| Message ID | 20250214062443.9936-1-sebastian.huber@embedded-brains.de (mailing list archive) |
|---|---|
| Headers | show |
| Series | Improve Microchip Polarfire SoC customization | expand |
+cc qemu-riscv, Alistar. On Fri, Feb 14, 2025 at 07:24:37AM +0100, Sebastian Huber wrote: > Booting the microchip-icicle-kit machine using the latest PolarFire SoC > Hart Software Services (HSS) no longer works since Qemu lacks support > for several registers (clocks, DRAM controller). Also reading from the > SDCard does not work currently. On that note, I think the inaccurate docs about polarfire should be removed. There's a wiki page here with dead links, or links to things that do not work anymore: https://wiki.qemu.org/Documentation/Platforms/RISCV#Microchip_PolarFire_SoC_Icicle_Kit I think the whole section should be removed, find it kinda odd that there's a polarfire section but not for any other board. Either way, it's talking about something that just does not work, the current HSS and Yocto don't boot. There's also a docs page here: https://www.qemu.org/docs/master/system/riscv/microchip-icicle-kit.html that has a copy of the table your patch 4 modifies, that probably should be updated to match your changes. In a similar vein to the wiki, it talks about the HSS and booting a yocto wic image. I think those should be deleted since they don't work. Alistar/Other RISC-V folks, what do you think? Bin wrote the port but seems to be AFK and I don't have the capacity to fix any of that stuff on top of what I already do in my spare time - do you agree that deleting the now inaccurate docs makes sense? > In order to allow tests runs for real-time kernels such as RTEMS and > Zephyr, improve the boot customization. This patch set enables a direct > run of kernel executables, for example: > > qemu-system-riscv64 -no-reboot -nographic \ > -serial null -serial mon:stdio \ > -smp 2 \ > -bios none \ > -machine microchip-icicle-kit,clint-timebase-frequency=10000000 \ > -kernel rtos.elf The series breaks my usage: qemu//build/qemu-system-riscv64 -M microchip-icicle-kit \ -m 3G -smp 5 \ -kernel vmlinux.bin \ -dtb riscvpc.dtb \ -initrd initramfs.cpio.gz \ -display none -serial null \ -serial mon:stdio \ -D qemu.log -d unimp opensbi-riscv64-generic-fw_dynamic.bin: No such file or directory qemu-system-riscv64: could not load firmware 'opensbi-riscv64-generic-fw_dynamic.bin' make: *** [Makefile:305: qemu-icicle] Error 1 Figure it is likely to be your patch 4? The file does exist, so probably some sort of path-to-it issues? Cheers, Conor. > > Sebastian Huber (5): > hw/misc: Add MPFS system reset support > hw/riscv: More flexible FDT placement for MPFS > hw/riscv: Make FDT optional for MPFS > hw/riscv: Allow direct start of kernel for MPFS > hw/riscv: Configurable MPFS CLINT timebase freq > > hw/misc/mchp_pfsoc_sysreg.c | 7 ++ > hw/riscv/microchip_pfsoc.c | 147 +++++++++++++++++++++-------- > include/hw/riscv/microchip_pfsoc.h | 1 + > 3 files changed, 115 insertions(+), 40 deletions(-) > > -- > 2.43.0 > >
Hi Conor, On 20/2/25 19:30, Conor Dooley wrote: > +cc qemu-riscv, Alistar. > > On Fri, Feb 14, 2025 at 07:24:37AM +0100, Sebastian Huber wrote: >> Booting the microchip-icicle-kit machine using the latest PolarFire SoC >> Hart Software Services (HSS) no longer works since Qemu lacks support >> for several registers (clocks, DRAM controller). Also reading from the >> SDCard does not work currently. > > On that note, I think the inaccurate docs about polarfire should be > removed. There's a wiki page here with dead links, or links to things > that do not work anymore: > https://wiki.qemu.org/Documentation/Platforms/RISCV#Microchip_PolarFire_SoC_Icicle_Kit > I think the whole section should be removed, find it kinda odd that > there's a polarfire section but not for any other board. Either way, > it's talking about something that just does not work, the current HSS > and Yocto don't boot. > > There's also a docs page here: > https://www.qemu.org/docs/master/system/riscv/microchip-icicle-kit.html > that has a copy of the table your patch 4 modifies, that probably should > be updated to match your changes. > > In a similar vein to the wiki, it talks about the HSS and booting a > yocto wic image. I