Commit 1efa30d0 authored by Sarangdhar Joshi's avatar Sarangdhar Joshi Committed by Bjorn Andersson

remoteproc: Introduce rproc_{start,stop}() functions

In the context of recovering from crash, rproc_trigger_recovery() does
rproc_shutdown() followed by rproc_boot(). The remoteproc resources are
cleaned up in rproc_shutdown() and immediately reallocated in
rproc_boot() which is an unnecessary overhead. Furthermore, we want the
memory regions to be accessible after stopping the remote processor, to
be able to extract the memory content for a coredump.

This patch factors out the code in rproc_boot() and rproc_shutdown()
path and introduces rproc_{start,stop}() in order to avoid resource
allocation overhead.
Signed-off-by: default avatarSarangdhar Joshi <spjoshi@codeaurora.org>
Signed-off-by: default avatarBjorn Andersson <bjorn.andersson@linaro.org>
parent 2ea659a9
......@@ -847,6 +847,63 @@ static void rproc_resource_cleanup(struct rproc *rproc)
kref_put(&rvdev->refcount, rproc_vdev_release);
}
static int rproc_start(struct rproc *rproc, const struct firmware *fw)
{
struct resource_table *table, *loaded_table;
struct device *dev = &rproc->dev;
int ret, tablesz;
/* look for the resource table */
table = rproc_find_rsc_table(rproc, fw, &tablesz);
if (!table) {
dev_err(dev, "Resouce table look up failed\n");
return -EINVAL;
}
/* load the ELF segments to memory */
ret = rproc_load_segments(rproc, fw);
if (ret) {
dev_err(dev, "Failed to load program segments: %d\n", ret);
return ret;
}
/*
* The starting device has been given the rproc->cached_table as the
* resource table. The address of the vring along with the other
* allocated resources (carveouts etc) is stored in cached_table.
* In order to pass this information to the remote device we must copy
* this information to device memory. We also update the table_ptr so
* that any subsequent changes will be applied to the loaded version.
*/
loaded_table = rproc_find_loaded_rsc_table(rproc, fw);
if (loaded_table) {
memcpy(loaded_table, rproc->cached_table, tablesz);
rproc->table_ptr = loaded_table;
}
/* power up the remote processor */
ret = rproc->ops->start(rproc);
if (ret) {
dev_err(dev, "can't start rproc %s: %d\n", rproc->name, ret);
return ret;
}
/* probe any subdevices for the remote processor */
ret = rproc_probe_subdevices(rproc);
if (ret) {
dev_err(dev, "failed to probe subdevices for %s: %d\n",
rproc->name, ret);
rproc->ops->stop(rproc);
return ret;
}
rproc->state = RPROC_RUNNING;
dev_info(dev, "remote processor %s is now up\n", rproc->name);
return 0;
}
/*
* take a firmware and boot a remote processor with it.
*/
......@@ -854,7 +911,7 @@ static int rproc_fw_boot(struct rproc *rproc, const struct firmware *fw)
{
struct device *dev = &rproc->dev;
const char *name = rproc->firmware;
struct resource_table *table, *loaded_table;
struct resource_table *table;
int ret, tablesz;
ret = rproc_fw_sanity_check(rproc, fw);
......@@ -905,50 +962,12 @@ static int rproc_fw_boot(struct rproc *rproc, const struct firmware *fw)
goto clean_up_resources;
}
/* load the ELF segments to memory */
ret = rproc_load_segments(rproc, fw);
if (ret) {
dev_err(dev, "Failed to load program segments: %d\n", ret);
goto clean_up_resources;
}
/*
* The starting device has been given the rproc->cached_table as the
* resource table. The address of the vring along with the other
* allocated resources (carveouts etc) is stored in cached_table.
* In order to pass this information to the remote device we must copy
* this information to device memory. We also update the table_ptr so
* that any subsequent changes will be applied to the loaded version.
*/
loaded_table = rproc_find_loaded_rsc_table(rproc, fw);
if (loaded_table) {
memcpy(loaded_table, rproc->cached_table, tablesz);
rproc->table_ptr = loaded_table;
}
/* power up the remote processor */
ret = rproc->ops->start(rproc);
if (ret) {
dev_err(dev, "can't start rproc %s: %d\n", rproc->name, ret);
ret = rproc_start(rproc, fw);
if (ret)
goto clean_up_resources;
}
/* probe any subdevices for the remote processor */
ret = rproc_probe_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 up\n", rproc->name);
return 0;
stop_rproc:
rproc->ops->stop(rproc);
clean_up_resources:
rproc_resource_cleanup(rproc);
clean_up:
......@@ -994,6 +1013,32 @@ static int rproc_trigger_auto_boot(struct rproc *rproc)
return ret;
}
static int rproc_stop(struct rproc *rproc)
{
struct device *dev = &rproc->dev;
int ret;
/* remove any subdevices for the remote processor */
rproc_remove_subdevices(rproc);
/* power off the remote processor */
ret = rproc->ops->stop(rproc);
if (ret) {
dev_err(dev, "can't stop rproc: %d\n", ret);
return ret;
}
/* if in crash state, unlock crash handler */
if (rproc->state == RPROC_CRASHED)
complete_all(&rproc->crash_comp);
rproc->state = RPROC_OFFLINE;
dev_info(dev, "stopped remote processor %s\n", rproc->name);
return 0;
}
/**
* rproc_trigger_recovery() - recover a remoteproc
* @rproc: the remote processor
......@@ -1163,14 +1208,9 @@ void rproc_shutdown(struct rproc *rproc)
if (!atomic_dec_and_test(&rproc->power))
goto out;
/* remove any subdevices for the remote processor */
rproc_remove_subdevices(rproc);
/* power off the remote processor */
ret = rproc->ops->stop(rproc);
ret = rproc_stop(rproc);
if (ret) {
atomic_inc(&rproc->power);
dev_err(dev, "can't stop rproc: %d\n", ret);
goto out;
}
......@@ -1183,15 +1223,6 @@ void rproc_shutdown(struct rproc *rproc)
kfree(rproc->cached_table);
rproc->cached_table = NULL;
rproc->table_ptr = NULL;
/* if in crash state, unlock crash handler */
if (rproc->state == RPROC_CRASHED)
complete_all(&rproc->crash_comp);
rproc->state = RPROC_OFFLINE;
dev_info(dev, "stopped remote processor %s\n", rproc->name);
out:
mutex_unlock(&rproc->lock);
}
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment