Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-08 Thread Jesper Nilsson 
On Wed, Mar 07, 2018 at 03:09:27PM +0100, Arnd Bergmann wrote:
> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>  wrote:
> > On (03/07/18 13:48), Arnd Bergmann wrote:
> > [..]
> >> >
> >> > I can easily stop building blackfin - and, if the intention is to
> >> > remove it, there is not much point in wasting resources building it any
> >> > anyway.
> >>
> >> Right. At the moment, the plan is to remove metag, score, unicore32,
> >> m32r, frv and blackfin. If you are building any more of those, you can
> >> stop that as well.
> >>
> >> The fate of tile and mn10300 is still open, but I suspect they won't
> >> last long either.
> >
> > That's a huge list.
> >
> > I heard that CRIS is having some problems as well:
> > https://lkml.org/lkml/2018/1/11/403
> 
> Interesting. I thought I looked at cris when I made the list and found that
> it was still (semi-)regularly updated, so I assumed it was still needed.

Yeah, I've tried to keep it chugging along, but as I mentioned
in that thread, time and working hardware are running low.

I've not heard of anyone using real CRIS-hardware with mainline
releases for a couple of years now (except me :-)

> Jesper/Mikael: If you are considering to drop the port, I think it
> would make sense to do it at the same time as the others, so feel
> free to send me a removal pull request for 4.17 that I'll merge with
> the other removal patches, or I can offer to remove it myself.

I'm having mixed feelings, but I think it's time for Linux to
drop the CRIS-port. I'll prepare some patches for you in the CRIS-tree.

> Obviously if your plans have changed in the meantime and you
> prefer to keep it for a while longer, just let me know and I'll document
> the state in the tag description for the arch removal for future reference.
> 
> If we remove cris, it's probably also appropriate to ask the question
> about microblaze: Michal, I know your work on microblaze linux has
> been winding down over the years as the user base shrinks (just
> like for all others above). Should we keep it for a few more years for
> the remaining users, or is it time to let go? I would guess we still
> have active users on modern kernels, but that's what I thought
> about cris as well ;-)
> 
>   Arnd

/^JN - Jesper Nilsson
-- 
   Jesper Nilsson -- jesper.nils...@axis.com

Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-08 Thread Jesper Nilsson 
On Wed, Mar 07, 2018 at 03:09:27PM +0100, Arnd Bergmann wrote:
> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>  wrote:
> > On (03/07/18 13:48), Arnd Bergmann wrote:
> > [..]
> >> >
> >> > I can easily stop building blackfin - and, if the intention is to
> >> > remove it, there is not much point in wasting resources building it any
> >> > anyway.
> >>
> >> Right. At the moment, the plan is to remove metag, score, unicore32,
> >> m32r, frv and blackfin. If you are building any more of those, you can
> >> stop that as well.
> >>
> >> The fate of tile and mn10300 is still open, but I suspect they won't
> >> last long either.
> >
> > That's a huge list.
> >
> > I heard that CRIS is having some problems as well:
> > https://lkml.org/lkml/2018/1/11/403
> 
> Interesting. I thought I looked at cris when I made the list and found that
> it was still (semi-)regularly updated, so I assumed it was still needed.

Yeah, I've tried to keep it chugging along, but as I mentioned
in that thread, time and working hardware are running low.

I've not heard of anyone using real CRIS-hardware with mainline
releases for a couple of years now (except me :-)

> Jesper/Mikael: If you are considering to drop the port, I think it
> would make sense to do it at the same time as the others, so feel
> free to send me a removal pull request for 4.17 that I'll merge with
> the other removal patches, or I can offer to remove it myself.

I'm having mixed feelings, but I think it's time for Linux to
drop the CRIS-port. I'll prepare some patches for you in the CRIS-tree.

> Obviously if your plans have changed in the meantime and you
> prefer to keep it for a while longer, just let me know and I'll document
> the state in the tag description for the arch removal for future reference.
> 
> If we remove cris, it's probably also appropriate to ask the question
> about microblaze: Michal, I know your work on microblaze linux has
> been winding down over the years as the user base shrinks (just
> like for all others above). Should we keep it for a few more years for
> the remaining users, or is it time to let go? I would guess we still
> have active users on modern kernels, but that's what I thought
> about cris as well ;-)
> 
>   Arnd

/^JN - Jesper Nilsson
-- 
   Jesper Nilsson -- jesper.nils...@axis.com

Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Arnd Bergmann 
On Wed, Mar 7, 2018 at 4:01 PM, Michal Simek  wrote:
> On 7.3.2018 15:09, Arnd Bergmann wrote:
>> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>>  wrote:

>> If we remove cris, it's probably also appropriate to ask the question
>> about microblaze: Michal, I know your work on microblaze linux has
>> been winding down over the years as the user base shrinks (just
>> like for all others above). Should we keep it for a few more years for
>> the remaining users, or is it time to let go? I would guess we still
>> have active users on modern kernels, but that's what I thought
>> about cris as well ;-)
>
> It is not like that. I definitely have less time for Microblaze to
> develop new stuff but Microblaze HW is still regularly updated and new
> hw features are coming there and new versions are created. We are also
> updating and testing newer kernels and using latest kernel at Xilinx. It
> is definitely not in state that we should just remove it.
>
> There are still a lot of users which are happy with current state and
> there is no pressure to add new features like SMP/address extension/etc
> that's why there is no huge development in arch/microblaze.

