On Tue, 05 Aug 2025 19:45:33 +0200, Ahmad Fatoum wrote:
> The loop remapping the memory banks looks at reserved memory
> regions and then maps everything eXecute Never up to the start of the
> region. If the region happens to be in the same bank as the text area
> and it comes after it, this mean
On Tue, 05 Aug 2025 20:53:44 +0200, Ahmad Fatoum wrote:
> A TLV blob may contain more MAC addresses than barebox can assign for
> the network devices it knows about. In that case, barebox should assign
> as many MAC addresses as it can and fix up the remainder, so the OS can
> deal with it.
>
>
On Wed, Aug 06, 2025 at 08:19:14AM +0200, Sascha Hauer wrote:
> Supply qemu options only for qemu targets. Otherwise
>
> labgrid-pytest --lg-env test/sandbox/sandbox_defconfig.yaml
>
> fails with:
>
> _pytest.outcomes.Exit: Qemu option supplied for non-Qemu targe
>
> Signed-off-by: Sascha Hauer
Supply qemu options only for qemu targets. Otherwise
labgrid-pytest --lg-env test/sandbox/sandbox_defconfig.yaml
fails with:
_pytest.outcomes.Exit: Qemu option supplied for non-Qemu targe
Signed-off-by: Sascha Hauer
---
conftest.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff -
On Tue, Aug 05, 2025 at 10:07:54PM +0200, Marco Felsch wrote:
> On 25-08-05, Sascha Hauer wrote:
> > +Cc Basti
> >
> > On Thu, Jul 24, 2025 at 03:50:56PM +0200, Marco Felsch wrote:
> > > Currently the env provided keyspec is resolved during the gen_key() step
> > > by making use of the try_resolve
On Tue, Aug 05, 2025 at 10:34:35PM +0200, Marco Felsch wrote:
> On 24-07-15, Sascha Hauer wrote:
> > On Mon, Jul 15, 2024 at 01:30:00PM +0200, Marco Felsch wrote:
> > > Hi Sascha,
> > >
> > > On 24-07-15, Sascha Hauer wrote:
> > > > On Wed, Jul 03, 2024 at 06:58:34PM +0200, Marco Felsch wrote:
> >
On Tue, 05 Aug 2025 09:25:37 +0200, Sascha Hauer wrote:
> devfs-core.c implements the partition handling only for devices that are
> not MTD devices. MTD devices have their own partition handling in
> drivers/mtd/partition.c, see the special case in __devfs_add_partition()
> for MTD devices. This
On Tue, 05 Aug 2025 16:38:48 +0200, Stefan Kerkmann wrote:
> CONFIG_USB_ROLE_SWITCH is only defined in the kernel and is always
> undefined in barebox.
>
>
Applied, thanks!
[1/1] usb: dwc3: remove CONFIG_USB_ROLE_SWITCH in condition
https://git.pengutronix.de/cgit/barebox/commit/?id=fcd
On Tue, 05 Aug 2025 15:45:29 +0200, Stefano Manni wrote:
>
Applied, thanks!
[1/1] mtd: nand: add manufacturer ISSI
https://git.pengutronix.de/cgit/barebox/commit/?id=16c355edf21d (link may
not be stable)
Best regards,
--
Sascha Hauer
On 24-07-15, Sascha Hauer wrote:
> On Mon, Jul 15, 2024 at 01:30:00PM +0200, Marco Felsch wrote:
> > Hi Sascha,
> >
> > On 24-07-15, Sascha Hauer wrote:
> > > On Wed, Jul 03, 2024 at 06:58:34PM +0200, Marco Felsch wrote:
> > > > This adds the support to load devicetree overlays from an FIT image.
On 25-08-05, Sascha Hauer wrote:
> On Tue, Jul 29, 2025 at 05:34:32PM +0200, Marco Felsch wrote:
> > This adds the support to load devicetree overlays from a FIT image.
> > There are a few options to handle FIT overlays since the FIT overlay
> > spec is not very strict.
