Re: [PATCH 04/12] habanalabs: add unsupported functions

[Oded Gabbay]

On Tue, Jun 28, 2022 at 12:12 PM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Tue, Jun 28, 2022 at 11:21:24AM +0300, Oded Gabbay wrote:
> > On Tue, Jun 28, 2022 at 9:34 AM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > >
> > > On Mon, Jun 27, 2022 at 11:26:12PM +0300, Oded Gabbay wrote:
> > > > There are a number of new ASIC-specific functions that were added
> > > > for Gaudi2. To make the common code work, we need to define empty
> > > > implementations of those functions for Goya and Gaudi.
> > > >
> > > > Some functions will return error if called with Goya/Gaudi.
> > > >
> > > > Signed-off-by: Oded Gabbay <ogabbay@xxxxxxxxxx>
> > > > ---
> > > >  drivers/misc/habanalabs/gaudi/gaudi.c | 24 ++++++++++++++++++++++++
> > > >  drivers/misc/habanalabs/goya/goya.c   | 24 ++++++++++++++++++++++++
> > > >  2 files changed, 48 insertions(+)
> > > >
> > > > diff --git a/drivers/misc/habanalabs/gaudi/gaudi.c b/drivers/misc/habanalabs/gaudi/gaudi.c
> > > > index ae894335e9f8..f4581220ecd5 100644
> > > > --- a/drivers/misc/habanalabs/gaudi/gaudi.c
> > > > +++ b/drivers/misc/habanalabs/gaudi/gaudi.c
> > > > @@ -8588,6 +8588,11 @@ static void gaudi_ctx_fini(struct hl_ctx *ctx)
> > > >       gaudi_internal_cb_pool_fini(ctx->hdev, ctx);
> > > >  }
> > > >
> > > > +int gaudi_pre_schedule_cs(struct hl_cs *cs)
> > > > +{
> > > > +     return 0;
> > > > +}
> > > > +
> > > >  static u32 gaudi_get_queue_id_for_cq(struct hl_device *hdev, u32 cq_idx)
> > > >  {
> > > >       return gaudi_cq_assignment[cq_idx];
> > > > @@ -8959,6 +8964,14 @@ static void gaudi_enable_events_from_fw(struct hl_device *hdev)
> > > >               gaudi_irq_map_table[GAUDI_EVENT_INTS_REGISTER].cpu_id);
> > > >  }
> > > >
> > > > +int gaudi_ack_mmu_page_fault_or_access_error(struct hl_device *hdev,
> > > > +                                                     u64 mmu_cap_mask)
> > > > +{
> > > > +     dev_err(hdev->dev, "mmu_error function is not supported\n");
> > >
> > > Can userspace trigger this?  if so, make it debug, as you don't want to
> > > give userspace a way to spam the logs.
> >
> > Only via a debugfs node which is exposed only to root user.
> > What is your recommendation in that case ?
>
> What does this error message help out with?  Why not just return an
> error and be done with it?  No need to spam the kernel log, right?
Sure, np.  I'm with you on that.

Thanks,
Oded
>
> thanks,
>
> greg k-h