mbox

[GIT,PULL] i.MX fixes for 4.9

Message ID 20161024140622.GM30578@tiger
State New
Headers show

Pull-request

git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git tags/imx-fixes-4.9

Message

Shawn Guo Oct. 24, 2016, 2:06 p.m. UTC
The following changes since commit 1001354ca34179f3db924eb66672442a173147dc:

  Linux 4.9-rc1 (2016-10-15 12:17:50 -0700)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git tags/imx-fixes-4.9

for you to fetch changes up to 4edd601c5a9c5094daa714e65063e623826f3bcc:

  ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031 (2016-10-24 21:26:01 +0800)

----------------------------------------------------------------
The i.MX fixes for 4.9:
 - A couple of patches from Fabio to fix the GPC power domain regression
   which is caused by PM Domain core change 0159ec670763dd
   ("PM / Domains: Verify the PM domain is present when adding a
   provider"), and a related kernel crash seen with multi_v7_defconfig
   build.
 - Correct the PHY ID mask for AR8031 to match phy driver code.
 - Apply new added timer erratum A008585 for LS1043A and LS2080A SoC.
 - Correct vf610 global timer IRQ flag to avoid warning from gic driver
   after commit 992345a58e0c ("irqchip/gic: WARN if setting the
   interrupt type for a PPI fails").

----------------------------------------------------------------
Fabio Estevam (3):
      ARM: imx: gpc: Initialize all power domains
      ARM: imx: gpc: Fix the imx_gpc_genpd_init() error path
      ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031

Scott Wood (1):
      arm64: dts: Add timer erratum property for LS2080A and LS1043A

Stefan Agner (1):
      ARM: dts: vf610: fix IRQ flag of global timer

 arch/arm/boot/dts/vf500.dtsi                   |  2 +-
 arch/arm/mach-imx/gpc.c                        | 15 ++++++++++++---
 arch/arm/mach-imx/mach-imx6q.c                 |  2 +-
 arch/arm64/boot/dts/freescale/fsl-ls1043a.dtsi |  1 +
 arch/arm64/boot/dts/freescale/fsl-ls2080a.dtsi |  1 +
 5 files changed, 16 insertions(+), 5 deletions(-)

Comments

Olof Johansson Oct. 29, 2016, 6:08 p.m. UTC | #1
On Mon, Oct 24, 2016 at 10:06:22PM +0800, Shawn Guo wrote:
> The following changes since commit 1001354ca34179f3db924eb66672442a173147dc:
> 
>   Linux 4.9-rc1 (2016-10-15 12:17:50 -0700)
> 
> are available in the git repository at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git tags/imx-fixes-4.9
> 
> for you to fetch changes up to 4edd601c5a9c5094daa714e65063e623826f3bcc:
> 
>   ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031 (2016-10-24 21:26:01 +0800)
> 
> ----------------------------------------------------------------
> The i.MX fixes for 4.9:
>  - A couple of patches from Fabio to fix the GPC power domain regression
>    which is caused by PM Domain core change 0159ec670763dd
>    ("PM / Domains: Verify the PM domain is present when adding a
>    provider"), and a related kernel crash seen with multi_v7_defconfig
>    build.
>  - Correct the PHY ID mask for AR8031 to match phy driver code.
>  - Apply new added timer erratum A008585 for LS1043A and LS2080A SoC.
>  - Correct vf610 global timer IRQ flag to avoid warning from gic driver
>    after commit 992345a58e0c ("irqchip/gic: WARN if setting the
>    interrupt type for a PPI fails").

Merged, thanks.


-Olof
Olof Johansson Oct. 31, 2016, 10:39 p.m. UTC | #2
Fabio,

I've got multi_v7 booting now, but imx_v6_v7 is still broken on both
hummingboard and wandboard.

Hummingboard doesn't even output anything on serial during boot, bug
with DEBUG_LL I got some more info

(No output boot at
http://arm-soc.lixom.net/bootlogs/mainline/v4.9-rc3/hummingboard-arm-imx_v6_v7_defconfig.html,
not very useful).

This one seems to be due to CONFIG_PCI_IMX6, which is off on
multi_v7_defconfig. Hard hang(?), last output is:

[    0.490643] OF: PCI: host bridge /soc/pcie@0x01000000 ranges:
[    0.496540] OF: PCI:   No bus range found for /soc/pcie@0x01000000,
using [bus 00-ff]
[    0.504512] OF: PCI:    IO 0x01f80000..0x01f8ffff -> 0x00000000
[    0.510559] OF: PCI:   MEM 0x01000000..0x01efffff -> 0x01000000



Wandboard, has a peculiar MMC error. First I thought it was just the
card going bad, but now I'm not so sure since multi_v7 boots just fine
on the same hardware:

http://arm-soc.lixom.net/bootlogs/mainline/v4.9-rc3/wandboard-arm-imx_v6_v7_defconfig.html

Not sure what's going on here, since bisecting between two configs is
a pain. When I disabled CONFIG_PCI, the error seems to be less likely
to happen (some boots succeed, sometimes it happens later during
boot).


