* [PATCHv2 0/2] dmaengine: fsl_raid: fix resource leaks and simplify drvdata handling @ 2026-07-21 19:54 Rosen Penev 2026-07-21 19:54 ` [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata Rosen Penev 2026-07-21 19:54 ` [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe Rosen Penev 0 siblings, 2 replies; 7+ messages in thread From: Rosen Penev @ 2026-07-21 19:54 UTC (permalink / raw) To: dmaengine; +Cc: Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list This series fixes resource leaks in the FSL RAIDEngine driver and simplifies drvdata handling: 1) Replace dev_get_drvdata/dev_set_drvdata with platform_get_drvdata/ platform_set_drvdata, since a platform_device is always used. This avoids taking a pointer to a platform device member. 2) Add missing free_irq() in error unwind of fsl_re_chan_probe() and in fsl_re_remove_chan(). Also add tasklet_kill() and of_platform_device_destroy() to properly release all resources. v2: split into two patches Rosen Penev (2): dmaengine: fsl_raid: pass platform_device to drvdata dmaengine: fsl_raid: free resources in probe drivers/dma/fsl_raid.c | 33 ++++++++++++++++++++------------- drivers/dma/fsl_raid.h | 1 + 2 files changed, 21 insertions(+), 13 deletions(-) -- 2.55.0 ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata 2026-07-21 19:54 [PATCHv2 0/2] dmaengine: fsl_raid: fix resource leaks and simplify drvdata handling Rosen Penev @ 2026-07-21 19:54 ` Rosen Penev 2026-07-21 21:38 ` Frank Li 2026-07-21 19:54 ` [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe Rosen Penev 1 sibling, 1 reply; 7+ messages in thread From: Rosen Penev @ 2026-07-21 19:54 UTC (permalink / raw) To: dmaengine; +Cc: Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list Simplifies code slightly and avoids having to take pointer of a platform device member. A platform device is always used. Signed-off-by: Rosen Penev <rosenp@gmail.com> --- drivers/dma/fsl_raid.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c index da5228e121d1..2d54879cd6e1 100644 --- a/drivers/dma/fsl_raid.c +++ b/drivers/dma/fsl_raid.c @@ -219,10 +219,11 @@ static void fsl_re_dequeue(struct tasklet_struct *t) /* Per Job Ring interrupt handler */ static irqreturn_t fsl_re_isr(int irq, void *data) { + struct platform_device *pdev = data; struct fsl_re_chan *re_chan; u32 irqstate, status; - re_chan = dev_get_drvdata((struct device *)data); + re_chan = platform_get_drvdata(pdev); irqstate = in_be32(&re_chan->jrregs->jr_interrupt_status); if (!irqstate) @@ -649,7 +650,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, struct platform_device *chan_ofdev; dev = &ofdev->dev; - re_priv = dev_get_drvdata(dev); + re_priv = platform_get_drvdata(ofdev); dma_dev = &re_priv->dma_dev; chan = devm_kzalloc(dev, sizeof(*chan), GFP_KERNEL); @@ -687,7 +688,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, chandev = &chan_ofdev->dev; tasklet_setup(&chan->irqtask, fsl_re_dequeue); - ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chandev); + ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); if (ret) { dev_err(dev, "Unable to register interrupt for JR %d\n", q); ret = -EINVAL; @@ -743,7 +744,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, out_be32(&chan->jrregs->jr_config_1, FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); - dev_set_drvdata(chandev, chan); + platform_set_drvdata(chan_ofdev, chan); /* Enable RE/CHAN */ out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); @@ -834,7 +835,7 @@ static int fsl_re_probe(struct platform_device *ofdev) return -ENOMEM; } - dev_set_drvdata(dev, re_priv); + platform_set_drvdata(ofdev, re_priv); /* Parse Device tree to find out the total number of JQs present */ for_each_compatible_node_scoped(np, NULL, "fsl,raideng-v1.0-job-queue") { @@ -903,11 +904,9 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) static void fsl_re_remove(struct platform_device *ofdev) { struct fsl_re_drv_private *re_priv; - struct device *dev; int i; - dev = &ofdev->dev; - re_priv = dev_get_drvdata(dev); + re_priv = platform_get_drvdata(ofdev); /* Unregister the DMA device first so no client can still hold a * channel and submit new work while the channel rings are torn -- 2.55.