On Wed 28 Sep 03:42 PDT 2016, Sricharan R wrote:
> Patch [1] fixes the issues with interrupted SCM64 calls
> by returning the register r6 (Session ID) on the next try, but register
> r0 also needs to be preserved for the retry. r0 contains the
> result of the previous try. Without this i see that the SCM calls
> hang when retried.
>
> [1] https://patchwork.kernel.org/patch/9291589/
>
Andy, this is needed as well on 8996 - or I get rcu stalls during
firmware loading operations (so I guess it's busy spinning in that
loop).
Tested-by: Bjorn Andersson <bjorn.anders...@linaro.org>
Regards,
Bjorn
> Signed-off-by: Sricharan R <sricha...@codeaurora.org>
> ---
> drivers/firmware/qcom_scm-64.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/firmware/qcom_scm-64.c b/drivers/firmware/qcom_scm-64.c
> index 7ecd0e7..28b604e 100755
> --- a/drivers/firmware/qcom_scm-64.c
> +++ b/drivers/firmware/qcom_scm-64.c
> @@ -134,6 +134,9 @@ static int qcom_scm_call(struct device *dev, u32 svc_id,
> u32 cmd_id,
> res->a6 = 0;
>
> do {
> + if (res->a0 == QCOM_SCM_INTERRUPTED)
> + cmd = res->a0;
> +
> arm_smccc_smc(cmd, desc->arginfo, desc->args[0],
> desc->args[1], desc->args[2], x5,
> res->a6, 0,
> res);
> --
> QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of
> Code Aurora Forum, hosted by The Linux Foundation
>
