| Message ID | 79CD15C6BA57404B839C016229A409A83EC20D27@DBDE01.ent.ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Mar 07, 2013 at 03:50:01PM +0000, Vaibhav Hiremath wrote: > > > -----Original Message----- > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of Porter, Matt > > Sent: Thursday, March 07, 2013 8:34 PM > > To: Hiremath, Vaibhav > > Cc: Chris Ball; Russell King; Krishnamoorthy, Balaji T; Devicetree > > Discuss; Linux MMC List; Linux Kernel Mailing List; Linux OMAP List; > > Linux ARM Kernel List > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > On Thu, Mar 07, 2013 at 02:59:42PM +0000, Vaibhav Hiremath wrote: > > > > > > > -----Original Message----- > > > > From: Hiremath, Vaibhav > > > > Sent: Thursday, March 07, 2013 8:24 PM > > > > To: Porter, Matt > > > > Cc: Linux OMAP List; Russell King; Krishnamoorthy, Balaji T; > > Devicetree > > > > Discuss; Linux MMC List; Linux Kernel Mailing List; Chris Ball; > > Linux > > > > ARM Kernel List > > > > Subject: RE: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > -----Original Message----- > > > > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of Porter, > > > > Matt > > > > > Sent: Thursday, March 07, 2013 8:17 PM > > > > > To: Hiremath, Vaibhav > > > > > Cc: Linux OMAP List; Russell King; Krishnamoorthy, Balaji T; > > > > Devicetree > > > > > Discuss; Linux MMC List; Linux Kernel Mailing List; Chris Ball; > > Linux > > > > > ARM Kernel List > > > > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > On Thu, Mar 07, 2013 at 02:39:55PM +0000, Vaibhav Hiremath wrote: > > > > > > > > > > > > > -----Original Message----- > > > > > > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of > > Porter, > > > > > Matt > > > > > > > Sent: Thursday, March 07, 2013 7:43 PM > > > > > > > To: Hiremath, Vaibhav > > > > > > > Cc: Krishnamoorthy, Balaji T; Chris Ball; Cousson, Benoit; > > Tony > > > > > > > Lindgren; Russell King; Devicetree Discuss; Linux ARM Kernel > > > > List; > > > > > > > Linux OMAP List; Linux Kernel Mailing List; Linux MMC List > > > > > > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > > > > On Thu, Mar 07, 2013 at 05:29:24AM +0000, Vaibhav Hiremath > > wrote: > > > > > > > > > -----Original Message----- > > > > > > > > > From: linux-omap-owner@vger.kernel.org [mailto:linux- > > omap- > > > > > > > > > owner@vger.kernel.org] On Behalf Of Porter, Matt > > > > > > > > > Sent: Thursday, March 07, 2013 9:47 AM > > > > > > > > > To: Krishnamoorthy, Balaji T; Chris Ball; Cousson, > > Benoit; > > > > Tony > > > > > > > > > Lindgren; Russell King > > > > > > > > > Cc: Devicetree Discuss; Linux ARM Kernel List; Linux OMAP > > > > List; > > > > > > > Linux > > > > > > > > > Kernel Mailing List; Linux MMC List > > > > > > > > > Subject: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > > <snip> > > > > > > > > I believe you meant "CONFIG_TI_EDMA" right? > > > > > > > > Yes, I just enabled it and the result is still same. > > > > > > > > > > > > > > > > [root@arago /]# dmesg | grep -ir mmc > > > > [ 0.506844] vmmc: 1800 <--> 3300 mV at 3300 mV > > > > [ 0.506970] vmmc: supplied by vbat > > > > [root@arago /]# > > > > [root@arago /]# > > > > [root@arago /]# dmesg | grep -ir dma > > > > [ 0.217063] DMA: preallocated 256 KiB pool for atomic coherent > > > > allocations > > > > [ 0.236321] platform 49000000.edma: alias fck already exists > > > > [ 0.236360] platform 49000000.edma: alias fck already exists > > > > [ 0.236381] platform 49000000.edma: alias fck already exists > > > > [ 0.370705] edma-dma-engine