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata 2026-07-21 19:54 ` [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata Rosen Penev @ 2026-07-21 21:38 ` Frank Li 0 siblings, 0 replies; 7+ messages in thread From: Frank Li @ 2026-07-21 21:38 UTC (permalink / raw) To: Rosen Penev Cc: dmaengine, Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list On Tue, Jul 21, 2026 at 12:54:21PM -0700, Rosen Penev wrote: > Simplifies code slightly and avoids having to take pointer of a platform > device member. A platform device is always used. > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > --- Reviewed-by: Frank Li <Frank.Li@nxp.com> > drivers/dma/fsl_raid.c | 15 +++++++-------- > 1 file changed, 7 insertions(+), 8 deletions(-) > > diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c > index da5228e121d1..2d54879cd6e1 100644 > --- a/drivers/dma/fsl_raid.c > +++ b/drivers/dma/fsl_raid.c > @@ -219,10 +219,11 @@ static void fsl_re_dequeue(struct tasklet_struct *t) > /* Per Job Ring interrupt handler */ > static irqreturn_t fsl_re_isr(int irq, void *data) > { > + struct platform_device *pdev = data; > struct fsl_re_chan *re_chan; > u32 irqstate, status; > > - re_chan = dev_get_drvdata((struct device *)data); > + re_chan = platform_get_drvdata(pdev); > > irqstate = in_be32(&re_chan->jrregs->jr_interrupt_status); > if (!irqstate) > @@ -649,7 +650,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > struct platform_device *chan_ofdev; > > dev = &ofdev->dev; > - re_priv = dev_get_drvdata(dev); > + re_priv = platform_get_drvdata(ofdev); > dma_dev = &re_priv->dma_dev; > > chan = devm_kzalloc(dev, sizeof(*chan), GFP_KERNEL); > @@ -687,7 +688,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > chandev = &chan_ofdev->dev; > tasklet_setup(&chan->irqtask, fsl_re_dequeue); > > - ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chandev); > + ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); > if (ret) { > dev_err(dev, "Unable to register interrupt for JR %d\n", q); > ret = -EINVAL; > @@ -743,7 +744,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > out_be32(&chan->jrregs->jr_config_1, > FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); > > - dev_set_drvdata(chandev, chan); > + platform_set_drvdata(chan_ofdev, chan); > > /* Enable RE/CHAN */ > out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); > @@ -834,7 +835,7 @@ static int fsl_re_probe(struct platform_device *ofdev) > return -ENOMEM; > } > > - dev_set_drvdata(dev, re_priv); > + platform_set_drvdata(ofdev, re_priv); > > /* Parse Device tree to find out the total number of JQs present */ > for_each_compatible_node_scoped(np, NULL, "fsl,raideng-v1.0-job-queue") { > @@ -903,11 +904,9 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) > static void fsl_re_remove(struct platform_device *ofdev) > { > struct fsl_re_drv_private *re_priv; > - struct device *dev; > int i; > > - dev = &ofdev->dev; > - re_priv = dev_get_drvdata(dev); > + re_priv = platform_get_drvdata(ofdev); > > /* Unregister the DMA device first so no client can still hold a > * channel and submit new work while the channel rings are torn > -- > 2.55.0 > ^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe 2026-07-21 19:54 [PATCHv2 0/2] dmaengine: fsl_raid: fix resource leaks and simplify drvdata handling Rosen Penev 2026-07-21 19:54 ` [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata Rosen Penev @ 2026-07-21 19:54 ` Rosen Penev 2026-07-21 21:43 ` Frank Li 1 sibling, 1 reply; 7+ messages in thread From: Rosen Penev @ 2026-07-21 19:54 UTC (permalink / raw) To: dmaengine; +Cc: Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list