| Message ID | 20141105053414.GA24124@nvidia.com |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
This is very strange. Many people have tested this series; I myself boot my rev.D all the time without SATA connected and I have never encountered this. I just tested the commit you mentioned and my Jetson boots up fine with it. Based on my device's kernel log, the next message indeed should be from the ATA framework indicating a new detected interface. This would correspond to line drivers/ata/ahci_tegra.c:357, so the problem should lie between the lines ahci_tegra.c:299 and ahci_tegra.c:357. Perhaps you could do some printk bisecting to narrow that down? Thanks, Mikko On 11/05/2014 07:34 AM, Allen Martin wrote: > I tried booting 3.18-rc3 on jetson-tk1 and I'm seeing a hang during > boot right before user space normally comes up. I bisected the > regression back to this commit which just turns on AHCI. I don't have > any SATA device connected to the board. > > > commit 91816a7fe5a671b826d5288ed7b42e92117a8fe8 > Author: Mikko Perttunen <mperttunen@nvidia.com> > Date: Wed Jul 16 11:54:23 2014 +0300 > > ARM: tegra: Add options for Tegra AHCI support to tegra_defconfig > > This adds ATA, SATA_AHCI and AHCI_TEGRA support to tegra_defconfig > so that the SATA support will be automatically enabled. > > Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com> > Signed-off-by: Stephen Warren <swarren@nvidia.com> > > diff --git a/arch/arm/configs/tegra_defconfig > b/arch/arm/configs/tegra_defconfig > index 285c433..2b8be48 100644 > --- a/arch/arm/configs/tegra_defconfig > +++ b/arch/arm/configs/tegra_defconfig > @@ -102,6 +102,9 @@ CONFIG_BLK_DEV_SD=y > CONFIG_BLK_DEV_SR=y > CONFIG_SCSI_MULTI_LUN=y > # CONFIG_SCSI_LOWLEVEL is not set > +CONFIG_ATA=y > +CONFIG_SATA_AHCI=y > +CONFIG_AHCI_TEGRA=y > CONFIG_NETDEVICES=y > CONFIG_DUMMY=y > CONFIG_IGB=y > > > > Here's where it hangs, always at the same place: > > ... > [ 2.370671] +1.35V_LP0(sd2): supplied by +5V_SYS > [ 2.375732] +1.35V_LP0(sd3): 1350 mV > [ 2.379522] +1.35V_LP0(sd3): supplied by +5V_SYS > [ 2.384551] +1.05V_RUN: 1050 mV > [ 2.387918] +1.05V_RUN: supplied by +5V_SYS > [ 2.392482] +1.8V_VDDIO: 1800 mV > [ 2.395944] +1.8V_VDDIO: supplied by +5V_SYS > [ 2.401498] +VDD_GPU_AP: 650 <--> 1200 mV at 1000 mV 3500 mA > [ 2.407734] +1.05V_RUN_AVDD: 1050 mV at 300 mA > [ 2.412400] +1.05V_RUN_AVDD: supplied by +1.35V_LP0(sd2) > [ 2.418329] +1.8V_RUN_CAM: 1800 mV at 150 mA > [ 2.422838] +1.8V_RUN_CAM: supplied by +3.3V_RUN > [ 2.428665] +1.2V_GEN_AVDD: 1200 mV at 150 mA > [ 2.433268] +1.2V_GEN_AVDD: supplied by +1.8V_VDDIO > [ 2.439489] +1.05V_LP0_VDD_RTC: 1000 mV at 150 mA > [ 2.444426] +1.05V_LP0_VDD_RTC: supplied by +3.3V_SYS > [ 2.450185] +2.8V_RUN_CAM: 2800 mV at 150 mA > [ 2.454711] +2.8V_RUN_CAM: supplied by +3.3V_SYS > [ 2.460022] +1.2V_RUN_CAM_FRONT: 1200 mV at 150 mA > [ 2.465079] +1.2V_RUN_CAM_FRONT: supplied by +1.8V_VDDIO > [ 2.470958] +VDDIO_SDMMC3: 1800 <--> 3300 mV at 150 mA > [ 2.476354] +VDDIO_SDMMC3: supplied by +3.3V_RUN > [ 2.481675] +1.05V_RUN_CAM_REAR: 1050 mV at 150 mA > [ 2.486721] +1.05V_RUN_CAM_REAR: supplied by +1.8V_VDDIO > [ 2.492904] +3.3V_RUN_TOUCH: 2800 mV at 150 mA > [ 2.497608] +3.3V_RUN_TOUCH: supplied by +5V_SYS > [ 2.502933] +2.8V_RUN_CAM_AF: 2800 mV at 150 mA > [ 2.507724] +2.8V_RUN_CAM_AF: supplied by +5V_SYS > [ 2.513126] +1.8V_RUN_VPP_FUSE: 1800 mV at 150 mA > [ 2.518101] +1.8V_RUN_VPP_FUSE: supplied by +3.3V_RUN > *hang* > -- > To unsubscribe from this list: send the line "unsubscribe linux-tegra" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
