On Fri, Jul 05, 2024 at 11:55:52AM -0700, Elliot Berman wrote:
> On Fri, Jul 05, 2024 at 05:36:23PM +0530, Mukesh Ojha wrote:
> > SDI is enabled for most of the Qualcomm SoCs and as per commit
> > ff4aa3bc9825 ("firmware: qcom_scm: disable SDI if required")
> > it was recommended to disable SDI by mentioning it in device tree
> >
> > However, for some cases download mode tcsr register already configured
> > from boot firmware to collect dumps and in such cases if download
> > mode is set to zero(nodump mode) from kernel side and SDI is disabled
> ^^^^^^^^^^^^^
> that's not what download_mode=0 does currently, but it's what you're
> proposing it should be doing. I think that proposal is reasonable, but
> can you call out in the commit text that the current behavior of
> download_mode=0 is nop.
Correct..
>
> > via means of mentioning it in device tree we could end up with dump
> > collection.
> >
> > To disable complete dump collection mode, SDI and dload mode register
> > need to be set no dump mode.
> >
> > Signed-off-by: Mukesh Ojha <quic_mojha@xxxxxxxxxxx>
>
> Should this be?
>
> Fixes: ff4aa3bc9825 ("firmware: qcom_scm: disable SDI if required")
Yes, SDI disablement can still result in dump collection and can cause
a hang.
>
> > ---
> > drivers/firmware/qcom/qcom_scm.c | 6 ++----
> > 1 file changed, 2 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
> > index 00c379a3cceb..2e10f75a9cfd 100644
> > --- a/drivers/firmware/qcom/qcom_scm.c
> > +++ b/drivers/firmware/qcom/qcom_scm.c
> > @@ -1954,14 +1954,12 @@ static int qcom_scm_probe(struct platform_device *pdev)
> > * will cause the boot stages to enter download mode, unless
> > * disabled below by a clean shutdown/reboot.
> > */
> > - if (download_mode)
> > - qcom_scm_set_download_mode(true);
> > -
> > + qcom_scm_set_download_mode(download_mode ? true : false);
>
> Just: qcom_scm_set_download_mode(download_mode);
Sure.
-Mukesh
