| Message ID | 20200202222854.18409-15-digetx@gmail.com |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | NVIDIA Tegra APB DMA driver fixes and improvements | expand |
On 02/02/2020 22:28, Dmitry Osipenko wrote: > It's a bit impractical to enable hardware's clock at the time of DMA > channel's allocation because most of DMA client drivers allocate DMA > channel at the time of the driver's probing, and thus, DMA clock is kept > always-enabled in practice, defeating the whole purpose of runtime PM. > > Signed-off-by: Dmitry Osipenko <digetx@gmail.com> > --- > drivers/dma/tegra20-apb-dma.c | 35 ++++++++++++++++++++++++----------- > 1 file changed, 24 insertions(+), 11 deletions(-) What about something like ... diff --git a/drivers/dma/tegra20-apb-dma.c b/drivers/dma/tegra20-apb-dma.c index 22b88ccff05d..d60532f19a43 100644 --- a/drivers/dma/tegra20-apb-dma.c +++ b/drivers/dma/tegra20-apb-dma.c @@ -570,6 +570,7 @@ static bool handle_continuous_head_request(struct tegra_dma_channel *tdc, if (list_empty(&tdc->pending_sg_req)) { dev_err(tdc2dev(tdc), "DMA is running without req\n"); tegra_dma_stop(tdc); + pm_runtime_put(tdc->tdma->dev); return false; } @@ -581,6 +582,7 @@ static bool handle_continuous_head_request(struct tegra_dma_channel *tdc, hsgreq = list_first_entry(&tdc->pending_sg_req, typeof(*hsgreq), node); if (!hsgreq->configured) { tegra_dma_stop(tdc); + pm_runtime_put(tdc->tdma->dev); dev_err(tdc2dev(tdc), "Error in DMA transfer, aborting DMA\n"); tegra_dma_abort_all(tdc); return false; @@ -616,9 +618,14 @@ static void handle_once_dma_done(struct tegra_dma_channel *tdc, list_add_tail(&sgreq->node, &tdc->free_sg_req); /* Do not start DMA if it is going to be terminate */ - if (to_terminate || list_empty(&tdc->pending_sg_req)) + if (to_terminate) return; + if (list_empty(&tdc->pending_sg_req)) { + pm_runtime_put(tdc->tdma->dev); + return; + } + tdc_start_head_req(tdc); } @@ -731,6 +738,10 @@ static void tegra_dma_issue_pending(struct dma_chan *dc) goto end; } if (!tdc->busy) { + err = pm_runtime_get_sync(tdc->tdma->dev); + if (err < 0) + return err; + tdc_start_head_req(tdc); /* Continuous single mode: Configure next req */ @@ -786,6 +797,8 @@ static int tegra_dma_terminate_all(struct dma_chan *dc) } tegra_dma_resume(tdc); + pm_runtime_put(tdc->tdma->dev); + skip_dma_stop: tegra_dma_abort_all(tdc); @@ -1280,22 +1293,15 @@ tegra_dma_prep_dma_cyclic(struct dma_chan *dc, dma_addr_t buf_addr, static int tegra_dma_alloc_chan_resources(struct dma_chan *dc) { struct tegra_dma_channel *tdc = to_tegra_dma_chan(dc); - struct tegra_dma *tdma = tdc->tdma; - int ret; dma_cookie_init(&tdc->dma_chan); - ret = pm_runtime_get_sync(tdma->dev); - if (ret < 0) - return ret; - return 0; } static void tegra_dma_free_chan_resources(struct dma_chan *dc) { struct tegra_dma_channel *tdc = to_tegra_dma_chan(dc); - struct tegra_dma *tdma = tdc->tdma; struct tegra_dma_desc *dma_desc; struct tegra_dma_sg_req *sg_req; struct list_head dma_desc_list; @@ -1328,7 +1334,6 @@ static void tegra_dma_free_chan_resources(struct dma_chan *dc) list_del(&sg_req->node); kfree(sg_req); } - pm_runtime_put(tdma->dev); tdc->slave_id = TEGRA_APBDMA_SLAVE_ID_INVALID; }