The hang is the following register write in tegra_ahci_controller_init():
writel(BIT(0), tegra->sata_regs + SCFG_OFFSET + T_SATA0_INDEX);
here are some debug prints I added:
[ 2.893245] tegra_ahci_controller_init: SCFG
[ 2.897525] tegra->sata_regs: fe020000
[ 2.901273] SCFG_OFFSET: 0x1000
[ 2.904426] T_SATA0_INDEX: 0x680
[ 2.907653] writing 0x1 to fe021680
A hung PIO usually means the IP block is powered off or in reset, but the previous write to SATA_CONFIGURATION_0 goes through ok.
Which board revision do you have? I have tested the series on both rev.D and rev.C. I believe only rev.D is "officially" supported on upstream. On 11/05/2014 09:46 PM, Allen Martin wrote: > The hang is the following register write in tegra_ahci_controller_init(): > > writel(BIT(0), tegra->sata_regs + SCFG_OFFSET + T_SATA0_INDEX); > > here are some debug prints I added: > > [ 2.893245] tegra_ahci_controller_init: SCFG > [ 2.897525] tegra->sata_regs: fe020000 > [ 2.901273] SCFG_OFFSET: 0x1000 > [ 2.904426] T_SATA0_INDEX: 0x680 > [ 2.907653] writing 0x1 to fe021680 > > A hung PIO usually means the IP block is powered off or in reset, but the previous write to SATA_CONFIGURATION_0 goes through ok. > > ________________________________________ > From: Mikko Perttunen <mikko.perttunen@kapsi.fi> > Sent: Wednesday, November 5, 2014 9:42 AM > To: Allen Martin; Mikko Perttunen > Cc: linux-tegra@vger.kernel.org > Subject: Re: boot regression with AHCI on jetson-tk1 > > This is very strange. Many people have tested this series; I myself boot > my rev.D all the time without SATA connected and I have never > encountered this. I just tested the commit you mentioned and my Jetson > boots up fine with it. > > Based on my device's kernel log, the next message indeed should be from > the ATA framework indicating a new detected interface. This would > correspond to line drivers/ata/ahci_tegra.c:357, so the problem should > lie between the lines ahci_tegra.c:299 and ahci_tegra.c:357. Perhaps you > could do some printk bisecting to narrow that down? > > Thanks, > Mikko > > On 11/05/2014 07:34 AM, Allen Martin wrote: >> I tried booting 3.18-rc3 on jetson-tk1 and I'm seeing a hang during >> boot right before user space normally comes up. I bisected the >> regression back to this commit which just turns on AHCI. I don't have >> any SATA device connected to the board. >> >> >> commit 91816a7fe5a671b826d5288ed7b42e92117a8fe8 >> Author: Mikko Perttunen <mperttunen@nvidia.com> >> Date: Wed Jul 16 11:54:23 2014 +0300 >> >> ARM: tegra: Add options for Tegra AHCI support to tegra_defconfig >> >> This adds ATA, SATA_AHCI and AHCI_TEGRA support to tegra_defconfig >> so that the SATA support will be automatically enabled. >> >> Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com> >> Signed-off-by: Stephen Warren <swarren@nvidia.com> >> >> diff --git