Date
1 - 6 of 6
[isar-cip-core][PATCH v3 6/9] enhance qemu-riscv64 machine to be testable
|
Felix Moessbauer <felix.moessbauer@...>
This patch enhances the qemu-riscv64 machine by adding a reference to
u-boot. Further, we now use the qemu_riscv64 defconfig from cip-kernel-config. Signed-off-by: Felix Moessbauer <felix.moessbauer@...> --- conf/machine/qemu-riscv64.conf | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/conf/machine/qemu-riscv64.conf b/conf/machine/qemu-riscv64.conf index f1f3e87..8c1764b 100644 --- a/conf/machine/qemu-riscv64.conf +++ b/conf/machine/qemu-riscv64.conf @@ -12,4 +12,11 @@ DISTRO_ARCH = "riscv64" IMAGE_FSTYPES ?= "ext4" -KERNEL_DEFCONFIG ?= "defconfig" +USE_CIP_KERNEL_CONFIG = "1" + +KERNEL_DEFCONFIG ?= "cip-kernel-config/${KERNEL_DEFCONFIG_VERSION}/riscv/qemu_riscv64_defconfig" + +# for SWUpdate setups: watchdog is configured in U-Boot +WDOG_TIMEOUT = "0" + +PREFERRED_PROVIDER_u-boot-${MACHINE} = "u-boot-qemu-riscv64" -- 2.34.1 |
|
|
|
Jan Kiszka
On 02.03.23 16:26, Felix Moessbauer wrote:
This patch enhances the qemu-riscv64 machine by adding a reference toIs that true for qemu, or is that just a copy-and-paste statement? +WDOG_TIMEOUT = "0"Jan -- Siemens AG, Technology Competence Center Embedded Linux |
|
|
|
Jan Kiszka
On 03.03.23 09:35, Moessbauer, Felix (T CED INW-CN) wrote:
On Fri, 2023-03-03 at 08:13 +0100, Jan Kiszka wrote:Ah, I recall. There was a plan to have one, but it was shot down becauseOn 02.03.23 16:26, Felix Moessbauer wrote:Actually, I just copied that from the arm64 example. But having aThis patch enhances the qemu-riscv64 machine by adding a referenceIs that true for qemu, or is that just a copy-and-paste statement? I was not motivated to add that exotic watchdog driver to U-Boot: https://mail.gnu.org/archive/html/qemu-arm/2022-05/msg00154.html Jan -- Siemens AG, Technology Competence Center Embedded Linux |
|
|
|
Jan Kiszka
On 03.03.23 09:51, Moessbauer, Felix (T CED INW-CN) wrote:
On Fri, 2023-03-03 at 09:45 +0100, Jan Kiszka wrote:Fine with me.On 03.03.23 09:35, Moessbauer, Felix (T CED INW-CN) wrote:Thanks for the notice. How shall we proceed?On Fri, 2023-03-03 at 08:13 +0100, Jan Kiszka wrote:Ah, I recall. There was a plan to have one, but it was shot downOn 02.03.23 16:26, Felix Moessbauer wrote:Actually, I just copied that from the arm64 example. But having aThis patch enhances the qemu-riscv64 machine by adding aIs that true for qemu, or is that just a copy-and-paste Given that qemu-riscv64 would be the second "board" to benefit from a i6300esb driver in U-Boot, it might be worth looking into eventually. Provided both qemu-riscv64 and qemu-arm* can have PCI support as well in U-Boot. But not a topic for this series, for sure. Jan -- Siemens AG, Technology Competence Center Embedded Linux |
|
|
|
Moessbauer, Felix <felix.moessbauer@...>
On Fri, 2023-03-03 at 08:13 +0100, Jan Kiszka wrote:
On 02.03.23 16:26, Felix Moessbauer wrote:Actually, I just copied that from the arm64 example. But having aThis patch enhances the qemu-riscv64 machine by adding a referenceIs that true for qemu, or is that just a copy-and-paste statement? second look at the code, makes me think that it is also wrong there: - the u-boot option CONFIG_HW_WATCHDOG is not set - the start-qemu.sh script does not setup any watchdog in QEMU In short: It works, because we don't have a watchdog and nobody enables it. But I'm not an expert in this domain. Maybe Quirin can briefly comment on the intention of the original code. Felix +WDOG_TIMEOUT = "0"Jan |
|
|
|
Moessbauer, Felix <felix.moessbauer@...>
On Fri, 2023-03-03 at 09:45 +0100, Jan Kiszka wrote:
On 03.03.23 09:35, Moessbauer, Felix (T CED INW-CN) wrote:Thanks for the notice. How shall we proceed?On Fri, 2023-03-03 at 08:13 +0100, Jan Kiszka wrote:Ah, I recall. There was a plan to have one, but it was shot downOn 02.03.23 16:26, Felix Moessbauer wrote:Actually, I just copied that from the arm64 example. But having aThis patch enhances the qemu-riscv64 machine by adding aIs that true for qemu, or is that just a copy-and-paste I vote for keeping WDOG_TIMEOUT=0 but changing the comment to something like that: "Watchdog is not yet supported in our QEMU executor for this platform, disable it". Felix
|
|
|