On Wed, 12 Jun 2024 15:36:53 +0200, Ahmad Fatoum wrote:
> The usual use of barebox state is to store the bootstate for
> bootchooser. Thus change the example to use the minimum size of the
> bootstate variable set instead and relate the result to stride size and
> partition size.
>
>
Applied,
On Tue, 11 Jun 2024 10:27:14 +0200, Ahmad Fatoum wrote:
> A return code of -EPROBE_DEFER is not necessarily an error as the clock
> driver may indeed be probed later. Use dev_err_probe to silence the
> error message in this situation and to store it for later use if the
> probe is deferred perman
On Wed, 12 Jun 2024 15:05:08 +0200, Oleksij Rempel wrote:
>
Applied, thanks!
[1/2] arm: boards: Add support for Protonic MECIO1 board
https://git.pengutronix.de/cgit/barebox/commit/?id=979f8397a217 (link may
not be stable)
[2/2] arm: boards: Add support for Protonic MECT1S board
On Wed, 12 Jun 2024 15:02:56 +0200, Oleksij Rempel wrote:
> In case this PHY is bootstrapped for managed mode, we need to manually
> wake it. Otherwise no link will be detected.
>
>
Applied, thanks!
[1/1] net: phy: dp83tg720: wake up PHYs in managed mode
https://git.pengutronix.de/cgit/
On Tue, 11 Jun 2024 10:28:29 +0200, Ahmad Fatoum wrote:
> These address are not used and probably won't be soon, because they are
> either probed from DT or handled in the secure world.
>
> Given that they are needlessly generic (STM32_ instead of STM32MP1(5)),
> let's drop them.
>
>
> [...]
On Tue, 11 Jun 2024 08:59:17 +0200, Ahmad Fatoum wrote:
> So far, we limited semihosting support to 32-bit ARM, but it's also
> usable on ARM64 and even RISC-V.
>
> This series does the first step by making it compile for ARM64 and also
> adds DEBUG_LL and console drivers.
>
> Ahmad Fatoum (6):
On Tue, 11 Jun 2024 08:58:10 +0200, Ahmad Fatoum wrote:
> The watchdog peripheral on the Raspberry Pi can be used to shutdown by
> writing a magic value that the firmware checks for and rebooting.
>
> This is useful to close an emulated Raspberry Pi in QEMU, so add
> support.
>
>
> [...]
Appl
On Tue, 11 Jun 2024 08:58:00 +0200, Ahmad Fatoum wrote:
> Unlike clk_get, clk_get_for_console returns NULL when the clock provider
> lacks a driver, but the UART has already been configured by DEBUG_LL.
>
> This allows drivers to disable the set baudrate callback and continue
> probe, so the use
On Tue, 11 Jun 2024 08:57:15 +0200, Ahmad Fatoum wrote:
> In order not to clutter configs, we should hide symbols that are EFI
> related when CONFIG_EFI is disabled, especially as we are going to add
> more options with the loader support.
>
>
Applied, thanks!
[1/4] efi: hide EFI related symb
This is in preparation of buffering fit_open() calls to not load the
same fit twice if it is still open.
Signed-off-by: Marco Felsch
---
common/image-fit.c | 4
include/image-fit.h | 1 +
2 files changed, 5 insertions(+)
diff --git a/common/image-fit.c b/common/image-fit.c
index e93946a06
The FIT spec is not very specific when it comes to device-tree overlay
handling. Overlays can be added directely to an config node:
config-a {
compatible = "machine-compatible";
kernel = "kernel-img-name";
fdt = "fdt-base-name", "fdt-overlay1
The upcoming FIT overlay support requires these helpers for performing
check on a fit image.
Signed-off-by: Marco Felsch
---
common/image-fit.c | 4 ++--
include/image-fit.h | 2 ++
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/common/image-fit.c b/common/image-fit.c
index 7133
The current overlay mechanism can handle files only, so filepattern was
obvious. With the next commit this will change, so overlays can also
supplied via FIT images. To prepare the pattern filter to match FIT
config-nodes make the filter and global variable handling the pattern
more generic by drop
Add the a parameter which will be used by image-fit to decide if a
config node belongs to a devicetree-overlay.
Signed-off-by: Marco Felsch
---
drivers/of/overlay.c | 4
1 file changed, 4 insertions(+)
diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
index 73c7a91db9b5..20e12211d63
This is now done during of_overlay_global_fixup() so we can drop it
here.
