Re: [PATCH v6 0/4] Add Microchip IPC mailbox

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1639 bytes --]

On Mon, Jun 16, 2025 at 08:44:01PM +0200, Klara Modin wrote:
> On 2024-12-17 11:31:30 +0000, Valentina Fernandez wrote:
> > 
> > Valentina Fernandez (4):
> >   riscv: sbi: vendorid_list: Add Microchip Technology to the vendor list
> >   riscv: export __cpuid_to_hartid_map
> >   dt-bindings: mailbox: add binding for Microchip IPC mailbox controller
> >   mailbox: add Microchip IPC support
> > 
> >  .../bindings/mailbox/microchip,sbi-ipc.yaml   | 123 +++++
> >  arch/riscv/include/asm/vendorid_list.h        |   1 +
> >  arch/riscv/kernel/smp.c                       |   1 +
> >  drivers/mailbox/Kconfig                       |  13 +
> >  drivers/mailbox/Makefile                      |   2 +
> >  drivers/mailbox/mailbox-mchp-ipc-sbi.c        | 504 ++++++++++++++++++
> >  include/linux/mailbox/mchp-ipc.h              |  33 ++
> >  7 files changed, 677 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/mailbox/microchip,sbi-ipc.yaml
> >  create mode 100644 drivers/mailbox/mailbox-mchp-ipc-sbi.c
> >  create mode 100644 include/linux/mailbox/mchp-ipc.h
> > 
> > -- 
> > 2.34.1
> > 
> 
> Building this driver as a module with SMP=n fails becuase the non-SMP
> cpuid_to_hartid_map() depends on the non-exported boot_cpu_hartid
> (discovered via randconfig).
> 
> >From the description of the driver it doesn't seem to be very useful
> with SMP=n, would it make sense to have it depend on SMP rather than
> exporting boot_cpu_hartid?

No, I think boot_cpu_hartid needs to get exported, the driver is
intended for use in AMP situations so the kernel config may have SMP
disabled.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]