06.02.2020 16:50, Jon Hunter пишет: > > On 02/02/2020 22:28, Dmitry Osipenko wrote: >> It's a bit impractical to enable hardware's clock at the time of DMA >> channel's allocation because most of DMA client drivers allocate DMA >> channel at the time of the driver's probing, and thus, DMA clock is kept >> always-enabled in practice, defeating the whole purpose of runtime PM. >> >> Signed-off-by: Dmitry Osipenko <digetx@gmail.com> >> --- >> drivers/dma/tegra20-apb-dma.c | 35 ++++++++++++++++++++++++----------- >> 1 file changed, 24 insertions(+), 11 deletions(-) > What about something like ...... > @@ -581,6 +582,7 @@ static bool handle_continuous_head_request(struct tegra_dma_channel *tdc, > hsgreq = list_first_entry(&tdc->pending_sg_req, typeof(*hsgreq), node); > if (!hsgreq->configured) { > tegra_dma_stop(tdc); > + pm_runtime_put(tdc->tdma->dev); > dev_err(tdc2dev(tdc), "Error in DMA transfer, aborting DMA\n"); > tegra_dma_abort_all(tdc); > return false; Yes, that it's what you suggested to do in the reply to v6. Alright, I'll drop v7 patch #13 and add the put to this patch #14.
On 06/02/2020 14:31, Dmitry Osipenko wrote: > 06.02.2020 16:50, Jon Hunter пишет: >> >> On 02/02/2020 22:28, Dmitry Osipenko wrote: >>> It's a bit impractical to enable hardware's clock at the time of DMA >>> channel's allocation because most of DMA client drivers allocate DMA >>> channel at the time of the driver's probing, and thus, DMA clock is kept >>> always-enabled in practice, defeating the whole purpose of runtime PM. >>> >>> Signed-off-by: Dmitry Osipenko <digetx@gmail.com> >>> --- >>> drivers/dma/tegra20-apb-dma.c | 35 ++++++++++++++++++++++++----------- >>> 1 file changed, 24 insertions(+), 11 deletions(-) >> What about something like ...... >> @@ -581,6 +582,7 @@ static bool handle_continuous_head_request(struct tegra_dma_channel *tdc, >> hsgreq = list_first_entry(&tdc->pending_sg_req, typeof(*hsgreq), node); >> if (!hsgreq->configured) { >> tegra_dma_stop(tdc); >> + pm_runtime_put(tdc->tdma->dev); >> dev_err(tdc2dev(tdc), "Error in DMA transfer, aborting DMA\n"); >> tegra_dma_abort_all(tdc); >> return false; > > Yes, that it's what you suggested to do in the reply to v6. > > Alright, I'll drop v7 patch #13 and add the put to this patch #14. Yes I was not sure if it was clear. However, this seems a bit of an easier change and should keep the 'busy' status somewhat consistent with the rpm state. Jon
diff --git a/drivers/dma/tegra20-apb-dma.c b/drivers/dma/tegra20-apb-dma.c index 50abce608318..92535b962e64 100644 --- a/drivers/dma/tegra20-apb-dma.c +++ b/drivers/dma/tegra20-apb-dma.c @@ -605,9 +605,14 @@ static void handle_once_dma_done(struct tegra_dma_channel *tdc, list_add_tail(&sgreq->node, &tdc->free_sg_req); /* Do not start DMA if it is going to be terminate */ - if (to_terminate || list_empty(&tdc->pending_sg_req)) + if (to_terminate) return; + if (list_empty(&tdc->pending_sg_req)) { + pm_runtime_put(tdc->tdma->dev); + return; + } + tdc_start_head_req(tdc); } @@ -713,6 +718,7 @@ static void tegra_dma_issue_pending(struct dma_chan *dc) { struct tegra_dma_channel *tdc = to_tegra_dma_chan(dc); unsigned long flags; + int err; spin_lock_irqsave(&tdc->lock, flags); if (list_empty(&tdc->pending_sg_req)) { @@ -720,6 +726,12 @@ static void tegra_dma_issue_pending(struct dma_chan *dc) goto end; } if (!tdc->busy) { + err = pm_runtime_get_sync(tdc->tdma->dev); + if (err < 0) { + dev_err(tdc2dev(tdc), "Failed to enable DMA\n"); + goto end; + } + tdc_start_head_req(tdc); /* Continuous single mode: Configure next req */ @@ -778,6 +790,8 @@ static int tegra_dma_terminate_all(struct dma_chan *dc) } tegra_dma_resume(tdc); + pm_runtime_put(tdc->tdma->dev); + skip_dma_stop: tegra_dma_abort_all(tdc); @@ -1272,22 +1286,15 @@ tegra_dma_prep_dma_cyclic(struct dma_chan *dc, dma_addr_t buf_addr, static int tegra_dma_alloc_chan_resources(struct dma_chan *dc) { struct tegra_dma_channel *tdc = to_tegra_dma_chan(dc); - struct tegra_dma *tdma = tdc->tdma; - int ret; dma_cookie_init(&tdc->dma_chan); - ret = pm_runtime_get_sync(tdma->dev); - if (ret < 0) - return ret; - return 0; } static void tegra_dma_free_chan_resources(struct dma_chan *dc) { struct tegra_dma_channel *tdc = to_tegra_dma_chan(dc); - struct tegra_dma *tdma = tdc->tdma; struct tegra_dma_desc *dma_desc; struct tegra_dma_sg_req *sg_req; struct list_head dma_desc_list; @@ -1320,7 +1327,6 @@ static void tegra_dma_free_chan_resources(struct dma_chan *dc) list_del(&sg_req->node); kfree(sg_req); } - pm_runtime_put(tdma->dev); tdc->slave_id = TEGRA_APBDMA_SLAVE_ID_INVALID; } @@ -1420,6 +1426,11 @@ static int tegra_dma_probe(struct platform_device *pdev) spin_lock_init(&tdma->global_lock); + ret = clk_prepare(tdma->dma_clk); + if (ret) + return ret; + + pm_runtime_irq_safe(&pdev->dev); pm_runtime_enable(&pdev->dev); ret = pm_runtime_get_sync(&pdev->dev); @@ -1535,6 +1546,7 @@ static int tegra_dma_probe(struct platform_device *pdev) err_pm_disable: pm_runtime_disable(&pdev->dev); + clk_unprepare(tdma->dma_clk); return ret; } @@ -1545,6 +1557,7 @@ static int tegra_dma_remove(struct platform_device *pdev) dma_async_device_unregister(&tdma->dma_dev); pm_runtime_disable(&pdev->dev); + clk_unprepare(tdma->dma_clk); return 0; } @@ -1573,7 +1586,7 @@ static int tegra_dma_runtime_suspend(struct device *dev) TEGRA_APBDMA_CHAN_WCOUNT); } - clk_disable_unprepare(tdma->dma_clk); + clk_disable(tdma->dma_clk); return 0; } @@ -1584,7 +1597,7 @@ static int tegra_dma_runtime_resume(struct device *dev) unsigned int i; int ret; - ret = clk_prepare_enable(tdma->dma_clk); + ret = clk_enable(tdma->dma_clk); if (ret < 0) { dev_err(dev, "clk_enable failed: %d\n", ret); return ret;
It's a bit impractical to enable hardware's clock at the time of DMA channel's allocation because most of DMA client drivers allocate DMA channel at the time of the driver's probing, and thus, DMA clock is kept always-enabled in practice, defeating the whole purpose of runtime PM. Signed-off-by: Dmitry Osipenko <digetx@gmail.com> --- drivers/dma/tegra20-apb-dma.c | 35 ++++++++++++++++++++++++----------- 1 file changed, 24 insertions(+), 11 deletions(-)