On Fri, Aug 10, 2018 at 9:36 AM Roland Hieber wrote:
>
> Having three ints as parameter suggests that we can use up to 2^3
> power configurations for the system, but when we look at the code, the
> power setup is packaged in if {...} else if {...} else if {...} blocks,
> so setting more than one p
On Fri, Aug 10, 2018 at 9:35 AM Roland Hieber wrote:
>
> From: Sascha Hauer
>
> No need for the boards to manually add memory, we can do this
> automatically by reading back the memory size from the SDRAM
> controller.
>
> Signed-off-by: Sascha Hauer
> Signed-off-by: Roland Hieber
> ---
> .../
Signed-off-by: Andrey Smirnov
---
arch/arm/mach-imx/imx8mq.c | 6 +++---
arch/arm/mach-imx/include/mach/imx8mq.h | 11 ---
2 files changed, 11 insertions(+), 6 deletions(-)
diff --git a/arch/arm/mach-imx/imx8mq.c b/arch/arm/mach-imx/imx8mq.c
index bc463ee75..95839f31d 10064
Everyone:
This series is a number of fixes and changes I made for boot source
and reset reason detection on i.MX8MQ. Hopefully all of the changes
are self-explanatory.
Feedback is welcome!
Thanks,
Andrey Smirnov
Andrey Smirnov (6):
ARM: i.MX8M: Expose code to query cpu revision
ARM: i.MX: b
Reset reason bits and their meaning seem to be identical between i.MX7
and i.MX8MQ. Share the definitions for the former and used it for the
latter.
Signed-off-by: Andrey Smirnov
---
arch/arm/mach-imx/imx.c | 11 +++
arch/arm/mach-imx/imx7.c | 1
Signed-off-by: Andrey Smirnov
---
arch/arm/mach-imx/imx8mq.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/arch/arm/mach-imx/imx8mq.c b/arch/arm/mach-imx/imx8mq.c
index 1d0bbcfdb..4d00da5f0 100644
--- a/arch/arm/mach-imx/imx8mq.c
+++ b/arch/arm/mach-imx/imx8mq.c
@@ -54,6 +54,8 @@ int imx8
For both SoCs data found in SBMR registers reflects only the boot
source that was selected via pins of fuses and not the final boot
source that ended up being used by MaskROM code. Original i.MX7 boot
source detection implementation worked around that fact by having a
special code to correctly hand
Signed-off-by: Andrey Smirnov
---
arch/arm/mach-imx/boot.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm/mach-imx/boot.c b/arch/arm/mach-imx/boot.c
index ab25f75b2..45170ab10 100644
--- a/arch/arm/mach-imx/boot.c
+++ b/arch/arm/mach-imx/boot.c
@@ -512,7 +512,7 @@ vo
CPU revision information is needed for boot source detection, so
expose it as a small helper function and convert existing code to use
it.
Signed-off-by: Andrey Smirnov
---
arch/arm/mach-imx/imx8mq.c | 57 +++--
arch/arm/mach-imx/include/mach/imx8mq.h | 38 ++
Add code to parse "input-enable" and intepret it as SION bit. This is
needed for I2C configuration on i.MX8MQ EVK board.
Signed-off-by: Andrey Smirnov
---
drivers/pinctrl/imx-iomux-v3.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/pinctrl/imx-iomux-v3.c b/drivers/pinctrl/imx-io
Select DRIVER_SPI_IMX_2_3 for i.MX8MQ since that is the variant that's
being used
Signed-off-by: Andrey Smirnov
---
drivers/spi/Kconfig | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig
index 84945744e..fed628c58 100644
--- a/drivers/sp
Looks like that particular chunk of the code have not been built in a
while and bitrotted. Change function parameters appropriately to avoid
breaking the build.
Signed-off-by: Andrey Smirnov
---
drivers/i2c/busses/i2c-imx.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/
Together FEC driver and parts of IP stack might end up trying to
memcpy() small chunks of memory from uncached (that is Device memory)
addresses that are not properly aligned, leading to data abort.
To prevent such cases, add code to guard unaligned accesses and
redirect them to byte-wise implemen
Declare ar8031_phy_fixup as static since it is not being exported to
any other users.
---
arch/arm/boards/freescale-mx6sx-sabresdb/board.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm/boards/freescale-mx6sx-sabresdb/board.c
b/arch/arm/boards/freescale-mx6sx-sabresd
Allow to set not only the fractional divider, but also the prescaler for
the EMI clock in mxs_mem_init_clock(), and rename the parameters
accordingly to reflect the change. Port the existing board code to set
up the EMI clock explicitely with the old values. Also fix the
off-by-a-half error in the
Go more into detail about the relationship of the prepare stage and the
bootloader image, and about the generated images. Also don't use
mxs-usb-loader anymore since it has been removed.
Fixes: 3ca576cc4a4236a2556ab17 "scripts: imx/mxs remove mxs-usb-loader"
Signed-off-by: Roland Hieber
---
Docu
The cfg->bo_irq and cfg->bo_enirq (i.e. VDDx_BO_IRQ and ENIRQ_VDDx_BO)
flags are part of POWER_CTRL, so setting those flag in cfg->reg (i.e.
POWER_VDDxCTRL) makes no sense. Fortunately, those bits are reserved in
ENIRQ_VDDx_BO, so writing them did no harm, but also doesn't work as
intended.
Signed
The PINCTRL_*_DDR_MODE_* defines are now needed in global scope, so move
them to the respective include header.
