| Age | Commit message (Collapse) | Author |
|---|
|
This patch series enables full RVVM virtual machine support which was
earlier inconveniently provided as out-of-tree patchset.
This should be cleaner than a separate board config, since both
emulators provide similar feature set.
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
Enable framebuffer for better virtual machine integration.
Some guests need EFI FB to work properly.
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
QEMU supports NVMe devices, but U-Boot only tries to boot from Virtio.
This is problematic when explicitly using NVMe, so fix that.
Additionally, RVVM virtual machine is almost fully compatible with QEMU,
except it only implements NVMe drives instead of VirtIO.
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
Enable GPIO command and DWAPB_GPIO driver for LicheePi4A board.
Signed-off-by: Maksim Kiselev <[email protected]>
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
Starting with Linux commit 7569486d79ae ("gpio: dwapb: Add ngpios
DT-property support") the "snps,nr-gpios" property was marked
as deprecated.
And since all newly added dw-apb-gpio nodes are described using generic
"ngpios" property, it's time to add support for it.
Signed-off-by: Maksim Kiselev <[email protected]>
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
The SBI 3.0 specification [1] adds the following extensions:
* Firmware Features Extension
* Message Proxy Extension
Let the sbi command detect their availability.
The Firmware Features Extension is already implemented in OpenSBI.
Correct the text for the DBTR extension and move it to the same position
as in the specification.
[1] https://github.com/riscv-non-isa/riscv-sbi-doc/releases/download/vv3.0-rc1/riscv-sbi.pdf
Signed-off-by: Heinrich Schuchardt <[email protected]>
Reviewed-by: Leo Yu-Chi Liang <[email protected]>
|
|
The config setting CMD_BOOTEFI_HELLO_COMPILE was removed in favour
of BOOTEFI_HELLO_COMPILE but the usage in this Makefile was not
updated. Fix it.
Fixes: 6fe80876dcc7 ("efi_loader: Rename and move CMD_BOOTEFI_HELLO_COMPILE")
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
The config setting CMD_BOOTEFI_HELLO_COMPILE was removed in favour
of BOOTEFI_HELLO_COMPILE but the dependency for
test_efi_helloworld_net_http was not updated and so is now incorrect
preventing the test from ever running. Fix it.
Fixes: 6fe80876dcc7 ("efi_loader: Rename and move CMD_BOOTEFI_HELLO_COMPILE")
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
Reviewed-by: Ilias Apalodimas <[email protected]>
|
|
Patrick Rudolph <[email protected]> says:
Based on the existing work done by Simon Glass this series adds
support for booting aarch64 devices using ACPI only.
As first target QEMU SBSA support is added, which relies on ACPI
only to boot an OS. As secondary target the Raspberry Pi4 was used,
which is broadly available and allows easy testing of the proposed
solution.
The series is split into ACPI cleanups and code movements, adding
Arm specific ACPI tables and finally SoC and mainboard related
changes to boot a Linux on the QEMU SBSA and RPi4. Currently only the
mandatory ACPI tables are supported, allowing to boot into Linux
without errors.
The QEMU SBSA support is feature complete and provides the same
functionality as the EDK2 implementation.
The changes were tested on real hardware as well on QEMU v9.0:
qemu-system-aarch64 -machine sbsa-ref -nographic -cpu cortex-a57 \
-pflash secure-world.rom \
-pflash unsecure-world.rom
qemu-system-aarch64 -machine raspi4b -kernel u-boot.bin -cpu cortex-a72 \
-smp 4 -m 2G -drive file=raspbian.img,format=raw,index=0 \
-dtb bcm2711-rpi-4-b.dtb -nographic
Tested against FWTS V24.03.00.
Known issues:
- The QEMU rpi4 support is currently limited as it doesn't emulate PCI,
USB or ethernet devices!
- The SMP bringup doesn't work on RPi4, but works in QEMU (Possibly
cache related).
