| Message ID | 20200601175139.22097-4-mathieu.poirier@linaro.org (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | remoteproc: Add support for attaching with rproc | expand |
On Mon 01 Jun 10:51 PDT 2020, Mathieu Poirier wrote: > Introducing function rproc_attach() to enact the same actions as > rproc_start(), but without the steps related to the handling of > a firmware image. That way we can properly deal with scenarios > where the remoteproc core needs to attach with a remote processsor > that has been booted by another entity. > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > --- > drivers/remoteproc/remoteproc_core.c | 42 ++++++++++++++++++++++++++++ > 1 file changed, 42 insertions(+) > > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 9f04c30c4aaf..0b323f6b554b 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -1370,6 +1370,48 @@ static int rproc_start(struct rproc *rproc, const struct firmware *fw) > return ret; > } > > +static int __maybe_unused rproc_attach(struct rproc *rproc) > +{ > + struct device *dev = &rproc->dev; > + int ret; > + For the case where we're going DETACHED -> RUNNING - > OFFLINE we need to consider the pm_runtime (and prepare/unprepare) state of the device as well... Apart from that I think this looks good. Regards, Bjorn > + ret = rproc_prepare_subdevices(rproc); > + if (ret) { > + dev_err(dev, "failed to prepare subdevices for %s: %d\n", > + rproc->name, ret); > + goto out; > + } > + > + /* Attach to the remote processor */ > + ret = rproc_attach_device(rproc); > + if (ret) { > + dev_err(dev, "can't attach to rproc %s: %d\n", > + rproc->name, ret); > + goto unprepare_subdevices; > + } > + > + /* Start any subdevices for the remote processor */ > + ret = rproc_start_subdevices(rproc); > + if (ret) { > + dev_err(dev, "failed to probe subdevices for %s: %d\n", > + rproc->name, ret); > + goto stop_rproc; > + } > + > + rproc->state = RPROC_RUNNING; > + > + dev_info(dev, "remote processor %s is now attached\n", rproc->name); > + > + return 0; > + > +stop_rproc: > + rproc->ops->stop(rproc); > +unprepare_subdevices: > + rproc_unprepare_subdevices(rproc); > +out: > + return ret; > +} > + > /* > * take a firmware and boot a remote processor with it. > */ > -- > 2.20.1 >
On Mon 22 Jun 00:07 PDT 2020, Bjorn Andersson wrote: > On Mon 01 Jun 10:51 PDT 2020, Mathieu Poirier wrote: > > > Introducing function rproc_attach() to enact the same actions as > > rproc_start(), but without the steps related to the handling of > > a firmware image. That way we can properly deal with scenarios > > where the remoteproc core needs to attach with a remote processsor > > that has been booted by another entity. > > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > > --- > > drivers/remoteproc/remoteproc_core.c | 42 ++++++++++++++++++++++++++++ > > 1 file changed, 42 insertions(+) > > > > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > > index 9f04c30c4aaf..0b323f6b554b 100644 > > --- a/drivers/remoteproc/remoteproc_core.c > > +++ b/drivers/remoteproc/remoteproc_core.c > > @@ -1370,6 +1370,48 @@ static int rproc_start(struct rproc *rproc, const struct firmware *fw) > > return ret; > > } > > > > +static int __maybe_unused rproc_attach(struct rproc *rproc) > > +{ > > + struct device *dev = &rproc->dev; > > + int ret; > > + > > For the case where we're going DETACHED -> RUNNING - > OFFLINE we > need to consider the pm_runtime (and prepare/unprepare) state of the > device as well... > Missed that you do indeed pm_runtime_get() in the calling function, so I think we're good on that part. Still need how to actually implement that (and the prepare/unprepare), in particular if we're moving into detaching a remoteproc. > > Apart from that I think this looks good. > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> Regards, Bjorn > Regards, > Bjorn > > > + ret = rproc_prepare_subdevices(rproc); > > + if (ret) { > > + dev_err(dev, "failed to prepare subdevices for %s: %d\n", > > + rproc->name, ret); > > + goto out; > > + } > > + > > + /* Attach to the remote processor */ > > + ret = rproc_attach_device(rproc); > > + if (ret) { > > + dev_err(dev, "can't attach to rproc %s: %d\n", > > + rproc->name, ret); > > + goto unprepare_subdevices; > > + } > > + > > + /* Start any subdevices for the remote processor */ > > + ret = rproc_start_subdevices(rproc); > > + if (ret) { > > + dev_err(dev, "failed to probe subdevices for %s: %d\n", > > + rproc->name, ret); > > + goto stop_rproc; > > + } > > + > > + rproc->state = RPROC_RUNNING; > > + > > + dev_info(dev, "remote processor %s is now attached\n", rproc->name); > > + > > + return 0; > > + > > +stop_rproc: > > + rproc->ops->stop(rproc); > > +unprepare_subdevices: > > + rproc_unprepare_subdevices(rproc); > > +out: > > + return ret; > > +} > > + > > /* > > * take a firmware and boot a remote processor with it. > > */ > > -- > > 2.20.1 > >