Signed-off-by: Marco Felsch
---
drivers/of/overlay.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
index 941cd3d51883..dad4e05ca27f 100644
--- a/drivers/of/overlay.c
+++
Hi,
this series add the support to load FIT image supplied devicetree overlays.
The v1 of this series can be found here [1].
The overlay loading wasn't coupled to bootm due to the following
reasons:
- By making use of the common overlay handling we can specifiy a
different/separate FIT image
In preparation of adding support for FIT image overlay handling the
global.of.overlay.dir can be reused if we rename it to a more generic
global.of.overlay.path since it does not imply an directory.
This commit adds the ground work by deprecating the old
global.of.overlay.dir and moving the direct
This adds the support to load devicetree overlays from an FIT image.
There are quite a few options to handle FIT overlays since the FIT
overlay spec is not very strict.
This patch implement the most configurable case where each overlay does
have it's own config node (including the optional signatu
Cache the FIT image fit_open() calls to avoid loading the same FIT image
twice. This is very useful if the same FIT image is used to provide the
base devicetree, kernel and initrd as well as devicetree overlays.
Signed-off-by: Marco Felsch
---
common/image-fit.c | 38 +++
Currently we don't honor the setup done by the BootROM. Fix this by
reading the scr_el3 first and update the corresponding bits afterwards.
Furthermore also align the register settings:
- Remove the register with (RW) setup since this is done by the EL3
firmware (TF-A).
- Set IRQ/FIQ/EA bits
The i.MX8M* devices do have an sticky bit which indicates if the
field-return fuse can be written. Add support to query the lock bit.
To make it easy to read align the STICKY bit definitions as well.
Signed-off-by: Marco Felsch
---
drivers/nvmem/ocotp.c| 24
include
All current known i.MX8M devices: i.MX8MQ, i.MX8MP, i.MX8MN, i.MX8MM use
an other fuse to lock the SRK hash. Fix this by refactoring
imx_hab_write_srk_hash_ocotp() and make the lock fusing device specific.
Fixes: 6c4d5bb5acfe ("i.MX: HABv4: implement interface for i.MX8MQ")
Signed-off-by: Marco Fe
Add an helper to revoke an i.MX SRK key. At the moment the helper
supprts i.MX8M devices only, but adding support for other SoCs can be
done easily by providing the .revoke_key() hook.
Signed-off-by: Marco Felsch
---
drivers/hab/hab.c| 45
include
Add more common i.MX8M fuse defines so they can be reused by board code.
Signed-off-by: Marco Felsch
---
include/mach/imx/ocotp-fusemap.h | 10 ++
1 file changed, 10 insertions(+)
diff --git a/include/mach/imx/ocotp-fusemap.h b/include/mach/imx/ocotp-fusemap.h
index e6fa96b40ba9..1aece9
This ports U-Boot commit:
| commit 2f3c92060dcd6bc9cfd3e2e344a3e1745ca39f09
| Author: Peng Fan
| Date: Thu Jul 9 13:39:26 2020 +0800
|
| imx8m: workaround ROM serror
|
| ROM SError happens on two cases:
|
| 1. ERR050342, on iMX8MQ HDCP enabled parts ROM writes to GPV1 register,
but
Right now the core implementation always make use of DEVICE_ID_DYNAMIC.
This does not allow us having static devices supplied via the of-aliases
node. Therefore sync the code base with Linux to allow single ids,
albeit the id handling is still different.
While on it fix the nvmem_register_cdev() b
Expose the nvmem cells via cdevs which is our equivalent to the Linux
sysfs exposure. This allows the easier user queries for board code and
shell. Keep the Linux function name scheme for
nvmem_populate_sysfs_cells() to reduce the diff for nvmem_register()
function.
Signed-off-by: Marco Felsch
--
Re-sync our code base with Linux which moved quite a lot. This patch is
huge but still is only a partly sync. The main changes are:
- The nvmem_cell struct was splitted into nvmem_cell and
nvmem_cell_entry
- The cells are now parsed and registered during nvmem_register(), no
longer duri
Make use of the BIT() macro to define the flags, no functional change.
