| Message ID | 20210412073154.25011-1-dinghao.liu@zju.edu.cn (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] spi: spi-zynqmp-gqspi: Fix runtime PM imbalance in zynqmp_qspi_probe | expand |
Hi Dinghao, On Mon, Apr 12, 2021 at 03:31:54PM +0800, Dinghao Liu wrote: > There is a PM usage counter decrement after zynqmp_qspi_init_hw() > without any refcount increment, which leads to refcount leak.Add > a refcount increment to balance the refcount. Also set > auto_runtime_pm to resume suspended spi controller. > > Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn> > --- > changelog: > > v2: - Add a refcount increment to fix refcout leak instead of the > refcount decrement on error. > Set ctlr->auto_runtime_pm = true. > --- > drivers/spi/spi-zynqmp-gqspi.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/spi/spi-zynqmp-gqspi.c b/drivers/spi/spi-zynqmp-gqspi.c > index c8fa6ee18ae7..8b21c7b0e7eb 100644 > --- a/drivers/spi/spi-zynqmp-gqspi.c > +++ b/drivers/spi/spi-zynqmp-gqspi.c > @@ -1160,6 +1160,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) > pm_runtime_set_autosuspend_delay(&pdev->dev, SPI_AUTOSUSPEND_TIMEOUT); > pm_runtime_set_active(&pdev->dev); > pm_runtime_enable(&pdev->dev); > + pm_runtime_get_sync(&pdev->dev); Please check the return value here, if ret is "< 0", goto error label, and a pm_runtime_put_sync is needed in error label > /* QSPI controller initializations */ > zynqmp_qspi_init_hw(xqspi); > > @@ -1187,6 +1188,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) > ctlr->mode_bits = SPI_CPOL | SPI_CPHA | SPI_RX_DUAL | SPI_RX_QUAD | > SPI_TX_DUAL | SPI_TX_QUAD; > ctlr->dev.of_node = np; > + ctlr->auto_runtime_pm = true; > > ret = devm_spi_register_controller(&pdev->dev, ctlr); > if (ret) { These 2 function pm_runtime_mark_last_busy(&pdev->dev); pm_runtime_put_autosuspend(&pdev->dev); are the last operations in probe function since if they runs, spi_controller will enter suspend state and disable clks after 3s passing. So please move them just before "return 0". And would you please cc me when you send V3? I am preparing to send a patch series to fix clk and suspend/resume issues which bases on the pm_runtime issue. Thanks, Quanyang > -- > 2.17.1 >
> Hi Dinghao, > On Mon, Apr 12, 2021 at 03:31:54PM +0800, Dinghao Liu wrote: > > There is a PM usage counter decrement after zynqmp_qspi_init_hw() > > without any refcount increment, which leads to refcount leak.Add > > a refcount increment to balance the refcount. Also set > > auto_runtime_pm to resume suspended spi controller. > > > > Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn> > > --- > > changelog: > > > > v2: - Add a refcount increment to fix refcout leak instead of the > > refcount decrement on error. > > Set ctlr->auto_runtime_pm = true. > > --- > > drivers/spi/spi-zynqmp-gqspi.c | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/drivers/spi/spi-zynqmp-gqspi.c b/drivers/spi/spi-zynqmp-gqspi.c > > index c8fa6ee18ae7..8b21c7b0e7eb 100644 > > --- a/drivers/spi/spi-zynqmp-gqspi.c > > +++ b/drivers/spi/spi-zynqmp-gqspi.c > > @@ -1160,6 +1160,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) > > pm_runtime_set_autosuspend_delay(&pdev->dev, SPI_AUTOSUSPEND_TIMEOUT); > > pm_runtime_set_active(&pdev->dev); > > pm_runtime_enable(&pdev->dev); > > + pm_runtime_get_sync(&pdev->dev); > Please check the return value here, if ret is "< 0", goto error label, > and a pm_runtime_put_sync is needed in error label > > /* QSPI controller initializations */ > > zynqmp_qspi_init_hw(xqspi); > > > > @@ -1187,6 +1188,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) > > ctlr->mode_bits = SPI_CPOL | SPI_CPHA | SPI_RX_DUAL | SPI_RX_QUAD | > > SPI_TX_DUAL | SPI_TX_QUAD; > > ctlr->dev.of_node = np; > > + ctlr->auto_runtime_pm = true; > > > > ret = devm_spi_register_controller(&pdev->dev, ctlr); > > if (ret) { > These 2 function > pm_runtime_mark_last_busy(&pdev->dev); > pm_runtime_put_autosuspend(&pdev->dev); > are the last operations in probe function since if they runs, > spi_controller will enter suspend state and disable clks after 3s > passing. So please move them just before "return 0". > > And would you please cc me when you send V3? I am preparing to send a patch series > to fix clk and suspend/resume issues which bases on the pm_runtime issue. > Thanks for your advice and I will send a new patch soon. Regards, Dinghao
diff --git a/drivers/spi/spi-zynqmp-gqspi.c b/drivers/spi/spi-zynqmp-gqspi.c index c8fa6ee18ae7..8b21c7b0e7eb 100644 --- a/drivers/spi/spi-zynqmp-gqspi.c +++ b/drivers/spi/spi-zynqmp-gqspi.c @@ -1160,6 +1160,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) pm_runtime_set_autosuspend_delay(&pdev->dev, SPI_AUTOSUSPEND_TIMEOUT); pm_runtime_set_active(&pdev->dev); pm_runtime_enable(&pdev->dev); + pm_runtime_get_sync(&pdev->dev); /* QSPI controller initializations */ zynqmp_qspi_init_hw(xqspi); @@ -1187,6 +1188,7 @@ static int zynqmp_qspi_probe(struct platform_device *pdev) ctlr->mode_bits = SPI_CPOL | SPI_CPHA | SPI_RX_DUAL | SPI_RX_QUAD | SPI_TX_DUAL | SPI_TX_QUAD; ctlr->dev.of_node = np; + ctlr->auto_runtime_pm = true; ret = devm_spi_register_controller(&pdev->dev, ctlr); if (ret) {
There is a PM usage counter decrement after zynqmp_qspi_init_hw() without any refcount increment, which leads to refcount leak.Add a refcount increment to balance the refcount. Also set auto_runtime_pm to resume suspended spi controller. Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn> --- changelog: v2: - Add a refcount increment to fix refcout leak instead of the refcount decrement on error. Set ctlr->auto_runtime_pm = true. --- drivers/spi/spi-zynqmp-gqspi.c | 2 ++ 1 file changed, 2 insertions(+)