a/arch/arm/configs/tegra_defconfig >> b/arch/arm/configs/tegra_defconfig >> index 285c433..2b8be48 100644 >> --- a/arch/arm/configs/tegra_defconfig >> +++ b/arch/arm/configs/tegra_defconfig >> @@ -102,6 +102,9 @@ CONFIG_BLK_DEV_SD=y >> CONFIG_BLK_DEV_SR=y >> CONFIG_SCSI_MULTI_LUN=y >> # CONFIG_SCSI_LOWLEVEL is not set >> +CONFIG_ATA=y >> +CONFIG_SATA_AHCI=y >> +CONFIG_AHCI_TEGRA=y >> CONFIG_NETDEVICES=y >> CONFIG_DUMMY=y >> CONFIG_IGB=y >> >> >> >> Here's where it hangs, always at the same place: >> >> ... >> [ 2.370671] +1.35V_LP0(sd2): supplied by +5V_SYS >> [ 2.375732] +1.35V_LP0(sd3): 1350 mV >> [ 2.379522] +1.35V_LP0(sd3): supplied by +5V_SYS >> [ 2.384551] +1.05V_RUN: 1050 mV >> [ 2.387918] +1.05V_RUN: supplied by +5V_SYS >> [ 2.392482] +1.8V_VDDIO: 1800 mV >> [ 2.395944] +1.8V_VDDIO: supplied by +5V_SYS >> [ 2.401498] +VDD_GPU_AP: 650 <--> 1200 mV at 1000 mV 3500 mA >> [ 2.407734] +1.05V_RUN_AVDD: 1050 mV at 300 mA >> [ 2.412400] +1.05V_RUN_AVDD: supplied by +1.35V_LP0(sd2) >> [ 2.418329] +1.8V_RUN_CAM: 1800 mV at 150 mA >> [ 2.422838] +1.8V_RUN_CAM: supplied by +3.3V_RUN >> [ 2.428665] +1.2V_GEN_AVDD: 1200 mV at 150 mA >> [ 2.433268] +1.2V_GEN_AVDD: supplied by +1.8V_VDDIO >> [ 2.439489] +1.05V_LP0_VDD_RTC: 1000 mV at 150 mA >> [ 2.444426] +1.05V_LP0_VDD_RTC: supplied by +3.3V_SYS >> [ 2.450185] +2.8V_RUN_CAM: 2800 mV at 150 mA >> [ 2.454711] +2.8V_RUN_CAM: supplied by +3.3V_SYS >> [ 2.460022] +1.2V_RUN_CAM_FRONT: 1200 mV at 150 mA >> [ 2.465079] +1.2V_RUN_CAM_FRONT: supplied by +1.8V_VDDIO >> [ 2.470958] +VDDIO_SDMMC3: 1800 <--> 3300 mV at 150 mA >> [ 2.476354] +VDDIO_SDMMC3: supplied by +3.3V_RUN >> [ 2.481675] +1.05V_RUN_CAM_REAR: 1050 mV at 150 mA >> [ 2.486721] +1.05V_RUN_CAM_REAR: supplied by +1.8V_VDDIO >> [ 2.492904] +3.3V_RUN_TOUCH: 2800 mV at 150 mA >> [ 2.497608] +3.3V_RUN_TOUCH: supplied by +5V_SYS >> [ 2.502933] +2.8V_RUN_CAM_AF: 2800 mV at 150 mA >> [ 2.507724] +2.8V_RUN_CAM_AF: supplied by +5V_SYS >> [ 2.513126] +1.8V_RUN_VPP_FUSE: 1800 mV at 150 mA >> [ 2.518101] +1.8V_RUN_VPP_FUSE: supplied by +3.3V_RUN >> *hang* >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-tegra" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 11/04/2014 10:34 PM, Allen Martin wrote: > I tried booting 3.18-rc3 on jetson-tk1 and I'm seeing a hang during > boot right before user space normally comes up. I bisected the > regression back to this commit which just turns on AHCI. I don't have > any SATA device connected to the board. I just tried this, and everything works out OK for me with 3.18-rc3 or next-20141016 which I had lying around. Which board revision are you using? I'm using rev D. Which U-Boot revision are you using? Perhaps this is an issue with old U-Boots that don't initialize the full pinmux, now that the kernel doesn't and hence relies on U-Boot. I tested there weren't any kernel regressions when I updated it not to program the static pinmux, but perhaps SATA wasn't enabled when I did that. -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
I have a retail rev D jetson-tk1, I'm in the lab with Greg Roth, and I tried my sdcard on his rev C and it boots fine. I've never used SATA on this board, but the l4t kernel boots ok, which at least probes sata without issue.
