| Message ID | 1415096927-4097-1-git-send-email-festevam@gmail.com |
|---|---|
| State | New |
| Headers | show |
On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: > From: Fabio Estevam <fabio.estevam@freescale.com> > > Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit > 0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI > selection was dropped by savedefconfig for imx_v4_v5_defconfig and > imx_v6_v7_defconfig. > > In order to keep the same behaviour as previous kernel versions and avoid > regressions, let's add CONFIG_SPI option back. > > Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Applied, thanks.
Shawn, On Tue, Nov 4, 2014 at 9:07 AM, Shawn Guo <shawn.guo@linaro.org> wrote: > On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: >> From: Fabio Estevam <fabio.estevam@freescale.com> >> >> Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit >> 0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI >> selection was dropped by savedefconfig for imx_v4_v5_defconfig and >> imx_v6_v7_defconfig. >> >> In order to keep the same behaviour as previous kernel versions and avoid >> regressions, let's add CONFIG_SPI option back. >> >> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> > > Applied, thanks. I have just sent a v3, with an explanation as to why savedefconfig dropped CONFIG_SPI=y. Thanks to Uwe for his analysis!
[Note, this is a resend of a mail I accidently only sent to Fabio before.] Hello, On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: > From: Fabio Estevam <fabio.estevam@freescale.com> > > Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit > 0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI > selection was dropped by savedefconfig for imx_v4_v5_defconfig and > imx_v6_v7_defconfig. > > In order to keep the same behaviour as previous kernel versions and avoid > regressions, let's add CONFIG_SPI option back. > > Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> > --- > Changes since v1: > - Add commit id's that caused the issue. on top of 64546e9fe3a5b8c doing make imx_v6_v7_defconfig resulted in CONFIG_SPI=y. That's because MEDIA_SUBDRV_AUTOSELECT selects SPI and has default y. So the commit that really broke imx_v6_v7_defconfig is 7d24c514e8e0 ([media] Kconfig: do not select SPI bus on sub-driver auto-select). On a side note I wonder how that commit sneaked into v3.17-rc6. I think people need to be more aware of the results of Kconfig changes ... Best regards Uwe
Moikka! On 11/04/2014 05:26 PM, Uwe Kleine-König wrote: > [Note, this is a resend of a mail I accidently only sent to Fabio > before.] > > Hello, > > On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: >> From: Fabio Estevam <fabio.estevam@freescale.com> >> >> Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit >> 0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI >> selection was dropped by savedefconfig for imx_v4_v5_defconfig and >> imx_v6_v7_defconfig. >> >> In order to keep the same behaviour as previous kernel versions and avoid >> regressions, let's add CONFIG_SPI option back. >> >> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> >> --- >> Changes since v1: >> - Add commit id's that caused the issue. > on top of 64546e9fe3a5b8c doing > > make imx_v6_v7_defconfig > > resulted in CONFIG_SPI=y. That's because MEDIA_SUBDRV_AUTOSELECT selects > SPI and has default y. > > So the commit that really broke imx_v6_v7_defconfig is 7d24c514e8e0 > ([media] Kconfig: do not select SPI bus on sub-driver auto-select). > On a side note I wonder how that commit sneaked into v3.17-rc6. > > I think people need to be more aware of the results of Kconfig changes I added that SPI select patch for 3.17-rc1 and it was removed right after I got bug report it should not be there. But there was of course some delay and removal patch went to 3.17-rc6. But all in all, it was all happening during 3.17 period and I wonder how it could affect you? Here are related discussions: [GIT PULL for 3.17-rc1] media updates for next https://lkml.org/lkml/2014/8/5/646 Autoselecting SPI for MEDIA_SUBDRV_AUTOSELECT? https://lkml.org/lkml/2014/8/22/325 [PATCH] Kconfig: do not select SPI bus on sub-driver auto-select https://lkml.org/lkml/2014/8/22/452 [GIT PULL 3.17] FIX: remove SPI select from Kconfig http://permalink.gmane.org/gmane.linux.drivers.video-input-infrastructure/81346 [GIT PULL for v3.17-rc6] media fixes https://lkml.org/lkml/2014/9/21/44 regards Antti
Hello, On Tue, Nov 04, 2014 at 05:55:31PM +0200, Antti Palosaari wrote: > On 11/04/2014 05:26 PM, Uwe Kleine-König wrote: > >[Note, this is a resend of a mail I accidently only sent to Fabio > >before.] > > > >Hello, > > > >On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: > >>From: Fabio Estevam <fabio.estevam@freescale.com> > >> > >>Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit > >>0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI > >>selection was dropped by savedefconfig for imx_v4_v5_defconfig and > >>imx_v6_v7_defconfig. > >> > >>In order to keep the same behaviour as previous kernel versions and avoid > >>regressions, let's add CONFIG_SPI option back. > >> > >>Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> > >>--- > >>Changes since v1: > >>- Add commit id's that caused the issue. > >on top of 64546e9fe3a5b8c doing > > > > make imx_v6_v7_defconfig > > > >resulted in CONFIG_SPI=y. That's because MEDIA_SUBDRV_AUTOSELECT selects > >SPI and has default y. > > > >So the commit that really broke imx_v6_v7_defconfig is 7d24c514e8e0 > >([media] Kconfig: do not select SPI bus on sub-driver auto-select). > >On a side note I wonder how that commit sneaked into v3.17-rc6. > > > >I think people need to be more aware of the results of Kconfig changes > > I added that SPI select patch for 3.17-rc1 and it was removed right > after I got bug report it should not be there. But there was of > course some delay and removal patch went to 3.17-rc6. Still I think that -rc6 is quite late for such a patch. But obviously I'm not the relevant maintainer here and YMMV. > But all in all, it was all happening during 3.17 period and I wonder > how it could affect you? Commit 64546e9fe3a5 was based on v3.17-rc3 and so had the "select SPI". That's why savedefconfig removed CONFIG_SPI=y from the minimized config. Dropping the select for -rc6 resulted in imx_v6_v7_defconfig not including CONFIG_SPI any more. I checked all commits listed by: git log --oneline linus/master ^e4462ffc1602 -SCONFIG_SPI and the two imx defconfigs are the only ones being affected by this issue. Best regards Uwe
Moikka, On 11/04/2014 09:18 PM, Uwe Kleine-König wrote: > Hello, > > On Tue, Nov 04, 2014 at 05:55:31PM +0200, Antti Palosaari wrote: >> On 11/04/2014 05:26 PM, Uwe Kleine-König wrote: >>> [Note, this is a resend of a mail I accidently only sent to Fabio >>> before.] >>> >>> Hello, >>> >>> On Tue, Nov 04, 2014 at 08:28:47AM -0200, Fabio Estevam wrote: >>>> From: Fabio Estevam <fabio.estevam@freescale.com> >>>> >>>> Since 64546e9fe3a5b8c ("ARM: imx_v6_v7_defconfig updates") and commit >>>> 0650f855d2e4b0b9 ("ARM: imx_v4_v5_defconfig: Select CONFIG_IMX_WEIM") CONFIG_SPI >>>> selection was dropped by savedefconfig for imx_v4_v5_defconfig and >>>> imx_v6_v7_defconfig. >>>> >>>> In order to keep the same behaviour as previous kernel versions and avoid >>>> regressions, let's add CONFIG_SPI option back. >>>> >>>> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> >>>> --- >>>> Changes since v1: >>>> - Add commit id's that caused the issue. >>> on top of 64546e9fe3a5b8c doing >>> >>> make imx_v6_v7_defconfig >>> >>> resulted in CONFIG_SPI=y. That's because MEDIA_SUBDRV_AUTOSELECT selects >>> SPI and has default y. >>> >>> So the commit that really broke imx_v6_v7_defconfig is 7d24c514e8e0 >>> ([media] Kconfig: do not select SPI bus on sub-driver auto-select). >>> On a side note I wonder how that commit sneaked into v3.17-rc6. >>> >>> I think people need to be more aware of the results of Kconfig changes >> >> I added that SPI select patch for 3.17-rc1 and it was removed right >> after I got bug report it should not be there. But there was of >> course some delay and removal patch went to 3.17-rc6. > Still I think that -rc6 is quite late for such a patch. But obviously > I'm not the relevant maintainer here and YMMV. Yes, it is late. I send PULL request last day of rc1, but media fixes are not send for every rc, instead for rc1 and 1-2 times after that, usually very near end of merge window. > >> But all in all, it was all happening during 3.17 period and I wonder >> how it could affect you? > Commit 64546e9fe3a5 was based on v3.17-rc3 and so had the "select SPI". > That's why savedefconfig removed CONFIG_SPI=y from the minimized config. > > Dropping the select for -rc6 resulted in imx_v6_v7_defconfig not > including CONFIG_SPI any more. > > I checked all commits listed by: > > git log --oneline linus/master ^e4462ffc1602 -SCONFIG_SPI > > and the two imx defconfigs are the only ones being affected by this > issue. > > Best regards > Uwe > regards Antti
diff --git a/arch/arm/configs/imx_v4_v5_defconfig b/arch/arm/configs/imx_v4_v5_defconfig index e688741..e6b0007 100644 --- a/arch/arm/configs/imx_v4_v5_defconfig +++ b/arch/arm/configs/imx_v4_v5_defconfig @@ -97,6 +97,7 @@ CONFIG_SERIAL_IMX_CONSOLE=y # CONFIG_HW_RANDOM is not set CONFIG_I2C_CHARDEV=y CONFIG_I2C_IMX=y +CONFIG_SPI=y CONFIG_SPI_IMX=y CONFIG_SPI_SPIDEV=y CONFIG_GPIO_SYSFS=y diff --git a/arch/arm/configs/imx_v6_v7_defconfig b/arch/arm/configs/imx_v6_v7_defconfig index 8fca6e2..6790f1b 100644 --- a/arch/arm/configs/imx_v6_v7_defconfig +++ b/arch/arm/configs/imx_v6_v7_defconfig @@ -158,6 +158,7 @@ CONFIG_I2C_CHARDEV=y CONFIG_I2C_ALGOPCF=m CONFIG_I2C_ALGOPCA=m CONFIG_I2C_IMX=y +CONFIG_SPI=y CONFIG_SPI_IMX=y CONFIG_GPIO_SYSFS=y CONFIG_GPIO_MC9S08DZ60=y