edma-dma-engine.0: TI EDMA DMA > > engine > > > > driver > > > > [ 0.445156] omap-dma-engine omap-dma-engine: OMAP DMA engine > > driver > > > > [root@arago /]# > > > > [root@arago /]# > > > > > > > > > > > I have applied below patches from your recent post > > > > > > > > > [2/2] ARM: dts: add AM33XX MMC support > > > [1/2] mmc: omap_hsmmc: set max_segs based on dma engine limits > > > [v4,3/3] mmc: davinci: get SG segment limits with > > dma_get_slave_sg_limits() > > > [v4,2/3] dma: edma: add device_slave_sg_limits() support > > > [v4,1/3] dmaengine: add dma_get_slave_sg_limits() > > > [v9,9/9] ARM: dts: add AM33XX SPI DMA support > > > [v9,8/9] spi: omap2-mcspi: add generic DMA request support to the DT > > binding > > > [v9,7/9] spi: omap2-mcspi: convert to > > dma_request_slave_channel_compat() > > > [v9,6/9] ARM: dts: add AM33XX EDMA support > > > [v9,5/9] dmaengine: edma: Add TI EDMA device tree binding > > > [v9,4/9] dmaengine: edma: enable build for AM33XX > > > [v9,3/9] ARM: edma: add AM33XX support to the private EDMA API > > > [v9,2/9] ARM: edma: remove unused transfer controller handlers > > > [v9,1/9] ARM: davinci: move private EDMA API to arm/common > > > [v3,2/2] mmc: omap_hsmmc: add generic DMA request support to the DT > > binding > > > [v3,1/2] mmc: omap_hsmmc: convert to > > dma_request_slave_channel_compat() > > > > > > > > > > > > Am I missing anything here? > > > > Yes, you missed the > > http://www.spinics.net/lists/arm-kernel/msg227886.html dependency > > mentioned first in the cover letter. > > > > Matt, I manually edited the file with above patch and result is still the same. > Can you point me to branch where you have tested MMC code? git://github.com/ohporter/linux.git edma-dmaengine-am33xx-mmc-v1 omap2plus_defconfig + CONFIG_TI_EDMA I just doublechecked MMC rootfs on bone and evmsk as it's the standard smoke test. My EVM is intermittent now so trying to coax it to power up to reverify. -Matt > diff --git a/arch/arm/common/edma.c b/arch/arm/common/edma.c > index 8c3b1fa..e68ac38 100644 > --- a/arch/arm/common/edma.c > +++ b/arch/arm/common/edma.c > @@ -730,6 +730,9 @@ EXPORT_SYMBOL(edma_free_channel); > */ > int edma_alloc_slot(unsigned ctlr, int slot) > { > + if (!edma_cc[ctlr]) > + return -EINVAL; > + > if (slot >= 0) > slot = EDMA_CHAN_SLOT(slot); > > Thanks, > Vaibhav > _______________________________________________ > devicetree-discuss mailing list > devicetree-discuss@lists.ozlabs.org > https://lists.ozlabs.org/listinfo/devicetree-discuss
> -----Original Message----- > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of Porter, Matt > Sent: Thursday, March 07, 2013 9:33 PM > To: Hiremath, Vaibhav > Cc: Linux OMAP List; Russell King; Krishnamoorthy, Balaji T; Devicetree > Discuss; Linux MMC List; Linux Kernel Mailing List; Chris Ball; Linux > ARM Kernel List > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > On Thu, Mar 07, 2013 at 03:50:01PM +0000, Vaibhav Hiremath wrote: > > > > > -----Original Message----- > > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of Porter, > Matt > > > Sent: Thursday, March 07, 2013 8:34 PM > > > To: Hiremath, Vaibhav > > > Cc: Chris Ball; Russell King; Krishnamoorthy, Balaji T; Devicetree > > > Discuss; Linux MMC List; Linux Kernel Mailing List; Linux OMAP > List; > > > Linux ARM Kernel List > > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > On Thu, Mar 07, 2013 at 02:59:42PM +0000, Vaibhav Hiremath wrote: > > > > > > > > > -----Original Message----- > > > > > From: Hiremath, Vaibhav > > > > > Sent: Thursday, March 07, 2013 8:24 PM > > > > > To: Porter, Matt > > > > > Cc: Linux OMAP List; Russell King; Krishnamoorthy, Balaji T; > > > Devicetree > > > > > Discuss; Linux MMC List; Linux Kernel Mailing List; Chris Ball; > > > Linux > > > > > ARM Kernel List > > > > > Subject: RE: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > -----Original Message----- > > > > > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of > Porter, > > > > > Matt > > > > > > Sent: Thursday, March 07, 2013 8:17 PM > > > > > > To: Hiremath, Vaibhav > > > > > > Cc: Linux OMAP List; Russell King; Krishnamoorthy, Balaji T; > > > > > Devicetree > > > > > > Discuss; Linux MMC List; Linux Kernel Mailing List; Chris > Ball; > > > Linux > > > > > > ARM Kernel List > > > > > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > > On Thu, Mar 07, 2013 at 02:39:55PM +0000, Vaibhav Hiremath > wrote: > > > > > > > > > > > > > > > -----Original Message----- > > > > > > > > From: Matt Porter [mailto:ohiomdp@gmail.com] On Behalf Of > > > Porter, > > > > > > Matt > > > > > > > > Sent: Thursday, March 07, 2013 7:43 PM > > > > > > > > To: Hiremath, Vaibhav > > > > > > > > Cc: Krishnamoorthy, Balaji T; Chris Ball; Cousson, > Benoit; > > > Tony > > > > > > > > Lindgren; Russell King; Devicetree Discuss; Linux ARM > Kernel > > > > > List; > > > > > > > > Linux OMAP List; Linux Kernel Mailing List; Linux MMC > List > > > > > > > > Subject: Re: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > > > > > > On Thu, Mar 07, 2013 at 05:29:24AM +0000, Vaibhav > Hiremath > > > wrote: > > > > > > > > > > -----Original Message----- > > > > > > > > > > From: linux-omap-owner@vger.kernel.org [mailto:linux- > > > omap- > > > > > > > > > > owner@vger.kernel.org] On Behalf Of Porter, Matt > > > > > > > > > > Sent: Thursday, March 07, 2013 9:47 AM > > > > > > > > > > To: Krishnamoorthy, Balaji T; Chris Ball; Cousson, > > > Benoit; > > > > > Tony > > > > > > > > > > Lindgren; Russell King > > > > > > > > > > Cc: Devicetree Discuss; Linux ARM Kernel List; Linux > OMAP > > > > > List; > > > > > > > > Linux > > > > > > > > > > Kernel Mailing List; Linux MMC List > > > > > > > > > > Subject: [PATCH 2/2] ARM: dts: add AM33XX MMC support > > > > > > > > > > > > > > <snip> > > > > > > > > > > I believe you meant "CONFIG_TI_EDMA" right? > > > > > > > > > > Yes, I just enabled it and the result is still same. > > > > > > > > > > > > > > > > > > > > [root@arago /]# dmesg | grep -ir mmc > > > > > [ 0.506844] vmmc: 1800 <--> 3300 mV at 3300 mV > > > > > [ 0.506970] vmmc: supplied by vbat > > > > > [root@arago /]# > > > > > [root@arago /]# > > > > > [root@arago /]# dmesg | grep -ir dma > > > > > [ 0.217063] DMA: preallocated 256 KiB pool for atomic > coherent > > > > > allocations > > > > > [ 0.236321] platform 49000000.edma: alias fck already exists > > > > > [ 0.236360] platform 49000000.edma: alias fck already exists > > > > > [ 0.236381] platform 49000000.edma: alias fck already exists > > > > > [ 0.370705] edma-dma-engine edma-dma-engine.0: TI EDMA DMA > > > engine > > > > > driver > > > > > [ 0.445156] omap-dma-engine omap-dma-engine: OMAP DMA engine > > > driver > > > > > [root@arago /]# > > > > > [root@arago /]# > > > > > > > > > > > > > > I have applied below patches from your recent post > > > > > > > > > > > > [2/2] ARM: dts: add AM33XX MMC support > > > > [1/2] mmc: omap_hsmmc: set max_segs based on dma engine limits > > > > [v4,3/3] mmc: davinci: get SG segment limits with > > > dma_get_slave_sg_limits() > > > > [v4,2/3] dma: edma: add device_slave_sg_limits() support > > > > [v4,1/3] dmaengine: add dma_get_slave_sg_limits() > > > > [v9,9/9] ARM: dts: add AM33XX SPI DMA support > > > > [v9,8/9] spi: omap2-mcspi: add generic DMA request support to the > DT > > > binding > > > > [v9,7/9] spi: omap2-mcspi: convert to > > > dma_request_slave_channel_compat() > > > > [v9,6/9] ARM: dts: add AM33XX EDMA support > > > > [v9,5/9] dmaengine: edma: Add TI EDMA device tree binding > > > > [v9,4/9] dmaengine: edma: enable build for AM33XX > > > > [v9,3/9] ARM: edma: add AM33XX support to the private EDMA API > > > > [v9,2/9] ARM: edma: remove unused transfer controller handlers > > > > [v9,1/9] ARM: davinci: move private EDMA API to arm/common > > > > [v3,2/2] mmc: omap_hsmmc: add generic DMA request support to the > DT > > > binding > > > > [v3,1/2] mmc: omap_hsmmc: convert to > > > dma_request_slave_channel_compat() > > > > > > > > > > > > > > > > Am I missing anything here? > > > > > > Yes, you missed the > > > http://www.spinics.net/lists/arm-kernel/msg227886.html dependency > > > mentioned first in the cover letter. > > > > > > > Matt, I manually edited the file with above patch and result is still > the same. > > Can you point me to branch where you have tested MMC code? > > git://github.com/ohporter/linux.git edma-dmaengine-am33xx-mmc-v1 > > omap2plus_defconfig + CONFIG_TI_EDMA > > I just doublechecked MMC rootfs on bone and evmsk as it's the standard > smoke test. My EVM is intermittent now so trying to coax it to power up > to reverify. > Matt, Your branch is working for me, I tested it on EVM. Not sure what is wrong with manual rebasing I did yesterday. I will try to nail-down root-cause here. And you can submit the next version fixing the comments I have given. Thanks, Vaibhav
Hi Tony, Vaibhav, >> I just doublechecked MMC rootfs on bone and evmsk as it's the standard >> smoke test. My EVM is intermittent now so trying to coax it to power up >> to reverify. >> > > Matt, > > Your branch is working for me, I tested it on EVM. Not sure what is wrong with manual rebasing > I did yesterday. I will try to nail-down root-cause here. > > And you can submit the next version fixing the comments I have given. The only comment I could find in the whole discussion was adding of the interrupts property. I tested I don't have any problem booting with Matt's patch, without having the interrupts property. Can this patch be pulled in for 3.11? Tested-by: Joel A Fernandes <joelagnel@ti.com> Thanks, Joel
Hi, On Fri, Jun 14, 2013 at 02:54:33PM -0500, Joel A Fernandes wrote: > Hi Tony, Vaibhav, > > >> I just doublechecked MMC rootfs on bone and evmsk as it's the standard > >> smoke test. My EVM is intermittent now so trying to coax it to power up > >> to reverify. > >> > > > > Matt, > > > > Your branch is working for me, I tested it on EVM. Not sure what is wrong with manual rebasing > > I did yesterday. I will try to nail-down root-cause here. > > > > And you can submit the next version fixing the comments I have given. > > The only comment I could find in the whole discussion was adding of > the interrupts property. I tested I don't have any problem booting > with Matt's patch, without having the interrupts property. might very well be, but we want to figure out why adding "intrrupts" causes issues. Why does it have to come through hwmod ? You might have uncovered a real problem which needs fixing. Besides, at the end of the day, we want to remove the redundant hwmod data if we already have DTS bindings covering the same requirements.
Hi Felipe, On Fri, Jun 14, 2013 at 6:43 PM, Joel A Fernandes <agnel.joel@gmail.com> wrote: > Felipe, > > > On Friday, June 14, 2013, Felipe Balbi wrote: >> >> Hi, >> >> On Fri, Jun 14, 2013 at 02:54:33PM -0500, Joel A Fernandes wrote: >> > Hi Tony, Vaibhav, >> > >> > >> I just doublechecked MMC rootfs on bone and evmsk as it's the >> > >> standard >> > >> smoke test. My EVM is intermittent now so trying to coax it to power >> > >> up >> > >> to reverify. >> > >> >> > > >> > > Matt, >> > > >> > > Your branch is working for me, I tested it on EVM. Not sure what is >> > > wrong with manual rebasing >> > > I did yesterday. I will try to nail-down root-cause here. >> > > >> > > And you can submit the next version fixing the comments I have given. >> > >> > The only comment I could find in the whole discussion was adding of >> > the interrupts property. I tested I don't have any problem booting >> > with Matt's patch, without having the interrupts property. >> >> might very well be, but we want to figure out why adding "intrrupts" >> causes issues. Why does it have to come through hwmod ? You might have >> uncovered a real problem which needs fixing. Besides, at the end of the > You are right, adding interrupts property to DT node causes problems (driver probe doesn't take place). Will debug further on Monday. Thanks, Joel
Hi, On Fri, Jun 14, 2013 at 06:43:51PM -0500, Joel A Fernandes wrote: > > > >> I just doublechecked MMC rootfs on bone and evmsk as it's the standard > > > >> smoke test. My EVM is intermittent now so trying to coax it to power > > up > > > >> to reverify. > > > >> > > > > > > > > Matt, > > > > > > > > Your branch is working for me, I tested it on EVM. Not sure what is > > wrong with manual rebasing > > > > I did yesterday. I will try to nail-down root-cause here. > > > > > > > > And you can submit the next version fixing the comments I have given. > > > > > > The only comment I could find in the whole discussion was adding of > > > the interrupts property. I tested I don't have any problem booting > > > with Matt's patch, without having the interrupts property. > > > > might very well be, but we want to figure out why adding "intrrupts" > > causes issues. Why does it have to come through hwmod ? You might have > > uncovered a real problem which needs fixing. Besides, at the end of the > > > No actually I think I wasn't very descriptive in my post. There is no > problem here. Whether it comes from DT or not, it works in both cases . (I > haven't tried adding interrupt node to dts but what I gathered from > Vaibhav's post is that it works also when added to dts) > > But I can revisit Vaibhav's suggestion of adding it to dts for completeness > sake. My tree currently works and picks up the interrupt info from hwmod. I > just didn't see any activity on this thread for a while so I thought of > enquiring it was ok to pull it in for the 3.11 merge window or if the > consensus was that interrupt node be added to dts even though it works > without it. fair enough, then let's just add those interrupts to hwmod which will, eventually, let us drop all of that data from hwmod. cheers
* Felipe Balbi <balbi@ti.com> [130615 08:12]: > > fair enough, then let's just add those interrupts to hwmod which will, > eventually, let us drop all of that data from hwmod. Nope, we're already dropping the duplicate data from hwmod for am33xx and omap4. See omap-for-v3.11/cleanup branch. Regards, Tony
diff --git a/arch/arm/common/edma.c b/arch/arm/common/edma.c index 8c3b1fa..e68ac38 100644 --- a/arch/arm/common/edma.c +++ b/arch/arm/common/edma.c @@ -730,6 +730,9 @@ EXPORT_SYMBOL(edma_free_channel); */ int edma_alloc_slot(unsigned ctlr, int slot) { + if (!edma_cc[ctlr]) + return -EINVAL; + if (slot >= 0) slot = EDMA_CHAN_SLOT(slot);