SATA FSL and upstreaming

[Benjamin Herrenschmidt]

On Thu, 2013-05-16 at 06:24 +0000, Xie Shaohui-B21989 wrote:
> Hi, Ben,
> 
> Since the p5020ds you tested is a rev1 chip, I think the most possibility that SATA not work
> is due to a SATA erratum, which is fixed in rev1.1, we have a policy that patches for errata that 
> are present only on early silicon revisions, typically only rev1 silicon will not send upstream.
> You can find the patch at below link:
> http://git.freescale.com/git/cgit.cgi/ppc/sdk/linux.git/commit/?id=b79a8a0528b8a0008cb76f339806b1200b5d8f63

Well, the chip I have is rev1... since I'm the maintainer I might pull a
"Linus" and just apply the damn patch to make it work for me :-) Though
I'm being told that a rev2 chip is on its way to me ... I'll wait a few
more days see if that arrives.

Cheers,
Ben.

> Since it is not sent to upstream, it may need to be rebased if apply it to latest tree.
> 
> Also there might be another issue(SATA & USB) if you test 64 bit kernel on p5020ds with DRR > 4G, we have a patch
> Sent to upstream, but not accepted yet. Link as below:
> http://patchwork.ozlabs.org/patch/179828/
> 
> There are online documents of SDK, for the p5020ds hardware please see the link:
> http://www.freescale.com/infocenter/index.jsp?topic=%2FQORIQSDK%2F2989554.html
> 
> 
> Best Regards, 
> Shaohui Xie
> > -----Original Message-----
> > From: Benjamin Herrenschmidt [mailto:benh at kernel.crashing.org]
> > Sent: Thursday, May 16, 2013 12:48 PM
> > To: Liu Qiang-B32616; Xie Shaohui-B21989; Fleming Andy-AFLEMING
> > Cc: linuxppc-dev at lists.ozlabs.org; Kumar Gala
> > Subject: SATA FSL and upstreaming
> > 
> > Hi folks !
> > 
> > So I was trying to use my 5020ds to test some stuff today. Since I hadn't
> > used it in a while, I decided to "upgrade" it to the latest NOR etc...
> > 
> > Interestingly I discovered that the SATA (which was supposedly dead on
> > the rev1 chip) was actually working with the SDK kernel, while it's still
> > completely busted upstream.
> > 
> > A quick git compare shows about 5 or 6 commits in the SDK tree, some as
> > old as 2011, fixing various erratas in that chip, that never made their
> > way upstream.
> > 
> > Any reason for that ? Being GPL, I can submit them to Tejun myself but it
> > would be better form if you guys did :-)
> > 
> > BTW. Also what's the status with getting the network working upstream ?
> > Even if sub-standard the code could at least go into staging...
> > 
> > Cheers,
> > Ben.
> > 
> > 
>