Hi, On Mon, Jun 22, 2020 at 12:18:04AM -0700, Bjorn Andersson wrote: > On Mon 22 Jun 00:07 PDT 2020, Bjorn Andersson wrote: > > > On Mon 01 Jun 10:51 PDT 2020, Mathieu Poirier wrote: > > > > > Introducing function rproc_attach() to enact the same actions as > > > rproc_start(), but without the steps related to the handling of > > > a firmware image. That way we can properly deal with scenarios > > > where the remoteproc core needs to attach with a remote processsor > > > that has been booted by another entity. > > > > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > > > --- > > > drivers/remoteproc/remoteproc_core.c | 42 ++++++++++++++++++++++++++++ > > > 1 file changed, 42 insertions(+) > > > > > > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > > > index 9f04c30c4aaf..0b323f6b554b 100644 > > > --- a/drivers/remoteproc/remoteproc_core.c > > > +++ b/drivers/remoteproc/remoteproc_core.c > > > @@ -1370,6 +1370,48 @@ static int rproc_start(struct rproc *rproc, const struct firmware *fw) > > > return ret; > > > } > > > > > > +static int __maybe_unused rproc_attach(struct rproc *rproc) > > > +{ > > > + struct device *dev = &rproc->dev; > > > + int ret; > > > + > > > > For the case where we're going DETACHED -> RUNNING - > OFFLINE we > > need to consider the pm_runtime (and prepare/unprepare) state of the > > device as well... > > > > Missed that you do indeed pm_runtime_get() in the calling function, so I > think we're good on that part. Still need how to actually implement > that (and the prepare/unprepare), in particular if we're moving into > detaching a remoteproc. I had planned to look at the interaction between the PM runtime and prepare/unprepare callbacks later today. Depending on what I find I may end up modifying this patch... > > > > > Apart from that I think this looks good. > > > > Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org> > > Regards, > Bjorn > > > Regards, > > Bjorn > > > > > + ret = rproc_prepare_subdevices(rproc); > > > + if (ret) { > > > + dev_err(dev, "failed to prepare subdevices for %s: %d\n", > > > + rproc->name, ret); > > > + goto out; > > > + } > > > + > > > + /* Attach to the remote processor */ > > > + ret = rproc_attach_device(rproc); > > > + if (ret) { > > > + dev_err(dev, "can't attach to rproc %s: %d\n", > > > + rproc->name, ret); > > > + goto unprepare_subdevices; > > > + } > > > + > > > + /* Start any subdevices for the remote processor */ > > > + ret = rproc_start_subdevices(rproc); > > > + if (ret) { > > > + dev_err(dev, "failed to probe subdevices for %s: %d\n", > > > + rproc->name, ret); > > > + goto stop_rproc; > > > + } > > > + > > > + rproc->state = RPROC_RUNNING; > > > + > > > + dev_info(dev, "remote processor %s is now attached\n", rproc->name); > > > + > > > + return 0; > > > + > > > +stop_rproc: > > > + rproc->ops->stop(rproc); > > > +unprepare_subdevices: > > > + rproc_unprepare_subdevices(rproc); > > > +out: > > > + return ret; > > > +} > > > + > > > /* > > > * take a firmware and boot a remote processor with it. > > > */ > > > -- > > > 2.20.1 > > >
diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 9f04c30c4aaf..0b323f6b554b 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -1370,6 +1370,48 @@ static int rproc_start(struct rproc *rproc, const struct firmware *fw) return ret; } +static int __maybe_unused rproc_attach(struct rproc *rproc) +{ + struct device *dev = &rproc->dev; + int ret; + + ret = rproc_prepare_subdevices(rproc); + if (ret) { + dev_err(dev, "failed to prepare subdevices for %s: %d\n", + rproc->name, ret); + goto out; + } + + /* Attach to the remote processor */ + ret = rproc_attach_device(rproc); + if (ret) { + dev_err(dev, "can't attach to rproc %s: %d\n", + rproc->name, ret); + goto unprepare_subdevices; + } + + /* Start any subdevices for the remote processor */ + ret = rproc_start_subdevices(rproc); + if (ret) { + dev_err(dev, "failed to probe subdevices for %s: %d\n", + rproc->name, ret); + goto stop_rproc; + } + + rproc->state = RPROC_RUNNING; + + dev_info(dev, "remote processor %s is now attached\n", rproc->name); + + return 0; + +stop_rproc: + rproc->ops->stop(rproc); +unprepare_subdevices: + rproc_unprepare_subdevices(rproc); +out: + return ret; +} + /* * take a firmware and boot a remote processor with it. */
Introducing function rproc_attach() to enact the same actions as rproc_start(), but without the steps related to the handling of a firmware image. That way we can properly deal with scenarios where the remoteproc core needs to attach with a remote processsor that has been booted by another entity. Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> --- drivers/remoteproc/remoteproc_core.c | 42 ++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+)