On Thu, 2012-12-13 at 19:55 +0100, Hauke Mehrtens wrote:
> On 12/11/2012 05:56 AM, Nathan Hintz wrote:
> > I noticed these changes when reviewing the 2011 Broadcom SDK. I haven't
> > noticed any obvious changes in behavior with them applied; but thought I
> > should at least pass them on.
> >
> > Signed-off-by: Nathan Hintz <nlhi...@hotmail.com>
> >
>
> Thank you for your patch, it was committed in r34670.
>
> Could you also send this patch for linux mainline inclusion to the linux
> wireless mailing list. Split it into two patches before sending.
>
> Hauke
>
Hi Hauke:
The first part of this patch had been sent up stream once before; but
was never incorporated, based on the following discussion (the patch was
thought to be incomplete).
http://www.spinics.net/lists/linux-wireless/msg100664.html
What I took from Arend's comments was that a form of "flush_write"
should be implied in this instance, and perhaps others. I don't have
the background to be able to decipher which flush's are needed for all
devices, and which are only needed for certain BCM47xx SOCs. Do you
want to take a cut at this, or do you think this single patch is
adequate and should be pushed upstream as is?
Also, it seems odd to me that some of the writes to BCMA_CC_PLLCTL_ADDR
in driver_chipcommon_pmu.c are flushed (see
bcma_chipco_pll_read/write/maskset), while others are not
(bcma_pmu_spuravoid_pll_write). It is this way in the Broadcom SDK as
well. Should the calls to "bcma_pmu_spuravoid_pll_write" be replaced by
calls to "bcma_chipco_pll_write"?
If you want to send in the second part of the patch yourself (if it's
easy for you to do), I wouldn't mind; as I'm still a 'git' noob.
Thanks,
Nathan
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