Add free_irq() in the error unwind of fsl_re_chan_probe() and in fsl_re_remove_chan() so the interrupt is always released. Also add tasklet_kill(). Present in _remove but not _probe. Also add platform_device_put on failure as of_platform_device_create() increases the reference count and needs platform_device_put on failure. Requires placing the pointer in the struct for the _remove function. While at it, use platform_get/set_drvdata as there's no need for using a device pointer for that. Fixes: ad80da658bbc ("dmaengine: Driver support for FSL RaidEngine device.") Assisted-by: opencode:hy3-free Signed-off-by: Rosen Penev <rosenp@gmail.com> --- drivers/dma/fsl_raid.c | 20 ++++++++++++++------ drivers/dma/fsl_raid.h | 1 + 2 files changed, 15 insertions(+), 6 deletions(-) diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c index 2d54879cd6e1..cabc79e61e85 100644 --- a/drivers/dma/fsl_raid.c +++ b/drivers/dma/fsl_raid.c @@ -661,8 +661,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, chan_ofdev = of_platform_device_create(np, NULL, dev); if (!chan_ofdev) { dev_err(dev, "Not able to create ofdev for jr %d\n", q); - ret = -EINVAL; - goto err_free; + return -EINVAL; } /* read reg property from dts */ @@ -688,17 +687,19 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, chandev = &chan_ofdev->dev; tasklet_setup(&chan->irqtask, fsl_re_dequeue); + platform_set_drvdata(chan_ofdev, chan); ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); if (ret) { dev_err(dev, "Unable to register interrupt for JR %d\n", q); ret = -EINVAL; - goto err_free; + goto err_free_tasklet; } re_priv->re_jrs[q] = chan; chan->chan.device = dma_dev; chan->chan.private = chan; chan->dev = chandev; + chan->ofdev = chan_ofdev; chan->re_dev = re_priv; spin_lock_init(&chan->desc_lock); @@ -712,7 +713,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, if (!chan->inb_ring_virt_addr) { dev_err(dev, "No dma memory for inb_ring_virt_addr\n"); ret = -ENOMEM; - goto err_free; + goto err_free_irq; } chan->oub_ring_virt_addr = dma_pool_alloc(chan->re_dev->hw_desc_pool, @@ -744,8 +745,6 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, out_be32(&chan->jrregs->jr_config_1, FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); - platform_set_drvdata(chan_ofdev, chan); - /* Enable RE/CHAN */ out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); @@ -754,7 +753,12 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, err_free_1: dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, chan->inb_phys_addr); +err_free_irq: + free_irq(chan->irq, chan_ofdev); +err_free_tasklet: + tasklet_kill(&chan->irqtask); err_free: + of_platform_device_destroy(chan_ofdev); return ret; } @@ -892,6 +896,8 @@ static int fsl_re_probe(struct platform_device *ofdev) static void fsl_re_remove_chan(struct fsl_re_chan *chan) { + free_irq(chan->irq, chan->ofdev); + tasklet_kill(&chan->irqtask); dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, @@ -899,6 +905,8 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) dma_pool_free(chan->re_dev->hw_desc_pool, chan->oub_ring_virt_addr, chan->oub_phys_addr); + + of_platform_device_destroy(chan->ofdev); } static void fsl_re_remove(struct platform_device *ofdev) diff --git a/drivers/dma/fsl_raid.h b/drivers/dma/fsl_raid.h index adbfede330a7..34d362803610 100644 --- a/drivers/dma/fsl_raid.h +++ b/drivers/dma/fsl_raid.h @@ -270,6 +270,7 @@ struct fsl_re_chan { struct list_head active_q; /* already issued on hw, not completed */ struct list_head submit_q; struct list_head free_q; /* alloc available queue */ + struct platform_device *ofdev; struct device *dev; struct fsl_re_drv_private *re_dev; struct dma_chan chan; -- 2.55.0 ^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe 2026-07-21 19:54 ` [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe Rosen Penev @ 2026-07-21 21:43 ` Frank Li 2026-07-21 21:53 ` Rosen Penev 0 siblings, 1 reply; 7+ messages in thread From: Frank Li @ 2026-07-21 21:43 UTC (permalink / raw) To: Rosen Penev Cc: dmaengine, Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list On Tue, Jul 21, 2026 at 12:54:22PM -0700, Rosen Penev wrote: > Add free_irq() in the error unwind of fsl_re_chan_probe() and in > fsl_re_remove_chan() so the interrupt is always released. > > Also add tasklet_kill(). Present in _remove but not _probe. > > Also add platform_device_put on failure as of_platform_device_create() > increases the reference count and needs platform_device_put on failure. > Requires placing the pointer in the struct for the _remove function. > While at it, use platform_get/set_drvdata as there's no need for using a > device pointer for that. > > Fixes: ad80da658bbc ("dmaengine: Driver support for FSL RaidEngine device.") > Assisted-by: opencode:hy3-free > Signed-off-by: Rosen Penev <rosenp@gmail.com> > --- > drivers/dma/fsl_raid.c | 20 ++++++++++++++------ > drivers/dma/fsl_raid.h | 1 + > 2 files changed, 15 insertions(+), 6 deletions(-) > > diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c > index 2d54879cd6e1..cabc79e61e85 100644 > --- a/drivers/dma/fsl_raid.c > +++ b/drivers/dma/fsl_raid.c > @@ -661,8 +661,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > chan_ofdev = of_platform_device_create(np, NULL, dev); > if (!chan_ofdev) { > dev_err(dev, "Not able to create ofdev for jr %d\n", q); > - ret = -EINVAL; > - goto err_free; > + return -EINVAL; > } > > /* read reg property from dts */ > @@ -688,17 +687,19 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > chandev = &chan_ofdev->dev; > tasklet_setup(&chan->irqtask, fsl_re_dequeue); > > + platform_set_drvdata(chan_ofdev, chan); Previous patch already set, why need set again? Frank > ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); > if (ret) { > dev_err(dev, "Unable to register interrupt for JR %d\n", q); > ret = -EINVAL; > - goto err_free; > + goto err_free_tasklet; > } > > re_priv->re_jrs[q] = chan; > chan->chan.device = dma_dev; > chan->chan.private = chan; > chan->dev = chandev; > + chan->ofdev = chan_ofdev; > chan->re_dev = re_priv; > > spin_lock_init(&chan->desc_lock); > @@ -712,7 +713,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > if (!chan->inb_ring_virt_addr) { > dev_err(dev, "No dma memory for inb_ring_virt_addr\n"); > ret = -ENOMEM; > - goto err_free; > + goto err_free_irq; > } > > chan->oub_ring_virt_addr = dma_pool_alloc(chan->re_dev->hw_desc_pool, > @@ -744,8 +745,6 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > out_be32(&chan->jrregs->jr_config_1, > FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); > > - platform_set_drvdata(chan_ofdev, chan); > - > /* Enable RE/CHAN */ > out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); > > @@ -754,7 +753,12 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > err_free_1: > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > chan->inb_phys_addr); > +err_free_irq: > + free_irq(chan->irq, chan_ofdev); > +err_free_tasklet: > + tasklet_kill(&chan->irqtask); > err_free: > + of_platform_device_destroy(chan_ofdev); > return ret; > } > > @@ -892,6 +896,8 @@ static int fsl_re_probe(struct platform_device *ofdev) > > static void fsl_re_remove_chan(struct fsl_re_chan *chan) > { > + free_irq(chan->irq, chan->ofdev); > + > tasklet_kill(&chan->irqtask); > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > @@ -899,6 +905,8 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->oub_ring_virt_addr, > chan->oub_phys_addr); > + > + of_platform_device_destroy(chan->ofdev); > } > > static void fsl_re_remove(struct platform_device *ofdev) > diff --git a/drivers/dma/fsl_raid.h b/drivers/dma/fsl_raid.h > index adbfede330a7..34d362803610 100644 > --- a/drivers/dma/fsl_raid.h > +++ b/drivers/dma/fsl_raid.h > @@ -270,6 +270,7 @@ struct fsl_re_chan { > struct list_head active_q; /* already issued on hw, not completed */ > struct list_head