- PCI on RPI4 isn't working on real hardware since the pcie_brcmstb
Linux kernel module doesn't support ACPI yet.
Link: https://lore.kernel.org/r/[email protected]
|
|
Add QEMU's SBSA ref board to azure pipelines and gitlab CI to run tests on it.
TEST: Run on Azure pipelines and confirmed that tests succeed.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Allows to build the RPi4 with ACPI enabled.
TEST: - Boots on qemu-system-aarch64 -machine raspi4b
- Boots on real hardware with arm_64bit=1 in config.txt
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
|
|
Initialize addr to zero which allows to build on the CI
which is more strict.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
On FDT platforms the APs are brought out of reset using the spin-table
as defined in the DT. On ACPI enabled platforms there's no FDT, thus
enable the ARMV8_MULTIENTRY and BLOBLIST to use the ACPI parking
protocol instead.
TEST: - APs enter Linux when run on qemu-system-aarch64 -machine raspi4b
- APs on real hardware do not enter Linux, but continue spinning in
their ACPI parking protocol spinloop. To be investigated.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
When ACPI is enabled over FDT the APs cannot be brought out of reset
by the OS using the "FDT spin-table" mechanism, as no FDT is provided
to the OS. The APs must be released out of reset in u-boot and then
brought up in an ACPI compliant fashion.
When ARMV8_MULTIENTRY is specified, the APs are released from reset
and will enter U-Boot after it has been relocated as well.
By default ARMV8_MULTIENTRY is not selected, keeping existing behaviour.
TEST: All APs enter U-Boot when run on qemu-system-aarch64 and on
real hardware.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Implement read_mpidr() on armv7 to make use of it in generic
code that compiles on both armv7 and armv8.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Update the generic entry point code to support the ACPI parking protocol.
The ACPI parking protocol can be used when PSCI is not available to bring
up secondary CPU cores.
When enabled secondary CPUs will enter U-Boot proper and spin in their own
4KiB reserved memory page, which also acts as mailbox with the OS to
release the CPU.
TEST: Boots all CPUs on qemu-system-aarch64 -machine raspi4b
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
On Arm platforms that use ACPI they cannot rely on the "spin-table"
CPU bringup usually defined in the FDT. Thus implement the
'ACPI Multi-processor Startup for ARM Platforms', also referred to as
'ACPI parking protocol'.
The ACPI parking protocol works similar to the spin-table mechanism, but
the specification also covers lots of shortcomings of the spin-table
implementations.
Every CPU defined in the ACPI MADT table has it's own 4K page where the
spinloop code and the OS mailbox resides. When selected the U-Boot board
code must make sure that the secondary CPUs enter u-boot after relocation
as well, so that they can enter the spinloop code residing in the ACPI
parking protocol pages.
The OS will then write to the mailbox and generate an IPI to release the
CPUs from the spinloop code.
For now it's only implemented on ARMv8, but can easily be extended to
other platforms, like ARMv7.
TEST: Boots all CPUs on qemu-system-aarch64 -machine raspi4b
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Add the necessary DSDT files copied from tianocore to boot the RPi4.
In addition generate a board specific SSDT to dynamically enable/disable
ACPI devices based on FDT. This is required to support the various variants
and boot options. It also allows to test the code on QEMU 9.0 without
modifications, since it doesn't emulate PCIe yet.
Signed-off-by: Simon Glass <[email protected]>
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
|
|
Generate SoC specific ACPI tables for BCM2711:
- FADT
- PPTT
- GTDT
Board specific tables like DSDT and SSDT are added in a separate patch.
MADT is already properly generated from the FDT.
When ACPI is enabled for a different SoC compliation will fail by
design, indicating the required functions that needs to be implemented.
When ACPI is not enabled the added code does nothing, keeping existing
behaviour.
