On Thu, 10 Jan 2019 at 19:53, Daniel Thompson
<daniel.thompson@xxxxxxxxxx> wrote:
>
> On Thu, Jan 10, 2019 at 05:54:56PM +0530, Sumit Garg wrote:
> > This flag indicates that requests in this context should not wait for
> > tee-supplicant daemon to be started if not present and just return
> > with an error code. It is needed for requests which should be
> > non-blocking in nature like ones arising from TEE based kernel drivers
> > or any in kernel api that uses TEE internal client interface.
>
> This patch seems out of order in the patch set: Doesn't
> optee_enumarate_devices() require this feature if it is to
> work correctly with a TEE that does not implement the enumeate pTA.
>
> If so better to implement the feature first so that
> optee_emumerate_devices() works correctly when it first appears.
>
Agree, will make this patch as #2 in next version.
> >
> > Signed-off-by: Sumit Garg <sumit.garg@xxxxxxxxxx>
> > ---
> > drivers/tee/optee/supp.c | 10 +++++++++-
> > drivers/tee/tee_core.c | 2 ++
> > include/linux/tee_drv.h | 6 ++++++
> > 3 files changed, 17 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/tee/optee/supp.c b/drivers/tee/optee/supp.c
> > index 43626e1..92f56b8 100644
> > --- a/drivers/tee/optee/supp.c
> > +++ b/drivers/tee/optee/supp.c
> > @@ -88,10 +88,18 @@ u32 optee_supp_thrd_req(struct tee_context *ctx, u32 func, size_t num_params,
> > {
> > struct optee *optee = tee_get_drvdata(ctx->teedev);
> > struct optee_supp *supp = &optee->supp;
> > - struct optee_supp_req *req = kzalloc(sizeof(*req), GFP_KERNEL);
> > + struct optee_supp_req *req;
> > bool interruptable;
> > u32 ret;
> >
> > + /*
> > + * Return in case there is no supplicant available and
> > + * non-blocking request.
> > + */
> > + if (!supp->ctx && ctx->supp_nowait)
> > + return TEEC_ERROR_COMMUNICATION;
> > +
> > + req = kzalloc(sizeof(*req), GFP_KERNEL);
> > if (!req)
> > return TEEC_ERROR_OUT_OF_MEMORY;
> >
> > diff --git a/drivers/tee/tee_core.c b/drivers/tee/tee_core.c
> > index 9ddb89e..5d6c317 100644
> > --- a/drivers/tee/tee_core.c
> > +++ b/drivers/tee/tee_core.c
> > @@ -105,6 +105,7 @@ static int tee_open(struct inode *inode, struct file *filp)
> > if (IS_ERR(ctx))
> > return PTR_ERR(ctx);
> >
> > + ctx->supp_nowait = false;
> > filp->private_data = ctx;
> > return 0;
> > }
> > @@ -981,6 +982,7 @@ tee_client_open_context(struct tee_context *start,
> > } while (IS_ERR(ctx) && PTR_ERR(ctx) != -ENOMEM);
> >
> > put_device(put_dev);
> > + ctx->supp_nowait = true;
>
> Why automatically set supp_nowait inside open_context() ?
>
I think this is the default behaviour (non-blocking request) that any
in kernel client would expect. Also this flag could be configured
again before call to open_session() if any in kernel client requires
different behaviour.
-Sumit
>
> > return ctx;
> > }
> > EXPORT_SYMBOL_GPL(tee_client_open_context);
> > diff --git a/include/linux/tee_drv.h b/include/linux/tee_drv.h
> > index ed16bf1..fe1a920 100644
> > --- a/include/linux/tee_drv.h
> > +++ b/include/linux/tee_drv.h
> > @@ -49,6 +49,11 @@ struct tee_shm_pool;
> > * @releasing: flag that indicates if context is being released right now.
> > * It is needed to break circular dependency on context during
> > * shared memory release.
> > + * @supp_nowait: flag that indicates that requests in this context should not
> > + * wait for tee-supplicant daemon to be started if not present
> > + * and just return with an error code. It is needed for requests
> > + * that arises from TEE based kernel drivers that should be
> > + * non-blocking in nature.
> > */
> > struct tee_context {
> > struct tee_device *teedev;
> > @@ -56,6 +61,7 @@ struct tee_context {
> > void *data;
> > struct kref refcount;
> > bool releasing;
> > + bool supp_nowait;
> > };
> >
> > struct tee_param_memref {
> > --
> > 2.7.4
> >
