| Message ID | 20240521-qcom-firmware-name-v1-9-99a6d32b1e5e@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | arm64: qcom: autodetect firmware paths | expand |
On 5/21/2024 2:45 AM, Dmitry Baryshkov wrote: > Make the driver use qcom_fw_helper to autodetect the path to the > calibration data file. > > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> > --- > drivers/remoteproc/qcom_wcnss.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/drivers/remoteproc/qcom_wcnss.c b/drivers/remoteproc/qcom_wcnss.c > index 421a3943a90d..45fc578ae30b 100644 > --- a/drivers/remoteproc/qcom_wcnss.c > +++ b/drivers/remoteproc/qcom_wcnss.c > @@ -23,6 +23,7 @@ > #include <linux/regulator/consumer.h> > #include <linux/remoteproc.h> > #include <linux/soc/qcom/mdt_loader.h> > +#include <linux/soc/qcom/fw_helper.h> > #include <linux/soc/qcom/smem.h> > #include <linux/soc/qcom/smem_state.h> > > @@ -555,8 +556,13 @@ static int wcnss_probe(struct platform_device *pdev) > if (ret < 0 && ret != -EINVAL) > return ret; > > + fw_name = qcom_get_board_fw(fw_name); > + if (!fw_name) > + return -ENOMEM; > + > rproc = devm_rproc_alloc(&pdev->dev, pdev->name, &wcnss_ops, > fw_name, sizeof(*wcnss)); > + kfree(fw_name); > if (!rproc) { > dev_err(&pdev->dev, "unable to allocate remoteproc\n"); > return -ENOMEM; > can you cleanly bisect to this patch? seems it depends upon patch 10. should 09 & 10 be swapped, or perhaps squashed?
On Wed, 22 May 2024 at 22:22, Jeff Johnson <quic_jjohnson@quicinc.com> wrote: > > On 5/21/2024 2:45 AM, Dmitry Baryshkov wrote: > > Make the driver use qcom_fw_helper to autodetect the path to the > > calibration data file. > > > > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> > > --- > > drivers/remoteproc/qcom_wcnss.c | 6 ++++++ > > 1 file changed, 6 insertions(+) > > > > diff --git a/drivers/remoteproc/qcom_wcnss.c b/drivers/remoteproc/qcom_wcnss.c > > index 421a3943a90d..45fc578ae30b 100644 > > --- a/drivers/remoteproc/qcom_wcnss.c > > +++ b/drivers/remoteproc/qcom_wcnss.c > > @@ -23,6 +23,7 @@ > > #include <linux/regulator/consumer.h> > > #include <linux/remoteproc.h> > > #include <linux/soc/qcom/mdt_loader.h> > > +#include <linux/soc/qcom/fw_helper.h> > > #include <linux/soc/qcom/smem.h> > > #include <linux/soc/qcom/smem_state.h> > > > > @@ -555,8 +556,13 @@ static int wcnss_probe(struct platform_device *pdev) > > if (ret < 0 && ret != -EINVAL) > > return ret; > > > > + fw_name = qcom_get_board_fw(fw_name); > > + if (!fw_name) > > + return -ENOMEM; > > + > > rproc = devm_rproc_alloc(&pdev->dev, pdev->name, &wcnss_ops, > > fw_name, sizeof(*wcnss)); > > + kfree(fw_name); > > if (!rproc) { > > dev_err(&pdev->dev, "unable to allocate remoteproc\n"); > > return -ENOMEM; > > > > can you cleanly bisect to this patch? seems it depends upon patch 10. > should 09 & 10 be swapped, or perhaps squashed? Yes. I think I got this mixed during rebasing and squashing of the changes. For v2, if the approach is found to be generally acceptable, I'll squash them.
diff --git a/drivers/remoteproc/qcom_wcnss.c b/drivers/remoteproc/qcom_wcnss.c index 421a3943a90d..45fc578ae30b 100644 --- a/drivers/remoteproc/qcom_wcnss.c +++ b/drivers/remoteproc/qcom_wcnss.c @@ -23,6 +23,7 @@ #include <linux/regulator/consumer.h> #include <linux/remoteproc.h> #include <linux/soc/qcom/mdt_loader.h> +#include <linux/soc/qcom/fw_helper.h> #include <linux/soc/qcom/smem.h> #include <linux/soc/qcom/smem_state.h> @@ -555,8 +556,13 @@ static int wcnss_probe(struct platform_device *pdev) if (ret < 0 && ret != -EINVAL) return ret; + fw_name = qcom_get_board_fw(fw_name); + if (!fw_name) + return -ENOMEM; + rproc = devm_rproc_alloc(&pdev->dev, pdev->name, &wcnss_ops, fw_name, sizeof(*wcnss)); + kfree(fw_name); if (!rproc) { dev_err(&pdev->dev, "unable to allocate remoteproc\n"); return -ENOMEM;
Make the driver use qcom_fw_helper to autodetect the path to the calibration data file. Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> --- drivers/remoteproc/qcom_wcnss.c | 6 ++++++ 1 file changed, 6 insertions(+)