[RFC PATCH 2/2] lib: sbi: add priority for reset handler
Anup Patel
anup at brainfault.org
Wed Sep 29 20:51:01 PDT 2021
On Wed, Sep 29, 2021 at 3:12 PM Nikita Shubin <nikita.shubin at maquefel.me> wrote:
>
> From: Nikita Shubin <n.shubin at yadro.com>
>
> Let's make system_reset_check returning priority instead of only
> true/false. In that case 0 - means not supported, and anything above
> means priority that makes existing reset handlers being used in first
> place, unless it is decided to lower their priority.
>
> The handler with the least priority wins.
>
> Signed-off-by: Nikita Shubin <n.shubin at yadro.com>
> ---
> lib/sbi/sbi_system.c | 20 ++++++++++++++++----
> 1 file changed, 16 insertions(+), 4 deletions(-)
>
> diff --git a/lib/sbi/sbi_system.c b/lib/sbi/sbi_system.c
> index 92bc0ac..a82c4ef 100644
> --- a/lib/sbi/sbi_system.c
> +++ b/lib/sbi/sbi_system.c
> @@ -55,6 +55,8 @@ void __noreturn sbi_system_reset(u32 reset_type, u32 reset_reason)
Instead of changing sbi_system_reset(), the sbi_system_reset_get_device()
can return "sbi_system_reset_device *" with highest non-zero priority.
> struct sbi_domain *dom = sbi_domain_thishart_ptr();
> struct sbi_scratch *scratch = sbi_scratch_thishart_ptr();
> struct sbi_dlist *pos;
> + struct sbi_system_reset_device *reset_dev = 0;
> + int priority = 255;
>
> /* Send HALT IPI to every hart other than the current hart */
> while (!sbi_hsm_hart_interruptible_mask(dom, hbase, &hmask)) {
> @@ -70,12 +72,22 @@ void __noreturn sbi_system_reset(u32 reset_type, u32 reset_reason)
>
> /* Check each reset device registered for supported reset type */
> sbi_list_for_each(pos, &(reset_devices_list)) {
> - struct sbi_system_reset_device *dev =
> - to_system_reset_device(pos);
> - if (dev->system_reset_check(reset_type, reset_reason))
> - dev->system_reset(reset_type, reset_reason);
> + struct sbi_system_reset_device *dev
> + = to_system_reset_device(pos);
> + int status = dev->system_reset_check(reset_type, reset_reason);
> +
> + if (status == 0)
> + continue;
> +
> + if (status < priority) {
This is counter intuitive because lower value means higher priority.
Instead, I suggest to have following priority semantics:
1) 0 means not support
2) higher value means higher priority
> + reset_dev = dev;
> + priority = status;
> + }
> }
>
> + if (reset_dev)
> + reset_dev->system_reset(reset_type, reset_reason);
> +
> /* If platform specific reset did not work then do sbi_exit() */
> sbi_exit(scratch);
> }
> --
> 2.31.1
>
>
> --
> opensbi mailing list
> opensbi at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/opensbi
Regards,
Anup
More information about the opensbi
mailing list