TEST: Booted on RPi4 with only ACPI enabled, providing no FDT to the OS.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
These header files presumably duplicate things already in the U-Boot
devicetree. For now, bring them in to get the ASL code and ACPI table
code to compile.
Signed-off-by: Simon Glass <[email protected]>
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Acked-by: Matthias Brugger <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Cover the "ARM local MMIO" range as well in the default MMU mapping in
order to allow future code to access the GIC-400 without crashing. For
now the GIC is not touched in u-boot, thus this change is a noop.
See [1](BCM2711 ARM Peripherals) for reference.
TEST: Enabled CONFIG_GICV2 and accessed the GIC in C code without crash.
1: https://datasheets.raspberrypi.com/bcm2711/bcm2711-peripherals.pdf
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Acked-by: Matthias Brugger <[email protected]>
Cc: Matthias Brugger <[email protected]>
Cc: Peter Robinson <[email protected]>
|
|
Add support for Arm sbsa [1] v0.3+ that is supported by QEMU [2].
Unlike other Arm based platforms the machine only provides a minimal
FDT that contains number of CPUs, ammount of memory and machine-version.
The boot firmware has to provide ACPI tables to the OS.
Due to this design a full DTB is added here as well that allows U-Boot's
driver to properly function. The DTB is appended at the end of the U-Boot
image and will be merged with the QEMU provided DTB.
In addition provide documentation how to use, enable binman to fabricate both
ROMs that are required to boot and add ACPI tables to make it full compatible
to the EDK2 reference implementation.
The board was tested using Fedora 40 Aarch64 Workstation. It's able
to boot from USB and AHCI or network.
Tested and found working:
- serial
- PCI
- xHCI
- Bochs display
- AHCI
- network using e1000e
- CPU init
- Booting Fedora 40
1: Server Base System Architecture (SBSA)
2: https://www.qemu.org/docs/master/system/arm/sbsa.html
Signed-off-by: Patrick Rudolph <[email protected]>
Cc: Peter Robinson <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Allow to use BLOBLIST_TABLES on arm to store ACPI or other tables.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Fill the MADT table in the GIC driver and armv8 CPU driver to
drop SoC specific code. While the GIC only needs devicetree
data, the CPU driver needs additional information stored in
the cpu_plat struct.
While on it update the only board making use of the existing
drivers and writing ACPI MADT in mainboard code.
TEST: Booted on QEMU sbsa-ref using GICV3 driver model generated MADT.
Booted on QEMU raspb4 using GICV2 driver model generated MADT.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Support reading the "interrupts" property from the devicetree in case
the "interrupts-extended" property isn't found. As the "interrupts"
property is commonly used, this allows to parse all existing FDT and
makes irq_get_by_index() more useful.
The "interrupts" property doesn't contain a phandle as "interrupts-extended"
does, so implement a new method to locate the interrupt-parent called
irq_get_interrupt_parent().
TEST: Read the interrupts from the GIC node for ACPI MADT generation.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Moritz Fischer <[email protected]>
|
|
Add a generic GICV2 driver that:
- parses the DT and generates the ACPI MADT subtables
- implement of_xlate() and allows irq_get_by_index() to return the
correct interrupt mappings
Map DT interrupts to ARM GIC interrupts as follows:
- Interrupt numbers ID32-ID1019 are used for SPIs
- ID0-ID15 are used for SGIs
- ID16-ID31 are used for PPIs
TEST: Booted on QEMU raspb4 using GICV2 driver model generated MADT.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Translate IRQs by implementing of_xlate() as required by
irq_get_by_index() to parse interrupt properties.
Map DT interrupts to ARM GIC interrupts as follows:
- Interrupt numbers ID32-ID1019 are used for SPIs
- ID0-ID15 are used for SGIs
- ID16-ID31 are used for PPIs
TEST: Booted on qemu sbsa-ref that has a GICV3.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Moritz Fischer <[email protected]>
|
|
The code accesses the gic-v3 node, but not the gic-v3-its node,
thus rename the objects to clarify which node it operates on.