Signed-off-by: Marco Felsch
---
include/hab.h | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/include/hab.h b/include/hab.h
index da79a8ffea36..2cef3841d149 100644
--- a/include/hab.h
+++ b/include/
This adds the required Kconfig options which need to be enabled and
correctly set to build a custom device specific barebox image which can
be used to burn the FIELD_RETURN fuse.
The CST tool can't handle quoted UID strings so we need to define it on
the cmdline by using the -D switch. This remove
The HAB code needs an special [Unlock] instruction to keep the
SRK_REVOKE fuse bank unlocked. This is required if a key needs to be
revoked.
Signed-off-by: Marco Felsch
---
arch/arm/mach-imx/Kconfig| 8
include/mach/imx/habv4-imx8-gencsf.h | 6 ++
2 files changed, 14 ins
Add a convenient common helper to burn the field-return fuse which wraps
the platform specific hook. At the moment only i.MX8M devices are
supported. Adding support for other platforms can be done by providing
the platform specific hook.
Signed-off-by: Marco Felsch
---
drivers/hab/hab.c
The i.MX8M* devices do have an sticky bit which indicates if the
srk_revoke fuse can be written. Add support to query and to set the lock
bit.
Signed-off-by: Marco Felsch
---
drivers/nvmem/ocotp.c| 55
include/mach/imx/ocotp.h | 2 ++
2 files changed
Currently the nvmem memory is freed without removing/unregister the
nvmem->dev if nvmem_register_cdev() fails. Fix this by unregister the
device first, free the memory and return the error code.
Signed-off-by: Marco Felsch
---
v2:
- new patch
drivers/nvmem/core.c | 12
1 file chang
The at24 driver does not need to handle the ids by its own since the
nvmem_core does the handling too. This lead into issues where the eeprom
device is named: eeprom00.
Fix the alias handling too since the devie would never be named as
described in the alias, since we never told the nvmem-core to
Signed-off-by: Jules Maselbas
---
drivers/mci/mci-core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/mci/mci-core.c b/drivers/mci/mci-core.c
index b4139be11e..de1efe3bb9 100644
--- a/drivers/mci/mci-core.c
+++ b/drivers/mci/mci-core.c
@@ -1619,7 +1619,7 @@ stat
Free the handle if read_file_2() fails.
Signed-off-by: Marco Felsch
---
common/image-fit.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/common/image-fit.c b/common/image-fit.c
index 251fda97b3fc..008804e6a6c3 100644
--- a/common/image-fit.c
+++ b/common/image-fit.c
@@ -922,6 +922,7 @@ str
Since we do support FIT image overlays and file/dir based overlays the
filename variable is not sufficient anylonger. Therefore rename the
local variables as well as the filter function to hook.
To not break any systems keep the old filename based name too but mark
them as deprecated.
Signed-off-
The MMC version is read from the reserved bits 125:122 in the CSD register,
theses bits are specified as SPEC_VERS by JEDEC in the eMMC standard JESD84.
Signed-off-by: Jules Maselbas
---
drivers/mci/mci-core.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/mci/
Signed-off-by: Jules Maselbas
---
drivers/mci/mci-core.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/mci/mci-core.c b/drivers/mci/mci-core.c
index 5a8f81ecb8..0afb0d26c8 100644
--- a/drivers/mci/mci-core.c
+++ b/drivers/mci/mci-core.c
@@ -1070,8 +1070,12 @@ st
On Thu, 13 Jun 2024 17:27:24 +0200, Jules Maselbas wrote:
>
Applied, thanks!
[1/1] mci: core: clarify why write_bl_len is the same as read_bl_len
https://git.pengutronix.de/cgit/barebox/commit/?id=89bb083f5a63 (link may
not be stable)
Best regards,
--
Sascha Hauer
On Thu, 13 Jun 2024 15:53:38 +0200, Jules Maselbas wrote:
> The MMC version is read from the reserved bits 125:122 in the CSD register,
> theses bits are specified as SPEC_VERS by JEDEC in the eMMC standard JESD84.
>
>
Applied, thanks!
[1/1] mci: core: clarify comment on MMC version detection
On Thu, 13 Jun 2024 15:28:20 +0200, Jules Maselbas wrote:
>
Applied, thanks!
[1/1] mci: core: fix typos
https://git.pengutronix.de/cgit/barebox/commit/?id=cd74c163f8af (link may
not be stable)
Best regards,
--
Sascha Hauer
43 matches
Mail list logo