| Message ID | 1425322206-30115-1-git-send-email-mporter@konsulko.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Mar 02, 2015 at 01:50:06PM -0500, Matt Porter wrote: > The PCF8523 RTC is populated on the i2eX boards so > enable the i2c1 controller and enable the device on > that bus. Hmm, I wonder where my update patches went... oh, they're still sitting in my excessive backlog of patches in my tree as part of my "sgtl5000 hacks" patch...
On Mon, Mar 02, 2015 at 07:02:04PM +0000, Russell King - ARM Linux wrote: > On Mon, Mar 02, 2015 at 01:50:06PM -0500, Matt Porter wrote: > > The PCF8523 RTC is populated on the i2eX boards so > > enable the i2c1 controller and enable the device on > > that bus. > > Hmm, I wonder where my update patches went... oh, they're still sitting > in my excessive backlog of patches in my tree as part of my "sgtl5000 > hacks" patch... Also note that before we accept patches to the SolidRun stuff, we should first re-license the file as GPLv2/X11.
[added Rabeeh] On Mon, Mar 02, 2015 at 07:06:16PM +0000, Russell King wrote: > On Mon, Mar 02, 2015 at 07:02:04PM +0000, Russell King - ARM Linux wrote: > > On Mon, Mar 02, 2015 at 01:50:06PM -0500, Matt Porter wrote: > > > The PCF8523 RTC is populated on the i2eX boards so > > > enable the i2c1 controller and enable the device on > > > that bus. > > > > Hmm, I wonder where my update patches went... oh, they're still sitting > > in my excessive backlog of patches in my tree as part of my "sgtl5000 > > hacks" patch... > > Also note that before we accept patches to the SolidRun stuff, we should > first re-license the file as GPLv2/X11. Hi Russell, Given that there's no license information at all on any of these three dts files I think that's a good idea. I just realized I didn't copy Rabeeh on this and he's apparently the original submitter based on your work. Ideally Rabeeh should send a license update but I could do so if you'd like and gather some acks. OTOH, did you already have a license update somewhere in your backlog of patches? -Matt
On Mon, Mar 02, 2015 at 02:21:16PM -0500, Matt Porter wrote: > [added Rabeeh] > > On Mon, Mar 02, 2015 at 07:06:16PM +0000, Russell King wrote: > > On Mon, Mar 02, 2015 at 07:02:04PM +0000, Russell King - ARM Linux wrote: > > > On Mon, Mar 02, 2015 at 01:50:06PM -0500, Matt Porter wrote: > > > > The PCF8523 RTC is populated on the i2eX boards so > > > > enable the i2c1 controller and enable the device on > > > > that bus. > > > > > > Hmm, I wonder where my update patches went... oh, they're still sitting > > > in my excessive backlog of patches in my tree as part of my "sgtl5000 > > > hacks" patch... > > > > Also note that before we accept patches to the SolidRun stuff, we should > > first re-license the file as GPLv2/X11. > > Hi Russell, > > Given that there's no license information at all on any of these three > dts files I think that's a good idea. I just realized I didn't copy > Rabeeh on this and he's apparently the original submitter based on > your work. Ideally Rabeeh should send a license update but I could do > so if you'd like and gather some acks. It's more up to me, as I ended up being the initial author of the mainline files - and if you look at: http://www.home.arm.linux.org.uk/~rmk/cubox/hummingboard-cubox-i-v3.16-rc2-20140624/0041-sgtl5000-hacks.patch you'll see the patch I refer to above, which is my own work, and it's hardly changed to the 3.18 patch: http://www.home.arm.linux.org.uk/~rmk/cubox/hummingboard-cubox-i-v3.18-20141216/0113-sgtl5000-hacks.patch > OTOH, did you already have a license update somewhere in your backlog > of patches? Well, I do now. :) I had dropped the ball on a load of kernel stuff as it seemed to be going nowhere, no one seemed interested in it, and my massive backlog of patches (especially things like the ethernet driver) just became far too difficult both to merge into mainline and to keep forward-porting. In any case, anyone who is serious about using the SolidRun hardware runs Jon Nettleton's fully featured kernel (which is the basis for SR's kernel) - sadly something which mainline likely is never going to be. It's a bit like the MMC stuff, I'm still carrying Olof's solution for the SDIO wifi/bt power and reset control stuff, and I'm not at present intending to do anything with it other than continue forward-porting it. I'm not interested in wasting free time trying to re-work that to suit some other solution, especially as people couldn't settle on a solution when I /did/ have an interest in it (not that I have much interest in wifi or BT - I tend to prefer old fashioned wired connections. It also doesn't help that the Broadcom driver seemed to be very flakey with brcm4329 hardware for quite some time.) Anyway, you can find all my kernel patches in a suitably trimmed version of the above URL, and on http://ftp.arm.linux.org.uk/cgit/ is a huge pile of effort I put into an accelerated X server with etnaviv and /my/ kernel version of etnaviv drm, complete with Xv support.
