running a kernel on a processor that
supports ISA 3.0 and/or VSX IEEE Float 128-bit. The kernel must also support
this processor.
I see 3 possible scenarios:
- Processor that doesn't support these features: as these bits are 0 by
default, they have the correct value.
- Both processor and kernel support these features: both bits are set,
which is the correct value.
- Processor support these features, but an old kernel doesn't support: these
bits would be 0. But, that's a valid scenario.
Am I missing another scenario?
However, I do agree with the concerns raised by Peter and Adhemerval: glibc
should be in sync with the kernel by the time of the release in order to
guarantee both bits are reserved for the exact same goal and we should have
both AT_HWCAP and AT_PLATFORM supporting the new processor.
With that said, I was planning to revert both commits d2de9ef7 and b1f19b8e
if we don't get the kernel patch accepted into the powerpc tree in time for
the release 2.23.
--
Tulio Magno
Agreed, thanks. I'll also add the POWER9 support to the GCC side
of the patch now that the glibc code is upstream.
I do not see these bits being added in kernel side yet and GLIBC usual
only sync these kind of bits *after* they are included in kernel side.
So I would advise to either get these pieces (kernel support and hwcap
advertise) in kernel before 2.23 release, otherwise revert the patches.
Please revert the changes from glibc until you checkin support to linux
kernel mainline.
Leaving these bits in increases the risk that someone uses to deploy a glibc
that then may have the wrong value.
Could you clarify this statement, please?
I fail to see how they could have the wrong value.
Until it is checked into the mainline kernel it is not canonical.
That's the rule. There are no other discussions to be had.
The single rule avoids discussions like "it can never be wrong because that's
what our ABI says it is."
However, I do agree with the concerns raised by Peter and Adhemerval: glibc
should be in sync with the kernel by the time of the release in order to
guarantee both bits are reserved for the exact same goal and we should have
both AT_HWCAP and AT_PLATFORM supporting the new processor.
With that said, I was planning to revert both commits d2de9ef7 and b1f19b8e
if we don't get the kernel patch accepted into the powerpc tree in time for
the release 2.23.
Exactly. That's perfect. We can backport them to 2.23.1 if you get in later.
Cheers,
Carlos.
Agreed, thanks. I'll also add the POWER9 support to the GCC side
of the patch now that the glibc code is upstream.
I do not see these bits being added in kernel side yet and GLIBC usual
only sync these kind of bits *after* they are included in kernel side.
So I would advise to either get these pieces (kernel support and hwcap
advertise) in kernel before 2.23 release, otherwise revert the patches.
Please revert the changes from glibc until you checkin support to linux
kernel mainline.
Leaving these bits in increases the risk that someone uses to deploy a glibc
that then may have the wrong value.
Could you clarify this statement, please?
I fail to see how they could have the wrong value.
Until it is checked into the mainline kernel it is not canonical.
That's the rule. There are no other discussions to be had.
Well is was posted to to powerpc next:
https://git.kernel.org/powerpc/c/e708c24cd01ce80b1609d8bacc
We have agreement between the kernel and GLIBC (and the ABI).
The issue is just coordination across communities and individuals that
may not being paying attention to other communities dead lines.
Have you ever tried to push a string, up hill. That is open source
development in nutshell. ;)
So it is in flight and glibc is soft/slush freeze. I would hate to
revert this one day just to add it back to the next. Especially if those
days straddle the hard freeze ...
So can we let this ride a day or too?
The single rule avoids discussions like "it can never be wrong because that's
what our ABI says it is."
From: Carlos O'Donell <hidden> Date: 2016-01-12 17:45:40
On 01/12/2016 11:39 AM, Steven Munroe wrote:
quoted
That's the rule. There are no other discussions to be had.
Well is was posted to to powerpc next:
https://git.kernel.org/powerpc/c/e708c24cd01ce80b1609d8bacc
We have agreement between the kernel and GLIBC (and the ABI).
The issue is just coordination across communities and individuals that
may not being paying attention to other communities dead lines.
Have you ever tried to push a string, up hill. That is open source
development in nutshell. ;)
I know exactly what this is like.
So it is in flight and glibc is soft/slush freeze. I would hate to
revert this one day just to add it back to the next. Especially if those
days straddle the hard freeze ...
So can we let this ride a day or too?
Sure. I'm not an unreasonable person.
My goal as a glibc steward is to remind IBM that our best practice is that
we *wait* until it goes into mainline before committing to glibc master.
There really isn't any reason to check this in to glibc master right now.
It could wait.
Adhemerval as a release manager is also not an unreasonable person.
I have already discussed with Tulio that he should have just waited to
commit these changes, but gotten an exception from Adhemerval to checkin
the fairly low-risk patches late in the freeze. That's exactly the purpose
of a release managers job, to grant you exceptions as we approach release,
particularly when schedules don't quite line up.
Cheers,
Carlos.
From: Carlos Eduardo Seo <hidden> Date: 2016-01-15 22:30:58
On 1/11/16 5:55 PM, Tulio Magno Quites Machado Filho wrote:
However, I do agree with the concerns raised by Peter and Adhemerval: glibc
should be in sync with the kernel by the time of the release in order to
guarantee both bits are reserved for the exact same goal and we should have
both AT_HWCAP and AT_PLATFORM supporting the new processor.
With that said, I was planning to revert both commits d2de9ef7 and b1f19b8e
if we don't get the kernel patch accepted into the powerpc tree in time for
the release 2.23.
Kernel patch is in, commit f689b742f217b2ffe7925f8a6521b208ee995309
Regards,
--
Carlos Eduardo Seo
Software Engineer - Linux on Power Toolchain
cseo@linux.vnet.ibm.com