-Olof

On Mon, Oct 24, 2016 at 7:06 AM, Shawn Guo <shawnguo@kernel.org> wrote:
> The following changes since commit 1001354ca34179f3db924eb66672442a173147dc:
>
>   Linux 4.9-rc1 (2016-10-15 12:17:50 -0700)
>
> are available in the git repository at:
>
>   git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git tags/imx-fixes-4.9
>
> for you to fetch changes up to 4edd601c5a9c5094daa714e65063e623826f3bcc:
>
>   ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031 (2016-10-24 21:26:01 +0800)
>
> ----------------------------------------------------------------
> The i.MX fixes for 4.9:
>  - A couple of patches from Fabio to fix the GPC power domain regression
>    which is caused by PM Domain core change 0159ec670763dd
>    ("PM / Domains: Verify the PM domain is present when adding a
>    provider"), and a related kernel crash seen with multi_v7_defconfig
>    build.
>  - Correct the PHY ID mask for AR8031 to match phy driver code.
>  - Apply new added timer erratum A008585 for LS1043A and LS2080A SoC.
>  - Correct vf610 global timer IRQ flag to avoid warning from gic driver
>    after commit 992345a58e0c ("irqchip/gic: WARN if setting the
>    interrupt type for a PPI fails").
>
> ----------------------------------------------------------------
> Fabio Estevam (3):
>       ARM: imx: gpc: Initialize all power domains
>       ARM: imx: gpc: Fix the imx_gpc_genpd_init() error path
>       ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031
>
> Scott Wood (1):
>       arm64: dts: Add timer erratum property for LS2080A and LS1043A
>
> Stefan Agner (1):
>       ARM: dts: vf610: fix IRQ flag of global timer
>
>  arch/arm/boot/dts/vf500.dtsi                   |  2 +-
>  arch/arm/mach-imx/gpc.c                        | 15 ++++++++++++---
>  arch/arm/mach-imx/mach-imx6q.c                 |  2 +-
>  arch/arm64/boot/dts/freescale/fsl-ls1043a.dtsi |  1 +
>  arch/arm64/boot/dts/freescale/fsl-ls2080a.dtsi |  1 +
>  5 files changed, 16 insertions(+), 5 deletions(-)
Fabio Estevam Oct. 31, 2016, 11:29 p.m. UTC | #3
Hi Olof,

On Mon, Oct 31, 2016 at 8:39 PM, Olof Johansson <olof@lixom.net> wrote:
> Fabio,
>
> I've got multi_v7 booting now, but imx_v6_v7 is still broken on both
> hummingboard and wandboard.
>
> Hummingboard doesn't even output anything on serial during boot, bug
> with DEBUG_LL I got some more info
>
> (No output boot at
> http://arm-soc.lixom.net/bootlogs/mainline/v4.9-rc3/hummingboard-arm-imx_v6_v7_defconfig.html,
> not very useful).
>
> This one seems to be due to CONFIG_PCI_IMX6, which is off on
> multi_v7_defconfig. Hard hang(?), last output is:
>
> [    0.490643] OF: PCI: host bridge /soc/pcie@0x01000000 ranges:
> [    0.496540] OF: PCI:   No bus range found for /soc/pcie@0x01000000,
> using [bus 00-ff]
> [    0.504512] OF: PCI:    IO 0x01f80000..0x01f8ffff -> 0x00000000
> [    0.510559] OF: PCI:   MEM 0x01000000..0x01efffff -> 0x01000000
>

I think this will be fixed with this pcie-designware patch:
http://www.spinics.net/lists/linux-pci/msg55244.html

> Wandboard, has a peculiar MMC error. First I thought it was just the
> card going bad, but now I'm not so sure since multi_v7 boots just fine
> on the same hardware:

Yes, I have been noticing this mmc error reported in your autobooter
and I also thought it was some kind of a bad SD card.

This particular error I have never been able to reproduce. Also
checked kernelci and do not see this mmc error there.

Interesting that it happens only with imx_v6_v7_defconfig and not with
multi_v7_defconfig.
Fabio Estevam Nov. 5, 2016, 12:02 p.m. UTC | #4
Hi Olof,

On Mon, Oct 31, 2016 at 9:29 PM, Fabio Estevam <festevam@gmail.com> wrote:

> I think this will be fixed with this pcie-designware patch:
> http://www.spinics.net/lists/linux-pci/msg55244.html

Bjorn sent this fix to Linus, so we should be fine now.

>
>> Wandboard, has a peculiar MMC error. First I thought it was just the
>> card going bad, but now I'm not so sure since multi_v7 boots just fine
>> on the same hardware:
>
> Yes, I have been noticing this mmc error reported in your autobooter
> and I also thought it was some kind of a bad SD card.
>
> This particular error I have never been able to reproduce. Also
> checked kernelci and do not see this mmc error there.
>
> Interesting that it happens only with imx_v6_v7_defconfig and not with
> multi_v7_defconfig.

Today wandboard does not show the mmc error with imx_v6_v7_defconfig:
http://arm-soc.lixom.net/bootlogs/mainline/v4.9-rc3-376-gfb415f2/wandboard-arm-imx_v6_v7_defconfig.html

Looks like the mmc error come and go?