On Mon, Mar 02, 2015 at 07:49:08PM +0000, Russell King wrote: > On Mon, Mar 02, 2015 at 02:21:16PM -0500, Matt Porter wrote: > > [added Rabeeh] > > > > On Mon, Mar 02, 2015 at 07:06:16PM +0000, Russell King wrote: > > > On Mon, Mar 02, 2015 at 07:02:04PM +0000, Russell King - ARM Linux wrote: > > > > On Mon, Mar 02, 2015 at 01:50:06PM -0500, Matt Porter wrote: > > > > > The PCF8523 RTC is populated on the i2eX boards so > > > > > enable the i2c1 controller and enable the device on > > > > > that bus. > > > > > > > > Hmm, I wonder where my update patches went... oh, they're still sitting > > > > in my excessive backlog of patches in my tree as part of my "sgtl5000 > > > > hacks" patch... > > > > > > Also note that before we accept patches to the SolidRun stuff, we should > > > first re-license the file as GPLv2/X11. > > > > Hi Russell, > > > > Given that there's no license information at all on any of these three > > dts files I think that's a good idea. I just realized I didn't copy > > Rabeeh on this and he's apparently the original submitter based on > > your work. Ideally Rabeeh should send a license update but I could do > > so if you'd like and gather some acks. > > It's more up to me, as I ended up being the initial author of the > mainline files - and if you look at: > > http://www.home.arm.linux.org.uk/~rmk/cubox/hummingboard-cubox-i-v3.16-rc2-20140624/0041-sgtl5000-hacks.patch > > you'll see the patch I refer to above, which is my own work, and it's > hardly changed to the 3.18 patch: > > http://www.home.arm.linux.org.uk/~rmk/cubox/hummingboard-cubox-i-v3.18-20141216/0113-sgtl5000-hacks.patch Ok > > OTOH, did you already have a license update somewhere in your backlog > > of patches? > > Well, I do now. :) Ok :) > I had dropped the ball on a load of kernel stuff as it seemed to be going > nowhere, no one seemed interested in it, and my massive backlog of patches > (especially things like the ethernet driver) just became far too difficult > both to merge into mainline and to keep forward-porting. Understood, I recall that thread but didn't have the hardware back then. ;) > In any case, anyone who is serious about using the SolidRun hardware runs > Jon Nettleton's fully featured kernel (which is the basis for SR's kernel) > - sadly something which mainline likely is never going to be. Right, I also make use of that kernel, of course. It's just too old to make me completely happy. > It's a bit like the MMC stuff, I'm still carrying Olof's solution for > the SDIO wifi/bt power and reset control stuff, and I'm not at present > intending to do anything with it other than continue forward-porting it. > I'm not interested in wasting free time trying to re-work that to suit > some other solution, especially as people couldn't settle on a solution > when I /did/ have an interest in it (not that I have much interest in > wifi or BT - I tend to prefer old fashioned wired connections. It also > doesn't help that the Broadcom driver seemed to be very flakey with > brcm4329 hardware for quite some time.) > > Anyway, you can find all my kernel patches in a suitably trimmed version > of the above URL, and on http://ftp.arm.linux.org.uk/cgit/ is a huge pile > of effort I put into an accelerated X server with etnaviv and /my/ kernel > version of etnaviv drm, complete with Xv support. Excellent, so I feel like you ultimately giving me mixed messages here. You said you need to take care of this licensing issue and I think you implied you'd take care of the audio and rtc stuff as you have that patch in your backlog. However, sounds like you've