Signed-off-by: Roland Hieber
---
arch/arm/boards/duckbill/lowlevel.c | 3 ++-
arch/arm/boards/freescale-mx28-evk/lowlevel.c | 3 ++-
arch/arm/boards/karo-tx28/lowlevel.c
Exemplary previous behaviour:
vddio: 3100mV (BO -200mV), Linreg enabled, Linreg offset: -1, FET enabled
vdda: 1800mV (BO -175mV), Linreg disabled, Linreg offset: -1, FET enabled
vddd: 1200mV (BO -175mV), Linreg enabled, Linreg offset: -1, FET disabled
vddmem: 1500mV (BO -25mV), Linreg
Having three ints as parameter suggests that we can use up to 2^3
power configurations for the system, but when we look at the code, the
power setup is packaged in if {...} else if {...} else if {...} blocks,
so setting more than one parameter to 1 is useless here.
Refactor the parameters into an
From: Sascha Hauer
We can detect the SDRAM automatically, no need to let the boards
do it manually.
This also fixes device tree based i.MX28 boards like the duckbill.
These boards once had a /memory node in the device tree, but now
that this was changed to /memory@4000 barebox didn't find it
From: Sascha Hauer
No need for the boards to manually add memory, we can do this
automatically by reading back the memory size from the SDRAM
controller.
Signed-off-by: Sascha Hauer
Signed-off-by: Roland Hieber
---
.../arm/boards/chumby_falconwing/falconwing.c | 8 -
arch/arm/boards/free
While porting barebox onto a new old MX28 board, I had to extend the
low-level code to set up SDRAM and PMIC. Most of those changes are
probably also applicable to MX23 boards since the code paths are the
same, and are only wrapped in mx28_*/mx23_*-specific wrapper functions.
(Also, if anyone want
Rename bo_int to prev_bo_enirq to make its purpose clearer, and throw in
a few comments for easier understanding.
Signed-off-by: Roland Hieber
---
arch/arm/mach-mxs/power-init.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/arch/arm/mach-mxs/power-init.c b/arch/a
Some boards don't need the 4.2V power source at all, so allow them to
keep it disabled for efficiency reasons.
For backwards compatibility, don't disable the 4P2 power source on the
existing boards which are supplied from battery input.
The POWER_USE_5V code path however always enables it up as i
From: Sascha Hauer
Normally MXS SoCs only allow to boot images which have been
encrypted with a zero key. If the ENABLE_UNENCTRYPTED_BOOT fuse
is blown then the SoC also allows to boot unencrypted images.
Add an option to the mxsimage tool to generate such images.
Signed-off-by: Sascha Hauer
Si
mxsimage can now build unencrypted images with -u, so make use of it in
the Makefile. To keep the existing rules simple, name the generated
images *.mxsbsu instead of *.mxsbs.
Signed-off-by: Roland Hieber
---
images/Makefile.mxs | 8
1 file changed, 8 insertions(+)
diff --git a/images/
... and in that process fix the erroneous usage of VDDA values when
printing the stats for POWER_VDDDCTRL_DISABLE_FET and all of the
*LINREG_OFFSET bitfields.
Signed-off-by: Roland Hieber
---
arch/arm/mach-mxs/power-init.c | 37 ++
1 file changed, 20 insertions(+)
Depending on the application, board code might want to configure
different power levels and brownout thresholds. They can do so now with
an additional struct mxs_power_ctrls* parameter to mx2*_power_init.
Note that VDDMEM was not set up explicitely in mx28_power_init, but can
now be configured. Ho
This is now recommended for all systems. Especially for the HAB usecase
disallowing alternative boot selection after WDOG reset is essential.
Signed-off-by: Bastian Stender
---
Changes since v1:
- removed accidentally added empty line
---
drivers/hab/hab.c | 6 ++
1 file changed, 6 inserti
Signed-off-by: Bastian Stender
---
arch/arm/mach-imx/include/mach/ocotp-fusemap.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/arch/arm/mach-imx/include/mach/ocotp-fusemap.h
b/arch/arm/mach-imx/include/mach/ocotp-fusemap.h
index 44b58ca6e8..aec50dbf8a 100644
--- a/arch/arm/mach-imx/
This is now recommended for all systems. Especially for the HAB usecase
disallowing alternative boot selection after WDOG reset is essential.
Signed-off-by: Bastian Stender
---
drivers/hab/hab.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a/drivers/hab/hab.c b/drivers/hab/hab.c
inde
Hello Leigh,
On Thu, Aug 09, 2018 at 04:14:48PM +0100, Leigh Brown wrote:
> On 2018-08-06 19:36, Uwe Kleine-König wrote:
> > Hello Leigh,
> >
> > On Tue, Jul 31, 2018 at 02:21:04PM +0100, Leigh Brown wrote:
> > > Have I missed any steps? Any help would be greatly appreciated.
> >
> > Looks good
On Fri, Aug 10, 2018 at 08:40:31AM +0200, Sascha Hauer wrote:
> Hi Sam,
>
> Some comments inside.
snip - many review good comments.
Thanks!
Will address all comments and submit a v3.
Sam
___
barebox mailing list
barebox@lists.infradead.org
ht
On Fri, Aug 10, 2018 at 08:22:47AM +0200, Sascha Hauer wrote:
> On Thu, Aug 09, 2018 at 10:33:12AM +0200, Uwe Kleine-König wrote:
> > This makes the following do the expected thing:
> >
> > barebox@barebox sandbox:/ ls -l lala
> > -rwxrwxrwx 4 lala
> > barebox@barebox sand
35 matches
Mail list logo