Ok, glad to hear that, thanks a lot for the clarification.

   Arnd

Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Arnd Bergmann 
On Wed, Mar 7, 2018 at 4:01 PM, Michal Simek  wrote:
> On 7.3.2018 15:09, Arnd Bergmann wrote:
>> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>>  wrote:

>> If we remove cris, it's probably also appropriate to ask the question
>> about microblaze: Michal, I know your work on microblaze linux has
>> been winding down over the years as the user base shrinks (just
>> like for all others above). Should we keep it for a few more years for
>> the remaining users, or is it time to let go? I would guess we still
>> have active users on modern kernels, but that's what I thought
>> about cris as well ;-)
>
> It is not like that. I definitely have less time for Microblaze to
> develop new stuff but Microblaze HW is still regularly updated and new
> hw features are coming there and new versions are created. We are also
> updating and testing newer kernels and using latest kernel at Xilinx. It
> is definitely not in state that we should just remove it.
>
> There are still a lot of users which are happy with current state and
> there is no pressure to add new features like SMP/address extension/etc
> that's why there is no huge development in arch/microblaze.

Ok, glad to hear that, thanks a lot for the clarification.

   Arnd

Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Michal Simek 
Hi Arnd,

On 7.3.2018 15:09, Arnd Bergmann wrote:
> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>  wrote:
>> On (03/07/18 13:48), Arnd Bergmann wrote:
>> [..]

 I can easily stop building blackfin - and, if the intention is to
 remove it, there is not much point in wasting resources building it any
 anyway.
>>>
>>> Right. At the moment, the plan is to remove metag, score, unicore32,
>>> m32r, frv and blackfin. If you are building any more of those, you can
>>> stop that as well.
>>>
>>> The fate of tile and mn10300 is still open, but I suspect they won't
>>> last long either.
>>
>> That's a huge list.
>>
>> I heard that CRIS is having some problems as well:
>> https://lkml.org/lkml/2018/1/11/403
> 
> Interesting. I thought I looked at cris when I made the list and found that
> it was still (semi-)regularly updated, so I assumed it was still needed.
> 
> (adding CRIS and microblaze maintainers)
> 
> Jesper/Mikael: If you are considering to drop the port, I think it
> would make sense to do it at the same time as the others, so feel
> free to send me a removal pull request for 4.17 that I'll merge with
> the other removal patches, or I can offer to remove it myself.
> 
> Obviously if your plans have changed in the meantime and you
> prefer to keep it for a while longer, just let me know and I'll document
> the state in the tag description for the arch removal for future reference.
> 
> If we remove cris, it's probably also appropriate to ask the question
> about microblaze: Michal, I know your work on microblaze linux has
> been winding down over the years as the user base shrinks (just
> like for all others above). Should we keep it for a few more years for
> the remaining users, or is it time to let go? I would guess we still
> have active users on modern kernels, but that's what I thought
> about cris as well ;-)

It is not like that. I definitely have less time for Microblaze to
develop new stuff but Microblaze HW is still regularly updated and new
hw features are coming there and new versions are created. We are also
updating and testing newer kernels and using latest kernel at Xilinx. It
is definitely not in state that we should just remove it.

There are still a lot of users which are happy with current state and
there is no pressure to add new features like SMP/address extension/etc
that's why there is no huge development in arch/microblaze.

Thanks,
Michal

-- 
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Xilinx Microblaze
Maintainer of Linux kernel - Xilinx Zynq ARM and ZynqMP ARM64 SoCs
U-Boot custodian - Xilinx Microblaze/Zynq/ZynqMP SoCs




signature.asc
Description: OpenPGP digital signature

Re: CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Michal Simek 
Hi Arnd,

On 7.3.2018 15:09, Arnd Bergmann wrote:
> On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
>  wrote:
>> On (03/07/18 13:48), Arnd Bergmann wrote:
>> [..]