> >
> > This implements the
On 25-08-05, Sascha Hauer wrote:
> +Cc Basti
>
> On Thu, Jul 24, 2025 at 03:50:56PM +0200, Marco Felsch wrote:
> > Currently the env provided keyspec is resolved during the gen_key() step
> > by making use of the try_resolve_env(). This is wrong because it will
> > set the complete : keyspec strin
A TLV blob may contain more MAC addresses than barebox can assign for
the network devices it knows about. In that case, barebox should assign
as many MAC addresses as it can and fix up the remainder, so the OS can
deal with it.
This got broken during refactoring in reply to upstreaming feedback:
i
In preparation for moving the remapping of memory banks into the common
code, rename vectors_init to setup_trap_pages and export it for both
32-bit and 64-bit ARM, so it can be called from the common code as well.
This needs to happen after the remapping, because otherwise the trap
pages would be
The code is identical between ARM32 and 64 and is going to get more
complex with the addition of finer grained MMU permissions.
Let's move it to a common code file in anticipation.
Signed-off-by: Ahmad Fatoum
---
arch/arm/cpu/mmu-common.c | 46 +++
arch/arm/c
The loop in mmu_remap_memory_banks first looks at reserved memory
regions and then maps everything eXecute Never up to the start of the
region. If the region happens to be in the same bank as the text area
and it comes after it, this means the text area is temporarily mapped
eXecute Never, while ba
Most ranges in barebox identified by start and end offsets are exclusive
at the end, for example, the linker symbols for the sections.
To make handling them a bit easier while avoiding off-by-one, let's add
a region_overlap_end_exclusive helper as well.
Unlike inclusive ranges, exclusive ranges c
The order matters on ARM32, because arm_fixup_vectors() actually rewrite
the vector table, which is in the text area.
On ARM64, the order doesn't matter. As we are going to make the memory
bank remapping code common between both ARM32 and ARM64, move
setup_trap_pages, so the code between ARM32 and
While quite verbose, off-by-one are very annoying, so it makes sense to
be very explicit about the expected input. Rename the function in
preparation for adding region_overlap_end_exclusive.
Signed-off-by: Ahmad Fatoum
---
commands/iomemport.c | 2 +-
common/partitions.c | 6 +++---
include/r
early_remap_range() is calling get_pre_attrs() internally already on the
flags argument, so passing PTE flags here is incorrect.
Fixes: 59c1288698b4 ("ARM: MMU64: map memory for barebox proper pagewise")
Signed-off-by: Ahmad Fatoum
---
arch/arm/cpu/mmu_64.c | 4 ++--
1 file changed, 2 insertions
The loop that remaps memory banks can end up calling remap_range with
zero size, when a reserved region is at the very start of the memory
bank.
This is handled correctly by the code, but does an unnecessary
invalidation of the whole TLB. Let's early exit instead to skip that.
Signed-off-by: Ahma
The loop remapping the memory banks looks at reserved memory
regions and then maps everything eXecute Never up to the start of the
region. If the region happens to be in the same bank as the text area
and it comes after it, this means the text area is temporarily mapped
eXecute Never, while barebox
role-switch")) &&
> !DWC3_VER_IS_PRIOR(DWC3, 330A))
> mode = USB_DR_MODE_PERIPHERAL;
> }
>
> ---
> base-commit: 7be8b1bcb8aeb249e9d55f70ccb5ea93711e8efa
> change-id: 20250805-fix-dwc3-dr-mode-99b4893c2531
>
> Be
d_bool(dwc->dev->of_node,
"usb-role-switch")) &&
!DWC3_VER_IS_PRIOR(DWC3, 330A))
mode = USB_DR_MODE_PERIPHERAL;
}
---
base-commit: 7be8b1bcb8aeb249e9d55f70ccb5ea93711e8efa
change-id: 20250805-fix-dwc3-dr-mode-99b4893c2531
Best regards,
--
Stefan Kerkmann
Signed-off-by: Stefano Manni
---
drivers/mtd/nand/raw/internals.h | 1 +
drivers/mtd/nand/raw/nand_ids.c | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/mtd/nand/raw/internals.h b/drivers/mtd/nand/raw/internals.h
index 191244801b..07129af7a7 100644
--- a/drivers/mtd/nand/raw/intern
On mar, ago 5 2025 at 09:25:37 +02:00:00, Sascha Hauer
wrote:
devfs-core.c implements the partition handling only for devices that
are
not MTD devices. MTD devices have their own partition handling in
drivers/mtd/partition.c, see the special case in
__devfs_add_partition()
for MTD devices.
