Hello Patrick,
On Fri, 11 Feb 2022 at 14:46, Patrick DELAUNAY patrick.delaunay@foss.st.com wrote:
Hi Etienne,
On 2/11/22 13:07, Etienne Carriere wrote:
As per DT bindings since Linux kernel v5.14, the device tree can define only 1 SCMI agent node that is named scmi [1]. As a consequence, change implementation of the SCMI driver test through sandbox architecture to reflect that.
This change updates sandbox test DT and sandbox SCMI driver accordingly since all these are impacted.
Cc: Simon Glass sjg@chromium.org Signed-off-by: Etienne Carriere etienne.carriere@linaro.org
arch/sandbox/dts/test.dts | 37 ++-- arch/sandbox/include/asm/scmi_test.h | 12 +- drivers/firmware/scmi/sandbox-scmi_agent.c | 167 ++++++------------- drivers/firmware/scmi/sandbox-scmi_devices.c | 4 +- test/dm/scmi.c | 117 +++++++------ 5 files changed, 129 insertions(+), 208 deletions(-) (snip)
static int sandbox_scmi_test_probe(struct udevice *dev) {
static const char basename[] = "sandbox-scmi-agent@";
static const char basename[] = "scmi"; struct sandbox_scmi_agent *agent = dev_get_priv(dev);
const size_t basename_size = sizeof(basename) - 1;if (strncmp(basename, dev->name, basename_size))
if (strcmp(basename, dev->name)) return -ENOENT;This check is still needed ?
you test in drivier if device is correct
- static const char basename[] = "scmi";
....
- if (strcmp(basename, dev->name))
return -ENOENT;
Indeed I missed that update. Thanks, i'll fix.
(snip)
diff --git a/test/dm/scmi.c b/test/dm/scmi.c index c938e6d4fc..2f19b99d7c 100644 --- a/test/dm/scmi.c +++ b/test/dm/scmi.c @@ -5,7 +5,7 @@
- Tests scmi_agent uclass and the SCMI drivers implemented in other
- uclass devices probe when a SCMI server exposes resources.
- Note in test.dts the protocol@10 node in agent 1. Protocol 0x10 is not
- Note in test.dts the protocol@10 node in scmi node. Protocol 0x10 is not
- implemented in U-Boot SCMI components but the implementation is exepected
- to not complain on unknown protocol IDs, as long as it is not used. Note
- in test.dts tests that SCMI drivers probing does not fail for such an
@@ -28,8 +28,7 @@ static int ut_assert_scmi_state_preprobe(struct unit_test_state *uts) struct sandbox_scmi_service *scmi_ctx = sandbox_scmi_service_ctx();
ut_assertnonnull(scmi_ctx);
if (scmi_ctx->agent_count)ut_asserteq(2, scmi_ctx->agent_count);
ut_assertnull(scmi_ctx->agent); return 0;}
@@ -39,35 +38,28 @@ static int ut_assert_scmi_state_postprobe(struct unit_test_state *uts, { struct sandbox_scmi_devices *scmi_devices; struct sandbox_scmi_service *scmi_ctx;
struct sandbox_scmi_agent *agent0;struct sandbox_scmi_agent *agent1;
struct sandbox_scmi_agent *agent; /* Device references to check context against test sequence */ scmi_devices = sandbox_scmi_devices_ctx(dev);
ut_assertnonnull(scmi_devices);ut_asserteq(3, scmi_devices->clk_count);
added line ?
will remove
ut_asserteq(2, scmi_devices->clk_count); ut_asserteq(1, scmi_devices->reset_count); ut_asserteq(2, scmi_devices->regul_count); /* State of the simulated SCMI server exposed */ scmi_ctx = sandbox_scmi_service_ctx();
agent0 = scmi_ctx->agent[0];agent1 = scmi_ctx->agent[1];ut_asserteq(2, scmi_ctx->agent_count);ut_assertnonnull(agent0);ut_asserteq(2, agent0->clk_count);ut_assertnonnull(agent0->clk);ut_asserteq(1, agent0->reset_count);ut_assertnonnull(agent0->reset);ut_asserteq(2, agent0->voltd_count);ut_assertnonnull(agent0->voltd);
ut_assertnonnull(scmi_ctx);agent = scmi_ctx->agent;ut_assertnonnull(agent);line to remove
ok
ut_assertnonnull(agent1);ut_assertnonnull(agent1->clk);ut_asserteq(1, agent1->clk_count);
ut_asserteq(2, agent->clk_count);ut_assertnonnull(agent->clk);ut_asserteq(1, agent->reset_count);ut_assertnonnull(agent->reset);ut_asserteq(2, agent->voltd_count);ut_assertnonnull(agent->voltd); return 0;}
(snip)
with few minor remarks
Reviewed-by: Patrick Delaunay patrick.delaunay@foss.st.com
I'll send the fixes. Many thanks.
Best regards, Etienne
Thanks Patrick