given up on pushing the bigger things upstream given the problem with getting agreement on those pieces. Are you just going to be submitting the less controversial portions like the audio and dts updates? My goal for HB initially is just to have all the low-hanging fruit (like rtc and audio) working on mainline...besides the stuff that's already there. Hrm, I also don't care about WiFi/BT on this board, I also prefer wired, but on a custom board I'll have to support a different WiFi/BT vendor so I'll take a look at that reset solution you are carrying. Thanks, Matt
On Mon, Mar 02, 2015 at 03:25:40PM -0500, Matt Porter wrote: > On Mon, Mar 02, 2015 at 07:49:08PM +0000, Russell King wrote: > > It's a bit like the MMC stuff, I'm still carrying Olof's solution for > > the SDIO wifi/bt power and reset control stuff, and I'm not at present > > intending to do anything with it other than continue forward-porting it. > > I'm not interested in wasting free time trying to re-work that to suit > > some other solution, especially as people couldn't settle on a solution > > when I /did/ have an interest in it (not that I have much interest in > > wifi or BT - I tend to prefer old fashioned wired connections. It also > > doesn't help that the Broadcom driver seemed to be very flakey with > > brcm4329 hardware for quite some time.) > > > > Anyway, you can find all my kernel patches in a suitably trimmed version > > of the above URL, and on http://ftp.arm.linux.org.uk/cgit/ is a huge pile > > of effort I put into an accelerated X server with etnaviv and /my/ kernel > > version of etnaviv drm, complete with Xv support. > > Excellent, so I feel like you ultimately giving me mixed messages here. Yes, I am - partly because I don't know what to do with many of the patches I seem to be endlessly carrying (which depresses me). I look back to the days when I could be sure of completely emptying my tree at each merge window, something which /never/ happens anymore. > You said you need to take care of this licensing issue and I think you > implied you'd take care of the audio and rtc stuff as you have that > patch in your backlog. As you may have seen, I've mailed out the licensing patch, and I've also mailed out two patches - one for PWM stuff and the second one being the RTC change broken out from my sgtl5000 hacks patch. I've also cut a v4.0-rc1 patch set, in the usual diff.bz2, tar.bz2, and separate patch forms which include those three patches. > However, sounds like you've given up on pushing > the bigger things upstream given the problem with getting agreement > on those pieces. Are you just going to be submitting the > less controversial portions like the audio and dts updates? My goal for > HB initially is just to have all the low-hanging fruit (like rtc and audio) > working on mainline...besides the stuff that's already there. RTC is the easiest bit of the problem, as you've discovered it's just a matter of uncommenting the bit in DT. (It's even better if you have a battery to plug into the little connector!) Audio needs more than just DT changes, and there's recently been some SGTL5000 patches submitted which change the way power is controlled in the SGTL5000. I don't yet know whether those patches solve the problems I was seeing with the kernel powering down the SGTL5000's internal regulator, thereby making the device totally inaccessible until the next power cycle or not. Whether I'll get a chance to look at that this week or not, I don't know... (see my privately shared G+ status from yesterday, which people in my G+ circles should be able to see.)