On Tue, Jul 29, 2025 at 05:34:32PM +0200, Marco Felsch wrote:
> This adds the support to load devicetree overlays from a FIT image.
> There are a few options to handle FIT overlays since the FIT overlay
> spec is not very strict.
>
> This implements the most configurable case where each overlay do
Hi Marco,
On Mon, Aug 04, 2025 at 04:36:46PM +0200, Marco Felsch wrote:
> Hi,
>
> with this patchset the barebox nvmem-core is prepared for the Linux
> nvmem-layout drivers. Layout drivers are used to describe the NVMEM
> storage format. The patchset also adds the support to read nvmem-cells
> vi
Hi Ahmad,
On Mon, Aug 04, 2025 at 07:22:20PM +0200, Ahmad Fatoum wrote:
> This series replaces 7 patches that are in next to fix a barebox hang
> when used together with OP-TEE.
Unfortunately I haven't seen this in time, so the offending patches are
in master now. Could you resend your changes as
On Mon, 07 Jul 2025 13:22:42 +0200, Steffen Trumtrar wrote:
> This series adds basic support for the Altera Agilex5 SoCFPGA series.
> Currently only booting from QSPI is supported. Flashing the image to the
> QSPI has to be done with quartus_pgm, therefore there are still some
> essential things
On Mon, 23 Jun 2025 15:57:45 +0200, Steffen Trumtrar wrote:
> This series adds basic support for the Altera Agilex5 SoCFPGA series.
> Currently only booting from QSPI is supported. Flashing the image to the
> QSPI has to be done with quartus_pgm, therefore there are still some
> essential things
On Tue, 29 Oct 2024 09:42:30 +0100, Steffen Trumtrar wrote:
> This series adds basic support for the Altera Agilex5 SoCFPGA series.
> Currently only booting from QSPI is supported. Flashing the image to the
> QSPI has to be done with quartus_pgm, therefore there are still some
> essential things
On Tue, 18 Feb 2025 10:21:25 +0100, Steffen Trumtrar wrote:
> This series adds basic support for the Altera Agilex5 SoCFPGA series.
> Currently only booting from QSPI is supported. Flashing the image to the
> QSPI has to be done with quartus_pgm, therefore there are still some
> essential things
On Mon, Jul 07, 2025 at 05:53:30AM +, Gwenhael Goavec-Merou wrote:
> From: Gwenhael Goavec-Merou
>
> As mentionned in [1] MMC sometime fails to be initialized/ is not
> working with this type of error:
>
> arasan-sdhci e010@e010.of: registered as mmc0
> malloc space: 0x17e0 -
On Mon, 07 Jul 2025 05:53:29 +, Gwenhael Goavec-Merou wrote:
> From: Gwenhael Goavec-Merou
>
> As mentionned in [1,2] sequence present in
> arch/arm/boards/avnet-zedboard/zedboard.zynqcfg
> causes boot failure without any messages.
>
> This patch removes all instructions but the last one
On Wed, 02 Jul 2025 14:03:41 +0200, Ahmad Fatoum wrote:
> flags is the generic map type (e.g. MAP_CACHED), while MAS2_I is a bit
> only relevant to PowerPC. Checking for MAS2_I in flags is thus never
> true and means we skipped cache maintenance so far when remapping a
> region as uncached.
>
>
On Wed, 02 Jul 2025 13:01:35 +0200, Ahmad Fatoum wrote:
> Stuff I noticed while looking into actual issues that have been
> submitted separately as fixes for master.
>
> Nothing critical in this series, so can go into next.
>
> Ahmad Fatoum (4):
> include: linux/iopoll.h: evaluate timeout_us
On Tue, 08 Jul 2025 21:35:56 +0200, Ahmad Fatoum wrote:
> The --env and --fs options are very useful for passing stuff into QEMU
> without having to assemble images or type stuff into the console.
>
> Document them and add a new --graphic option that spawns either a Bochs
> ISA, PCI or a ramfb g
On Wed, 09 Jul 2025 09:42:43 +0200, Stefan Kerkmann wrote:
> Before this commit writing the same srk hash twice failed with an error.