The following commit will make use of the gic-v3-its node for real.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
|
Add a generic driver that binds to armv8 CPU nodes. The generic driver allows
- to enumerate CPUs present in a system, even when no other driver binds it
- generates ACPI SSDT code for each CPU
- Fill the ACPI MADT table (implemented in a follow up patch)
The newly introduced code could also be reused on other CPU drivers that are
compatible with armv8.
TEST: Booted on QEMU sbsa and verify the driver binds to CPU nodes.
Confirmed with FWTS that all ACPI processor devices are present.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Rename 'ahci_mvebu' to 'ahci_generic' and select it by default.
The AHCI driver contains no SoC specific code and only expects the
base address to be passed, thus rename it to ahci_generic and add the
DT compatible string "generic-ahci".
Update existing defconfigs to use the new Kconfig name as well.
TEST: Booted on QEMU sbsa using the generic-ahci node.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Stefan Roese <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Stefan Roese <[email protected]>
Acked-by: Tony Dinh <[email protected]>
|
|
Add support for the generic XHCI driver that contains no SoC
specific code. It can be used on platforms that simply work out
of the box, like on emulated platforms.
TEST: Booted on QEMU sbsa machine using the generic xhci driver.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Marek Vasut <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Marek Vasut <[email protected]>
|
|
Add a new method to write the processor device identified by _HID
ACPI0007, that is preferred over the Processor OpCode since ACPI 6.0.
Fixes booting arm using ACPI only since the Processor OpCode isn't
found valid by the Linux kernel.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Allocate memory for ACPI tables in generic acpi code. When ACPI wasn't
installed in other places, install the ACPI table using BLOBLISTs.
This allows non x86 platforms to boot using ACPI only in case the
EFI loader is being used, since EFI is necessary to advertise the location
of the ACPI tables in memory.
TEST: Booted QEMU SBSA (no QFW) using EFI and ACPI only.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Since ACPI 2.0 the RSDT is deprecated and the XSDT should be preferred.
Until now the RSDT and XSDT entries were keept in sync as all platforms
that installed ACPI tables placed them below 4GiB and thus the address
would fit into the 32bit RSDT.
On platforms that do not have usable DRAM below 4GiB, like QEMU sbsa,
the RSDT cannot be used. Allow both RSDT and XSDT to be null and only
fill those tables that are present in acpi_add_table().
TEST: Fixes a crash on QEMU sbsa and allows to boot on QEMU sbsa.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
|
|
Allow other architectures to use acpi_create_mcfg_mmconfig as well
by moving the function prototype to common code.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
The SoC can implement acpi_fill_iort to update the IORT table.
Add a helper function to fill out the NAMED_COMPONENT node.
TEST=Run FWTS V24.03.00 on RPi4 and round no problems.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Return the ACPI table revision in acpi_get_table_revision() for
PPTT and GTDT. Match both to ACPI 6.2.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
The FADT structure found in U-Boot represents FADT revision 6 and the
GICC and GICD structures defined in U-Boot are based on ACPI revision
6.3.
Bump the table revision to fix FWTS failures seen on aarch64.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Bin Meng <[email protected]>
|
|
Add a new method to acpi_ops to let drivers fill out ACPI MADT.
The code is unused for now until drivers implement the new ops.
TEST: Booted on QEMU sbsa using driver model generated MADT.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Add generic ACPI code to generate
- MADT GICC
- MADT GICD
- MADT GICR
- MADT GIC ITS
- PPTT processor
- PPTT cache
as commonly used on arm platforms.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Add the interrupt flag used in ACPI GTDT table as define.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Reviewed-by: Moritz Fischer <[email protected]>
|
|
When ACPI is enabled on arm it will use the getinfo function to fill
the SPCR ACPI table.
