All of lore.kernel.org
 help / color / mirror / Atom feed
* [OpenRISC] Linux kernel upstreaming status
@ 2017-03-19 22:32 Stafford Horne
  2017-03-20  7:53 ` Jakob Viketoft
  0 siblings, 1 reply; 3+ messages in thread
From: Stafford Horne @ 2017-03-19 22:32 UTC (permalink / raw)
  To: openrisc

Hi All,

This weekend Linus merged a few minor compiler warning fixes I was working
on for OpenRISC. This should help spot more serious issues going forward.

That should be it for version 4.11, which should be released in a few
weeks, and saw the following features merged from OpenRISC backlogs.

 * Optimized memset and memcpy routines from Olof and Me (~20% boot time
   reduction)
 * Support for CPU idling from Sebastian Macke (improved power usage,
   unmeasured)
 * Also, various native atomic primatives support in prep for SMP

## Moving On

Going forward to 4.12 and beyond we have 3 types of patches pending:

 * SMP support - SMP support from Stefan
 * Drivers - char-scet, usb-ohs900, gpio-jbtrivial, ps2-oc, spi-oc-simple,
             dma-wbstreamer
 * Configs (device tree and kernel configs for board support)

### SMP

For SMP there is a bit of work to get these patches ready for upstream.

 * Documentation of smp (IPI) in the openrisc spec
 * Consider fixing the timer hacks noted in code by Stefan
 * Convert spinlocks/rwlocks to qspinlocks and rwlocks suggested by Peter Z
 * Getting Stefan's example SMP boards into FuseSoc

### Drivers

If there are any drivers you would like to see pushed upstream let me know.
Or even better I can help you get them upstream.  At the moment, I can only
test the GPIO and SPI drivers.  I will need to buy a new board or wire
things to my board get the others working.  Some points:

 * I think USB is important for many users, if someone want to volunteer to
   working upstreaming that and has hardware it will be great.  If not I
   will get to it later.
 * What is the scet driver?

### Configs

There has been some discussion on IRC before about where the dts configs
should live.  Should they be in the kernel source or in FuseSOC?

I think we should have some basic configs in the kernel for the major
boards and configurations (i.e. de0 nano, sockit and atlys boards).  But
definitely not for every board config out there.  Having them in the kernel
enables them to stay up-to-date and allows other users to know what we
support.

## Branches

I have split all the remaining changes into branches on my github repo, you
can find them below, they are all based on v4.11-rc2:

  https://github.com/stffrdhrn/linux

  openrisc-4.12-smp
  openrisc-4.12-config
  (drivers)
    openrisc-4.12-char-scet
    openrisc-4.12-dma-wbstreamer
    openrisc-4.12-gpio-jbtrivial
    openrisc-4.12-ps2-oc
    openrisc-4.12-spi-oc-simple
    openrisc-4.12-usb-ohs900

-Stafford

^ permalink raw reply	[flat|nested] 3+ messages in thread

* [OpenRISC] Linux kernel upstreaming status
  2017-03-19 22:32 [OpenRISC] Linux kernel upstreaming status Stafford Horne
@ 2017-03-20  7:53 ` Jakob Viketoft
  2017-03-20 14:12   ` Stafford Horne
  0 siblings, 1 reply; 3+ messages in thread
From: Jakob Viketoft @ 2017-03-20  7:53 UTC (permalink / raw)
  To: openrisc

Hello,

________________________________________
From: OpenRISC [openrisc-bounces at lists.librecores.org] on behalf of Stafford Horne [shorne at gmail.com]
Sent: Sunday, March 19, 2017 23:32
To: Openrisc
Subject: [OpenRISC] Linux kernel upstreaming status

>  * What is the scet driver?

I might be able to shed some light on this at least. It's the SpaceCraft Elapsed Timer, essentially a dead reckoning counter from the moment of power-on. This can be adjusted to match the current clock and so provide a system time, although it only got a resolution of 15.3 us since it's using subseconds (1/2^16th of a second) and not normal milli/microseconds or something similar. As for its inclusion in the Linux tree, I don't know if anyone else uses this. We have switched to RTEMS for all future work and bare-metal for all the previous ones and aren't really using any Linux parts anymore. The SCET implementation has also changed and that driver doesn't match our hardware at all anymore.

What I'm saying is that if there's no one else interested in this, I'd suggest you scrap it.

Best regards,

       /Jakob

Jakob Viketoft
Senior Engineer in RTL and embedded software

ÅAC Microtec AB
Dag Hammarskjölds väg 48
SE-751 83 Uppsala, Sweden

T: +46 702 80 95 97
http://www.aacmicrotec.com



^ permalink raw reply	[flat|nested] 3+ messages in thread

* [OpenRISC] Linux kernel upstreaming status
  2017-03-20  7:53 ` Jakob Viketoft
@ 2017-03-20 14:12   ` Stafford Horne
  0 siblings, 0 replies; 3+ messages in thread
From: Stafford Horne @ 2017-03-20 14:12 UTC (permalink / raw)
  To: openrisc

On Mon, Mar 20, 2017 at 07:53:36AM +0000, Jakob Viketoft wrote:
> Hello,
> 
> ________________________________________
> From: OpenRISC [openrisc-bounces at lists.librecores.org] on behalf of Stafford Horne [shorne at gmail.com]
> Sent: Sunday, March 19, 2017 23:32
> To: Openrisc
> Subject: [OpenRISC] Linux kernel upstreaming status
> 
> >  * What is the scet driver?
> 
> I might be able to shed some light on this at least. It's the SpaceCraft Elapsed Timer, essentially a dead reckoning counter from the moment of power-on. This can be adjusted to match the current clock and so provide a system time, although it only got a resolution of 15.3 us since it's using subseconds (1/2^16th of a second) and not normal milli/microseconds or something similar. As for its inclusion in the Linux tree, I don't know if anyone else uses this. We have switched to RTEMS for all future work and bare-metal for all the previous ones and aren't really using any Linux parts anymore. The SCET implementation has also changed and that driver doesn't match our hardware at all anymore.
> 
> What I'm saying is that if there's no one else interested in this, I'd suggest you scrap it.

Thanks Jakob,

That's very informative.

-Stafford

> Best regards,
> 
>        /Jakob
> 
> Jakob Viketoft
> Senior Engineer in RTL and embedded software
> 
> ÅAC Microtec AB
> Dag Hammarskjölds väg 48
> SE-751 83 Uppsala, Sweden
> 
> T: +46 702 80 95 97
> http://www.aacmicrotec.com
> 
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2017-03-20 14:12 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-03-19 22:32 [OpenRISC] Linux kernel upstreaming status Stafford Horne
2017-03-20  7:53 ` Jakob Viketoft
2017-03-20 14:12   ` Stafford Horne

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.