 I can easily stop building blackfin - and, if the intention is to
 remove it, there is not much point in wasting resources building it any
 anyway.
>>>
>>> Right. At the moment, the plan is to remove metag, score, unicore32,
>>> m32r, frv and blackfin. If you are building any more of those, you can
>>> stop that as well.
>>>
>>> The fate of tile and mn10300 is still open, but I suspect they won't
>>> last long either.
>>
>> That's a huge list.
>>
>> I heard that CRIS is having some problems as well:
>> https://lkml.org/lkml/2018/1/11/403
> 
> Interesting. I thought I looked at cris when I made the list and found that
> it was still (semi-)regularly updated, so I assumed it was still needed.
> 
> (adding CRIS and microblaze maintainers)
> 
> Jesper/Mikael: If you are considering to drop the port, I think it
> would make sense to do it at the same time as the others, so feel
> free to send me a removal pull request for 4.17 that I'll merge with
> the other removal patches, or I can offer to remove it myself.
> 
> Obviously if your plans have changed in the meantime and you
> prefer to keep it for a while longer, just let me know and I'll document
> the state in the tag description for the arch removal for future reference.
> 
> If we remove cris, it's probably also appropriate to ask the question
> about microblaze: Michal, I know your work on microblaze linux has
> been winding down over the years as the user base shrinks (just
> like for all others above). Should we keep it for a few more years for
> the remaining users, or is it time to let go? I would guess we still
> have active users on modern kernels, but that's what I thought
> about cris as well ;-)

It is not like that. I definitely have less time for Microblaze to
develop new stuff but Microblaze HW is still regularly updated and new
hw features are coming there and new versions are created. We are also
updating and testing newer kernels and using latest kernel at Xilinx. It
is definitely not in state that we should just remove it.

There are still a lot of users which are happy with current state and
there is no pressure to add new features like SMP/address extension/etc
that's why there is no huge development in arch/microblaze.

Thanks,
Michal

-- 
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Xilinx Microblaze
Maintainer of Linux kernel - Xilinx Zynq ARM and ZynqMP ARM64 SoCs
U-Boot custodian - Xilinx Microblaze/Zynq/ZynqMP SoCs




signature.asc
Description: OpenPGP digital signature

CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Arnd Bergmann 
On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
 wrote:
> On (03/07/18 13:48), Arnd Bergmann wrote:
> [..]
>> >
>> > I can easily stop building blackfin - and, if the intention is to
>> > remove it, there is not much point in wasting resources building it any
>> > anyway.
>>
>> Right. At the moment, the plan is to remove metag, score, unicore32,
>> m32r, frv and blackfin. If you are building any more of those, you can
>> stop that as well.
>>
>> The fate of tile and mn10300 is still open, but I suspect they won't
>> last long either.
>
> That's a huge list.
>
> I heard that CRIS is having some problems as well:
> https://lkml.org/lkml/2018/1/11/403

Interesting. I thought I looked at cris when I made the list and found that
it was still (semi-)regularly updated, so I assumed it was still needed.

(adding CRIS and microblaze maintainers)

Jesper/Mikael: If you are considering to drop the port, I think it
would make sense to do it at the same time as the others, so feel
free to send me a removal pull request for 4.17 that I'll merge with
the other removal patches, or I can offer to remove it myself.

Obviously if your plans have changed in the meantime and you
prefer to keep it for a while longer, just let me know and I'll document
the state in the tag description for the arch removal for future reference.

If we remove cris, it's probably also appropriate to ask the question
about microblaze: Michal, I know your work on microblaze linux has
been winding down over the years as the user base shrinks (just
like for all others above). Should we keep it for a few more years for
the remaining users, or is it time to let go? I would guess we still
have active users on modern kernels, but that's what I thought
about cris as well ;-)

  Arnd

CRIS port status, was: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

2018-03-07 Thread Arnd Bergmann 
On Wed, Mar 7, 2018 at 2:40 PM, Sergey Senozhatsky
 wrote:
> On (03/07/18 13:48), Arnd Bergmann wrote:
> [..]
>> >
>> > I can easily stop building blackfin - and, if the intention is to
>> > remove it, there is not much point in wasting resources building it any
>> > anyway.
>>
>> Right. At the moment, the plan is to remove metag, score, unicore32,
>> m32r, frv and blackfin. If you are building any more of those, you can
>> stop that as well.
>>
>> The fate of tile and mn10300 is still open, but I suspect they won't
>> last long either.
>
> That's a huge list.
>
> I heard that CRIS is having some problems as well:
> https://lkml.org/lkml/2018/1/11/403

Interesting. I thought I looked at cris when I made the list and found that
it was still (semi-)regularly updated, so I assumed it was still needed.

(adding CRIS and microblaze maintainers)

Jesper/Mikael: If you are considering to drop the port, I think it
would make sense to do it at the same time as the others, so feel
free to send me a removal pull request for 4.17 that I'll merge with
the other removal patches, or I can offer to remove it myself.

Obviously if your plans have changed in the meantime and you
prefer to keep it for a while longer, just let me know and I'll document
the state in the tag description for the arch removal for future reference.

If we remove cris, it's probably also appropriate to ask the question
about microblaze: Michal, I know your work on microblaze linux has
been winding down over the years as the user base shrinks (just
like for all others above). Should we keep it for a few more years for
the remaining users, or is it time to let go? I would guess we still
have active users on modern kernels, but that's what I thought
about cris as well ;-)

  Arnd