I thought of u-boot dependencies as well, but I updated to:
commit 46b7b2e8025770fb29f7810c2d7d15a6650b9643
Merge: a8b972b 265072b
Author: Tom Rini <trini@ti.com>
Date: Mon Nov 3 12:46:12 2014 -0500
Merge branch 'master' of git://git.denx.de/u-boot-mips
and it still fails
diff --git a/arch/arm/configs/tegra_defconfig b/arch/arm/configs/tegra_defconfig index 285c433..2b8be48 100644 --- a/arch/arm/configs/tegra_defconfig +++ b/arch/arm/configs/tegra_defconfig @@ -102,6 +102,9 @@ CONFIG_BLK_DEV_SD=y CONFIG_BLK_DEV_SR=y CONFIG_SCSI_MULTI_LUN=y # CONFIG_SCSI_LOWLEVEL is not set +CONFIG_ATA=y +CONFIG_SATA_AHCI=y +CONFIG_AHCI_TEGRA=y CONFIG_NETDEVICES=y CONFIG_DUMMY=y CONFIG_IGB=y
I tried booting 3.18-rc3 on jetson-tk1 and I'm seeing a hang during boot right before user space normally comes up. I bisected the regression back to this commit which just turns on AHCI. I don't have any SATA device connected to the board. commit 91816a7fe5a671b826d5288ed7b42e92117a8fe8 Author: Mikko Perttunen <mperttunen@nvidia.com> Date: Wed Jul 16 11:54:23 2014 +0300 ARM: tegra: Add options for Tegra AHCI support to tegra_defconfig This adds ATA, SATA_AHCI and AHCI_TEGRA support to tegra_defconfig so that the SATA support will be automatically enabled. Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com> Signed-off-by: Stephen Warren <swarren@nvidia.com> Here's where it hangs, always at the same place: ... [ 2.370671] +1.35V_LP0(sd2): supplied by +5V_SYS [ 2.375732] +1.35V_LP0(sd3): 1350 mV [ 2.379522] +1.35V_LP0(sd3): supplied by +5V_SYS [ 2.384551] +1.05V_RUN: 1050 mV [ 2.387918] +1.05V_RUN: supplied by +5V_SYS [ 2.392482] +1.8V_VDDIO: 1800 mV [ 2.395944] +1.8V_VDDIO: supplied by +5V_SYS [ 2.401498] +VDD_GPU_AP: 650 <--> 1200 mV at 1000 mV 3500 mA [ 2.407734] +1.05V_RUN_AVDD: 1050 mV at 300 mA [ 2.412400] +1.05V_RUN_AVDD: supplied by +1.35V_LP0(sd2) [ 2.418329] +1.8V_RUN_CAM: 1800 mV at 150 mA [ 2.422838] +1.8V_RUN_CAM: supplied by +3.3V_RUN [ 2.428665] +1.2V_GEN_AVDD: 1200 mV at 150 mA [ 2.433268] +1.2V_GEN_AVDD: supplied by +1.8V_VDDIO [ 2.439489] +1.05V_LP0_VDD_RTC: 1000 mV at 150 mA [ 2.444426] +1.05V_LP0_VDD_RTC: supplied by +3.3V_SYS [ 2.450185] +2.8V_RUN_CAM: 2800 mV at 150 mA [ 2.454711] +2.8V_RUN_CAM: supplied by +3.3V_SYS [ 2.460022] +1.2V_RUN_CAM_FRONT: 1200 mV at 150 mA [ 2.465079] +1.2V_RUN_CAM_FRONT: supplied by +1.8V_VDDIO [ 2.470958] +VDDIO_SDMMC3: 1800 <--> 3300 mV at 150 mA [ 2.476354] +VDDIO_SDMMC3: supplied by +3.3V_RUN [ 2.481675] +1.05V_RUN_CAM_REAR: 1050 mV at 150 mA [ 2.486721] +1.05V_RUN_CAM_REAR: supplied by +1.8V_VDDIO [ 2.492904] +3.3V_RUN_TOUCH: 2800 mV at 150 mA [ 2.497608] +3.3V_RUN_TOUCH: supplied by +5V_SYS [ 2.502933] +2.8V_RUN_CAM_AF: 2800 mV at 150 mA [ 2.507724] +2.8V_RUN_CAM_AF: supplied by +5V_SYS [ 2.513126] +1.8V_RUN_VPP_FUSE: 1800 mV at 150 mA [ 2.518101] +1.8V_RUN_VPP_FUSE: supplied by +3.3V_RUN *hang* -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html