think those should be deleted since they don't work. > > Alistar/Other RISC-V folks, what do you think? Bin wrote the port but > seems to be AFK and I don't have the capacity to fix any of that stuff > on top of what I already do in my spare time - do you agree that > deleting the now inaccurate docs makes sense? > >> In order to allow tests runs for real-time kernels such as RTEMS and >> Zephyr, improve the boot customization. This patch set enables a direct >> run of kernel executables, for example: >> >> qemu-system-riscv64 -no-reboot -nographic \ >> -serial null -serial mon:stdio \ >> -smp 2 \ >> -bios none \ >> -machine microchip-icicle-kit,clint-timebase-frequency=10000000 \ >> -kernel rtos.elf > > The series breaks my usage: > qemu//build/qemu-system-riscv64 -M microchip-icicle-kit \ > -m 3G -smp 5 \ > -kernel vmlinux.bin \ > -dtb riscvpc.dtb \ > -initrd initramfs.cpio.gz \ > -display none -serial null \ > -serial mon:stdio \ > -D qemu.log -d unimp > opensbi-riscv64-generic-fw_dynamic.bin: No such file or directory > qemu-system-riscv64: could not load firmware 'opensbi-riscv64-generic-fw_dynamic.bin' > make: *** [Makefile:305: qemu-icicle] Error 1 > > Figure it is likely to be your patch 4? The file does exist, so probably > some sort of path-to-it issues? Maybe missing the -L option? -L path set the directory for the BIOS, VGA BIOS and keymaps Regards, Phil.
----- Am 20. Feb 2025 um 23:29 schrieb Philippe Mathieu-Daudé philmd@linaro.org: > Hi Conor, > > On 20/2/25 19:30, Conor Dooley wrote: >> +cc qemu-riscv, Alistar. >> >> On Fri, Feb 14, 2025 at 07:24:37AM +0100, Sebastian Huber wrote: >>> Booting the microchip-icicle-kit machine using the latest PolarFire SoC >>> Hart Software Services (HSS) no longer works since Qemu lacks support >>> for several registers (clocks, DRAM controller). Also reading from the >>> SDCard does not work currently. >> >> On that note, I think the inaccurate docs about polarfire should be >> removed. There's a wiki page here with dead links, or links to things >> that do not work anymore: >> https://wiki.qemu.org/Documentation/Platforms/RISCV#Microchip_PolarFire_SoC_Icicle_Kit >> I think the whole section should be removed, find it kinda odd that >> there's a polarfire section but not for any other board. Either way, >> it's talking about something that just does not work, the current HSS >> and Yocto don't boot. >> >> There's also a docs page here: >> https://www.qemu.org/docs/master/system/riscv/microchip-icicle-kit.html >> that has a copy of the table your patch 4 modifies, that probably should >> be updated to match your changes. >> >> In a similar vein to the wiki, it talks about the HSS and booting a >> yocto wic image. I think those should be deleted since they don't work. >> >> Alistar/Other RISC-V folks, what do you think? Bin wrote the port but >> seems to be AFK and I don't have the capacity to fix any of that stuff >> on top of what I already do in my spare time - do you agree that >> deleting the now inaccurate docs makes sense? >> >>> In order to allow tests runs for real-time kernels such as RTEMS and >>> Zephyr, improve the boot customization. This patch set enables a direct >>> run of kernel executables, for example: >>> >>> qemu-system-riscv64 -no-reboot -nographic \ >>> -serial null -serial mon:stdio \ >>> -smp 2 \ >>> -bios none \ >>> -machine microchip-icicle-kit,clint-timebase-frequency=10000000 \ >>> -kernel rtos.elf >> >> The series breaks my usage: >> qemu//build/qemu-system-riscv64 -M microchip-icicle-kit \ >> -m 3G -smp 5 \ >> -kernel vmlinux.bin \ >> -dtb riscvpc.dtb \ >> -initrd initramfs.cpio.gz \ >> -display none -serial null \ >> -serial mon:stdio \ >> -D qemu.log -d unimp >> opensbi-riscv64-generic-fw_dynamic.bin: No such file or directory >> qemu-system-riscv64: could not load firmware >> 'opensbi-riscv64-generic-fw_dynamic.bin' >> make: *** [Makefile:305: qemu-icicle] Error 1 >> >> Figure it is likely to be your patch 4? The file does exist, so probably >> some sort of path-to-it issues? > > Maybe missing the -L option? > > -L path set the directory for the BIOS, VGA BIOS and keymaps It was an error in patch 4/5. I sent a v2 version of it. You have to find the firmware, before you can load it.