On Mon, Mar 02, 2015 at 08:43:26PM +0000, Russell King wrote: > On Mon, Mar 02, 2015 at 03:25:40PM -0500, Matt Porter wrote: > > On Mon, Mar 02, 2015 at 07:49:08PM +0000, Russell King wrote: > > > It's a bit like the MMC stuff, I'm still carrying Olof's solution for > > > the SDIO wifi/bt power and reset control stuff, and I'm not at present > > > intending to do anything with it other than continue forward-porting it. > > > I'm not interested in wasting free time trying to re-work that to suit > > > some other solution, especially as people couldn't settle on a solution > > > when I /did/ have an interest in it (not that I have much interest in > > > wifi or BT - I tend to prefer old fashioned wired connections. It also > > > doesn't help that the Broadcom driver seemed to be very flakey with > > > brcm4329 hardware for quite some time.) > > > > > > Anyway, you can find all my kernel patches in a suitably trimmed version > > > of the above URL, and on http://ftp.arm.linux.org.uk/cgit/ is a huge pile > > > of effort I put into an accelerated X server with etnaviv and /my/ kernel > > > version of etnaviv drm, complete with Xv support. > > > > Excellent, so I feel like you ultimately giving me mixed messages here. > > Yes, I am - partly because I don't know what to do with many of the > patches I seem to be endlessly carrying (which depresses me). I look > back to the days when I could be sure of completely emptying my tree > at each merge window, something which /never/ happens anymore. > > > You said you need to take care of this licensing issue and I think you > > implied you'd take care of the audio and rtc stuff as you have that > > patch in your backlog. > > As you may have seen, I've mailed out the licensing patch, and I've > also mailed out two patches - one for PWM stuff and the second one > being the RTC change broken out from my sgtl5000 hacks patch. > > I've also cut a v4.0-rc1 patch set, in the usual diff.bz2, tar.bz2, > and separate patch forms which include those three patches. Ok, caught up on those now, thanks. > > However, sounds like you've given up on pushing > > the bigger things upstream given the problem with getting agreement > > on those pieces. Are you just going to be submitting the > > less controversial portions like the audio and dts updates? My goal for > > HB initially is just to have all the low-hanging fruit (like rtc and audio) > > working on mainline...besides the stuff that's already there. > > RTC is the easiest bit of the problem, as you've discovered it's just > a matter of uncommenting the bit in DT. (It's even better if you have > a battery to plug into the little connector!) > > Audio needs more than just DT changes, and there's recently been some > SGTL5000 patches submitted which change the way power is controlled > in the SGTL5000. I don't yet know whether those patches solve the > problems I was seeing with the kernel powering down the SGTL5000's > internal regulator, thereby making the device totally inaccessible > until the next power cycle or not. Ok, I'll take a look at those here. > Whether I'll get a chance to look at that this week or not, I don't > know... (see my privately shared G+ status from yesterday, which > people in my G+ circles should be able to see.) Aha, ok, I only saw your last public mention of the 4.0-rc1 breakage on the ethernet clock. Now I know why. -Matt
diff --git a/arch/arm/boot/dts/imx6qdl-hummingboard.dtsi b/arch/arm/boot/dts/imx6qdl-hummingboard.dtsi index 62841e8..38ebba9 100644 --- a/arch/arm/boot/dts/imx6qdl-hummingboard.dtsi +++ b/arch/arm/boot/dts/imx6qdl-hummingboard.dtsi @@ -73,18 +73,15 @@ }; &i2c1 { + clock-frequency = <400000>; pinctrl-names = "default"; pinctrl-0 = <&pinctrl_hummingboard_i2c1>; - - /* - * Not fitted on Carrier-1 board... yet status = "okay"; rtc: pcf8523@68 { compatible = "nxp,pcf8523"; reg = <0x68>; }; - */ }; &i2c2 {
The PCF8523 RTC is populated on the i2eX boards so enable the i2c1 controller and enable the device on that bus. Signed-off-by: Matt Porter <mporter@konsulko.com> --- arch/arm/boot/dts/imx6qdl-hummingboard.dtsi | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)