> This is inconsistent with other invocations of the hab tool e.g.: `hab
> -l -p` doesn't fail if the lockdown bit is already set.
>
> So let's align the behavior
On Wed, 09 Jul 2025 11:03:22 +0200, Stefan Kerkmann wrote:
> This allows using the string representation in other parts of barebox
> e.g. board code.
>
>
Applied, thanks!
[1/1] ARM: i.MX: ele: make ele_life_cycle public
https://git.pengutronix.de/cgit/barebox/commit/?id=0b229dff1471 (li
On Thu, 10 Jul 2025 11:06:44 +0200, Bastian Krause wrote:
> The header is necessary to add CSF data for signed barebox images in
> case HABV4_IMAGE_SIGNED=y is set. It should not hurt otherwise.
>
>
Applied, thanks!
[1/1] ARM: riotboard: add habv4-imx6-gencsf.h to flash header
https://g
On Fri, 11 Jul 2025 14:53:49 +0200, Ahmad Fatoum wrote:
> Building some commands like miitool results in a warning:
>
> miitool.rst:25: WARNING: Option list ends without a blank
> line; unexpected unindent.
>
> Have the script emit a new line after every command to silence it.
>
> [...]
A
On Wed, 09 Jul 2025 12:18:47 +0200, Ahmad Fatoum wrote:
> We have had guard page support for close to two years now and it's time
> we start enabling it by default to catch stack overflows.
>
>
Applied, thanks!
[1/2] hardening: make stack guard page the default
https://git.pengutronix.d
On Wed, 09 Jul 2025 12:12:49 +0200, Ahmad Fatoum wrote:
> ARM: interrupts: fix guard page linker error in PBL
>
> We set up the stack guard page only in barebox proper. Attempting to
> compile in support in barebox PBL leads to a compile error, because
> arm_mem_guard_page_get() references arm_m
On Mon, 14 Jul 2025 09:12:48 +0200, Ahmad Fatoum wrote:
> If we tell the serial driver core about the names of the Linux
> console and earlycon driver, barebox can fix up the console= and
> earlycon= arguments without having to hardcode Linux boot arguments.
>
>
Applied, thanks!
[1/2] serial:
On Thu, 10 Jul 2025 12:49:49 +0200, Ahmad Fatoum wrote:
> The parameters in the comment are correct, but the heading wasn't.
>
>
Applied, thanks!
[1/1] partition: fix kernel doc for region_identical
https://git.pengutronix.de/cgit/barebox/commit/?id=7afd9b64ee8a (link may
not be stable
On Fri, 11 Jul 2025 09:32:04 +0200, Ahmad Fatoum wrote:
> A consequence of running bare metal is that early failures are difficult
> to diagnose. Let's add a troubleshooting section to help users take
> the first step in diagnosing issues.
>
> Ahmad Fatoum (4):
> Documentation: devel: porting:
On Thu, 17 Jul 2025 21:13:39 +0800, Bo Sun wrote:
> Convert the comments of struct device from the old style to standard
> kernel-doc style.
>
> No functional changes.
>
>
Applied, thanks!
[1/1] driver: convert struct device comments to kernel-doc style
https://git.pengutronix.de/cgit/
On Fri, 18 Jul 2025 20:12:42 +0200, Mathieu Anquetin via B4 Relay wrote:
> On i.MX9, clock selection for DDR PHY is done by setting/clearing bit 0
> of GPR_SHARED2 register.
>
> This is done using the generic function ccm_shared_gpr_set() which takes
> two arguments, the GPR number and the value
On Fri, 11 Jul 2025 20:27:27 +0800, Bo Sun wrote:
> - Fixed a missing closing parenthesis in the comment for DEVICE_ID_SINGLE
> - Cleaned up indentation and alignment of multi-line comments in struct device
> - Removed redundant spaces in function pointer declarations
>
> No functional changes.
On Wed, 16 Jul 2025 08:27:38 +0200, Ahmad Fatoum wrote:
> We don't define, multi-search or real-search, so variables using them
> are always empty. So drop them to avoid confusion.
> We can bring them back and fix them, when we start applying overlays
> at build time.
>
>
Applied, thanks!