Signed-off-by: Maximilian Brune <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Reviewed-by: Moritz Fischer <[email protected]>
|
|
Rename ACPI tables MADR to MADT.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
|
|
Write MADT in common code and let the SoC fill out the body by
calling acpi_fill_madt() which must be implemented at SoC level.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Bin Meng <[email protected]>
|
|
Write the FADT in common code since it's used on all architectures.
Since the FADT is mandatory all SoCs or mainboards must implement the
introduced function acpi_fill_fadt() and properly update the FADT.
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Bin Meng <[email protected]>
|
|
This moves the SPCR and DBG2 table generation into common code, so that
they can be used by architectures other than x86.
Signed-off-by: Maximilian Brune <[email protected]>
Signed-off-by: Patrick Rudolph <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Cc: Simon Glass <[email protected]>
Cc: Bin Meng <[email protected]>
|
|
This patch series by Hanyuan Zhao <[email protected]> provides a number of
improvements to the dc2114x driver.
Link: https://lore.kernel.org/r/[email protected]
|
|
bus"
Romain Naour <[email protected]> says:
We are using a custom board where an ethernet switch device KSZ9896 is
available. This family of devices can use several types of serial bus
as management interface: mdio, i2c or SPI. Due to board design
constraints and because we initially planned to use this device only
from Linux, the SPI bus was used.
Luckily we are using a recent enough u-boot release where KSZ9477
driver is available... but only for the i2c interface. Indeed, unlike
the kernel driver, the KSZ9477 driver doesn't use the regmap API to
access the underlying bus since the regmap API is limited to direct
memory access [1].
Until regmap API with bus support is available in U-boot, we introduced
struct ksz_phy_ops to store low-level ksz bus operations (I2C or SPI).
This series has been tested on the current master branch (after v2024.10
release).
[1] https://source.denx.de/u-boot/u-boot/-/blob/v2024.10-rc5/drivers/core/Kconfig?ref_type=tags#L188
Link: https://lore.kernel.org/r/[email protected]
|
|
Frank Sae <[email protected]> says:
YT8531 as Gigabit transceiver uses bit15:14(bit9 reserved default 0) as phy
speed mask, YT8821 as 2.5 Gigabit transceiver uses bit9 bit15:14 as phy
speed mask.
Be compatible to YT8821, reform phy speed mask and phy speed macro.
Based on update above, add YT8821 2.5G phy driver.
Link: https://lore.kernel.org/r/[email protected]
|
|
Some drivers do not behave properly when free_pkt() is called with a
length of zero. It is an issue I observed when developing the lwIP
series [1] (see "QEMU CI tests for r2dplus_i82557c, r2dplus_rtl8139"
in the change log) and which I fixed incorrectly by not calling
free_pkt() when recv() returns 0. That turned out to be wrong for two
reasons:
1. The DM documentation [2] clearly requires it:
"The **recv** function polls for availability of a new packet. [...]
If there is an error [...], return 0 if you require the packet to
be cleaned up normally, or a negative error code otherwise (cleanup
not necessary or already done).
If **free_pkt** is defined, U-Boot will call it after a received
packet has been processed [...]. free_pkt() will be called after
recv(), for the same packet [...]"
2. The imx8mp_evk platform will fail with OOM errors if free_pkt() is
not called after recv() returns 0:
u-boot=> tftp 192.168.0.16:50M
Using ethernet@30be0000 device
TFTP from server 192.168.0.16; our IP address is 192.168.0.48
Filename '50M'.
Load address: 0x40480000
Loading: #######################fecmxc_recv: error allocating packetp
fecmxc_recv: error allocating packetp
fecmxc_recv: error allocating packetp
...
Therefore, make recv() return -EAGAIN instead of 0 when no packet is
available and the driver doesn't expect free_pkt() to be called
subsequently.
[1] https://lists.denx.de/pipermail/u-boot/2024-August/562861.html
[2] doc/develop/driver-model/ethernet.rst
Signed-off-by: Jerome Forissier <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
|