[1/
On Tue, 05 Aug 2025 09:33:45 +0200, Michael Tretter wrote:
> This series prepares the Rockchip tooling for creating and handling
> signed images. The patch series for adding code to sign the images will
> follow as a second step.
>
> Fix the wrong size when calculating the hash over the image he
On Mon, 21 Jul 2025 08:00:41 +0300, Antony Pavlov wrote:
> Antony Pavlov (4):
> include/stdio.h: drop serial_printf() declaration
> include/scsi.h: drop unused declarations
> include/of.h: drop of_parse_dtb() declaration
> include/driver.h: drop unused declarations
>
> include/driver.h |
On Fri, 11 Jul 2025 08:42:32 +0200, Ulrich Ölmann wrote:
>
Applied, thanks!
[1/1] ARM: i.MX: Kconfig: fix typos
https://git.pengutronix.de/cgit/barebox/commit/?id=0cc4e584f082 (link may
not be stable)
Best regards,
--
Sascha Hauer
On Fri, 04 Jul 2025 16:38:00 +0200, Ahmad Fatoum wrote:
> A consequence of running bare metal is that early failures are difficult
> to diagnose. Let's add a troubleshooting section to help users take
> the first step in diagnosing issues.
>
> Ahmad Fatoum (3):
> Documentation: devel: porting:
On Thu, 17 Jul 2025 21:14:31 +0800, Bo Sun wrote:
> The parameter represents a device tree node, not specifically a bus.
> Use more accurate naming following device tree conventions.
>
>
Applied, thanks!
[1/1] of: platform: rename parameter 'bus' to 'node' in of_platform_bus_create()
ht
On Fri, 18 Jul 2025 15:37:18 +0200, Jonas Rebmann wrote:
> GPIO-gated clocks wrongfully used barebox clock interface which leads to
> bugs due to clk not being initialized from init->ops.
>
>
Applied, thanks!
[1/1] clk: clk-gpio: Use clk_register instead of bclk_register
https://git.pen
On Thu, 03 Jul 2025 12:15:20 +0200, Ahmad Fatoum wrote:
> The indentation makes the heading a normal text and turns the dashes
> into a horizontal bar instead.
>
>
Applied, thanks!
[1/1] Documentation: migration-2025.07.0: fix formatting
https://git.pengutronix.de/cgit/barebox/commit/?i
On Fri, 04 Jul 2025 21:26:58 +0200, Johannes Roith wrote:
> Added documentation and build instructions for the Zynq 7000 and the
> Avnet Zedboard.
>
>
Applied, thanks!
[1/1] Added documentation for the Zynq 7000
https://git.pengutronix.de/cgit/barebox/commit/?id=7a0a2642ce05 (link may
On Fri, 04 Jul 2025 16:35:52 +0200, Ahmad Fatoum wrote:
> The closing backtick at the end was missing.
>
>
Applied, thanks!
[1/1] Documentation: devel: fuzzing: fix wrong ReST syntax
https://git.pengutronix.de/cgit/barebox/commit/?id=3c7862ab3962 (link may
not be stable)
Best regards,
On Mon, 07 Jul 2025 10:56:36 +0200, Ahmad Fatoum wrote:
> The file starts with defining and exporting CROSS_COMPILE_32 and
> CROSS_COMPILE_64 environment variables, which are used for building the
> individual components of the boot chain.
>
> CROSS_COMPILE should be set to the $variable, so it
On Tue, 15 Jul 2025 22:19:08 +0200, Ahmad Fatoum wrote:
> device_get_match_data returns NULL when no match data is available,
> which a cast to uintptr_t turns into a zero.
>
> Most drivers either put valid pointers into the match data or they
> reserve a literal value of zero to be "unknown".
>
+Cc Basti
On Thu, Jul 24, 2025 at 03:50:56PM +0200, Marco Felsch wrote:
> Currently the env provided keyspec is resolved during the gen_key() step
> by making use of the try_resolve_env(). This is wrong because it will
> set the complete : keyspec string for the 'keyname' and 'path'.
>
> To fix t
On Thu, 24 Jul 2025 11:13:39 +0200, Marco Felsch wrote:
> Before commit a6410cfa9aff ("block: mark underlying cdev with
> DEVFS_IS_BLOCK_DEV") the block device decision was made based on the
> block_ops. Since each parition-cdev inherits the parent (master)
> cdev-ops the parition-cdevs were dete
On Wed, 23 Jul 2025 17:55:45 +0200, Marco Felsch wrote:
> cdev_is_block_partition() fails if a cdev is passed which is a
> parition-link. Most of our barebox code is link-aware by resolving the
> links first.
>
> Make it more user-friendly and resolve the possible cdev parttion link
> first befo
On Tue, 22 Jul 2025 16:05:34 +0200, Ahmad Fatoum wrote:
> Uwe reports on IRC that fastboot tries to read some default variables
> that barebox doesn't define, which triggers host-side error output:
>
> ERROR: fastboot: no such variable: is-logical:kernel
>
> Flashing didn't seem impacted thou
On Fri, 25 Jul 2025 00:31:58 -0400, chalian...@gmail.com wrote:
>
Applied, thanks!
[1/1] pinctrl: tegra30: fix typo with variables to control the drive state on
nvidia pins.
https://git.pengutronix.de/cgit/barebox/commit/?id=47b3be77fb21 (link may
not be stable)
Best regards,
--
Sas
On Tue, 29 Jul 2025 00:56:07 -0400, chalian...@gmail.com wrote:
> not a string literal and no format
> arguments [-Werror=format-security]
>
>
Applied, thanks!
[1/1] clk: clkdev: fix format security.
https://git.pengutronix.de/cgit/barebox/commit/?id=ffeeae51a5cf (link may
not be stabl
On Tue, 29 Jul 2025 15:59:24 +0200, Michael Tretter wrote:
> EXT_CSD_RPMB_SIZE_MULT contains the number of 128 KB blocks in the RPMB.
> While the calculated RMPB is labeled as KB, it actually contains the
> size in bytes.
>
> Fix the calculation to report the RPMB size in KB.
>
> An alternative
On Tue, 29 Jul 2025 23:36:57 +0300, Ivaylo Ivanov wrote:
> Some platforms, like the Samsung Galaxy S8, leave CNTFRQ_EL0 unset in
> the previous stage bootloader. Therefore reading it causes a hang and
> the boot proccess is halted.
>
> Since on such retail devices there is no way to set it witho
The hash should include the entire header, which is actually 1536 bytes
long. It seems that the hash is not checked for unsigned images and thus
this wasn't an issue until now.
If the hash is used for signing the image, hashing 1535 bytes instead of
1536 bytes causes a verification failure.
Signe
RKSS is the magic for signed images. Add this magic to the rk-usb-loader
to be able to load signed images via USB.
Signed-off-by: Michael Tretter
---
Changes in v2:
- none
---
scripts/rk-usb-loader.c | 7 ---
scripts/rockchip.h | 1 +
2 files changed, 5 insertions(+), 3 deletions(-)
di
The idb contains a flag for the used hash type. Use the flag to
determine the used hash function and update the hash accordingly.
Signed-off-by: Michael Tretter
---
Changes in v2:
- Drop double negation in flag check
---
scripts/rkimage.c | 17 -
1 file changed, 12 insertions(+),
There are different magic values for different image types. Include the
type of magic in the name of the magic.
Signed-off-by: Michael Tretter
---
Changes in v2:
- none
---
scripts/rk-usb-loader.c | 4 ++--
scripts/rkimage.c | 2 +-
scripts/rockchip.h | 2 +-
3 files changed, 4 insert
The hashtype is actually a bitfield that contains flags for the
different hash types and other information. Rename the field to flags
and add constants for the flags in this field.
Signed-off-by: Michael Tretter
---
Changes in v2:
- none
---
scripts/rkimage.c | 5 +++--
scripts/rockchip.h | 5 +
This series prepares the Rockchip tooling for creating and handling
signed images. The patch series for adding code to sign the images will
follow as a second step.
Fix the wrong size when calculating the hash over the image header.
Rename the hashtype field in the header to flags, because there a
devfs-core.c implements the partition handling only for devices that are
not MTD devices. MTD devices have their own partition handling in
drivers/mtd/partition.c, see the special case in __devfs_add_partition()
for MTD devices. This means for MTD devices we have to pass the original
cdev to the op
77 matches
Mail list logo