submit_q; > struct list_head free_q; /* alloc available queue */ > + struct platform_device *ofdev; > struct device *dev; > struct fsl_re_drv_private *re_dev; > struct dma_chan chan; > -- > 2.55.0 > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe 2026-07-21 21:43 ` Frank Li @ 2026-07-21 21:53 ` Rosen Penev 2026-07-21 23:56 ` Frank Li 0 siblings, 1 reply; 7+ messages in thread From: Rosen Penev @ 2026-07-21 21:53 UTC (permalink / raw) To: Frank Li Cc: dmaengine, Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list On Tue, Jul 21, 2026 at 2:43 PM Frank Li <Frank.li@oss.nxp.com> wrote: > > On Tue, Jul 21, 2026 at 12:54:22PM -0700, Rosen Penev wrote: > > Add free_irq() in the error unwind of fsl_re_chan_probe() and in > > fsl_re_remove_chan() so the interrupt is always released. > > > > Also add tasklet_kill(). Present in _remove but not _probe. > > > > Also add platform_device_put on failure as of_platform_device_create() > > increases the reference count and needs platform_device_put on failure. > > Requires placing the pointer in the struct for the _remove function. > > While at it, use platform_get/set_drvdata as there's no need for using a > > device pointer for that. > > > > Fixes: ad80da658bbc ("dmaengine: Driver support for FSL RaidEngine device.") > > Assisted-by: opencode:hy3-free > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > --- > > drivers/dma/fsl_raid.c | 20 ++++++++++++++------ > > drivers/dma/fsl_raid.h | 1 + > > 2 files changed, 15 insertions(+), 6 deletions(-) > > > > diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c > > index 2d54879cd6e1..cabc79e61e85 100644 > > --- a/drivers/dma/fsl_raid.c > > +++ b/drivers/dma/fsl_raid.c > > @@ -661,8 +661,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > chan_ofdev = of_platform_device_create(np, NULL, dev); > > if (!chan_ofdev) { > > dev_err(dev, "Not able to create ofdev for jr %d\n", q); > > - ret = -EINVAL; > > - goto err_free; > > + return -EINVAL; > > } > > > > /* read reg property from dts */ > > @@ -688,17 +687,19 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > chandev = &chan_ofdev->dev; > > tasklet_setup(&chan->irqtask, fsl_re_dequeue); > > > > + platform_set_drvdata(chan_ofdev, chan); > > Previous patch already set, why need set again? This is because request_irq uses platform_get_drvdata. On further review, I think chan should be passed directly. > > Frank > > ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); > > if (ret) { > > dev_err(dev, "Unable to register interrupt for JR %d\n", q); > > ret = -EINVAL; > > - goto err_free; > > + goto err_free_tasklet; > > } > > > > re_priv->re_jrs[q] = chan; > > chan->chan.device = dma_dev; > > chan->chan.private = chan; > > chan->dev = chandev; > > + chan->ofdev = chan_ofdev; > > chan->re_dev = re_priv; > > > > spin_lock_init(&chan->desc_lock); > > @@ -712,7 +713,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > if (!chan->inb_ring_virt_addr) { > > dev_err(dev, "No dma memory for inb_ring_virt_addr\n"); > > ret = -ENOMEM; > > - goto err_free; > > + goto err_free_irq; > > } > > > > chan->oub_ring_virt_addr = dma_pool_alloc(chan->re_dev->hw_desc_pool, > > @@ -744,8 +745,6 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > out_be32(&chan->jrregs->jr_config_1, > > FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); > > > > - platform_set_drvdata(chan_ofdev, chan); > > - > > /* Enable RE/CHAN */ > > out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); > > > > @@ -754,7 +753,12 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > err_free_1: > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > > chan->inb_phys_addr); > > +err_free_irq: > > + free_irq(chan->irq, chan_ofdev); > > +err_free_tasklet: > > + tasklet_kill(&chan->irqtask); > > err_free: > > + of_platform_device_destroy(chan_ofdev); > > return ret; > > } > > > > @@ -892,6 +896,8 @@ static int fsl_re_probe(struct platform_device *ofdev) > > > > static void fsl_re_remove_chan(struct fsl_re_chan *chan) > > { > > + free_irq(chan->irq, chan->ofdev); > > + > > tasklet_kill(&chan->irqtask); > > > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > > @@ -899,6 +905,8 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) > > > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->oub_ring_virt_addr, > > chan->oub_phys_addr); > > + > > + of_platform_device_destroy(chan->ofdev); > > } > > > > static void fsl_re_remove(struct platform_device *ofdev) > > diff --git a/drivers/dma/fsl_raid.h b/drivers/dma/fsl_raid.h > > index adbfede330a7..34d362803610 100644 > > --- a/drivers/dma/fsl_raid.h > > +++ b/drivers/dma/fsl_raid.h > > @@ -270,6 +270,7 @@ struct fsl_re_chan { > > struct list_head active_q; /* already issued on hw, not completed */ > > struct list_head submit_q; > > struct list_head free_q; /* alloc available queue */ > > + struct platform_device *ofdev; > > struct device *dev; > > struct fsl_re_drv_private *re_dev; > > struct dma_chan chan; > > -- > > 2.55.0 > > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe 2026-07-21 21:53 ` Rosen Penev @ 2026-07-21 23:56 ` Frank Li 0 siblings, 0 replies; 7+ messages in thread From: Frank Li @ 2026-07-21 23:56 UTC (permalink / raw) To: Rosen Penev Cc: dmaengine, Vinod Koul, Frank Li, Xuelin Shi, Harninder Rai, open list On Tue, Jul 21, 2026 at 02:53:17PM -0700, Rosen Penev wrote: > On Tue, Jul 21, 2026 at 2:43 PM Frank Li <Frank.li@oss.nxp.com> wrote: > > > > On Tue, Jul 21, 2026 at 12:54:22PM -0700, Rosen Penev wrote: > > > Add free_irq() in the error unwind of fsl_re_chan_probe() and in > > > fsl_re_remove_chan() so the interrupt is always released. > > > > > > Also add tasklet_kill(). Present in _remove but not _probe. > > > > > > Also add platform_device_put on failure as of_platform_device_create() platform_device_put() > > > increases the reference count and needs platform_device_put on failure. > > > Requires placing the pointer in the struct for the _remove function. > > > While at it, use platform_get/set_drvdata as there's no need for using a > > > device pointer for that. Needn't "While at ..." > > > > > > Fixes: ad80da658bbc ("dmaengine: Driver support for FSL RaidEngine device.") > > > Assisted-by: opencode:hy3-free > > > Signed-off-by: Rosen Penev <rosenp@gmail.com> > > > --- > > > drivers/dma/fsl_raid.c | 20 ++++++++++++++------ > > > drivers/dma/fsl_raid.h | 1 + > > > 2 files changed, 15 insertions(+), 6 deletions(-) > > > > > > diff --git a/drivers/dma/fsl_raid.c b/drivers/dma/fsl_raid.c > > > index 2d54879cd6e1..cabc79e61e85 100644 > > > --- a/drivers/dma/fsl_raid.c > > > +++ b/drivers/dma/fsl_raid.c > > > @@ -661,8 +661,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > > chan_ofdev = of_platform_device_create(np, NULL, dev); > > > if (!chan_ofdev) { > > > dev_err(dev, "Not able to create ofdev for jr %d\n", q); > > > - ret = -EINVAL; > > > - goto err_free; > > > + return -EINVAL; > > > } > > > > > > /* read reg property from dts */ > > > @@ -688,17 +687,19 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > > chandev = &chan_ofdev->dev; > > > tasklet_setup(&chan->irqtask, fsl_re_dequeue); > > > > > > + platform_set_drvdata(chan_ofdev, chan); > > > > Previous patch already set, why need set again? > This is because request_irq uses platform_get_drvdata. > > On further review, I think chan should be passed directly. Yes Frank > > > > Frank > > > ret = request_irq(chan->irq, fsl_re_isr, 0, chan->name, chan_ofdev); > > > if (ret) { > > > dev_err(dev, "Unable to register interrupt for JR %d\n", q); > > > ret = -EINVAL; > > > - goto err_free; > > > + goto err_free_tasklet; > > > } > > > > > > re_priv->re_jrs[q] = chan; > > > chan->chan.device = dma_dev; > > > chan->chan.private = chan; > > > chan->dev = chandev; > > > + chan->ofdev = chan_ofdev; > > > chan->re_dev = re_priv; > > > > > > spin_lock_init(&chan->desc_lock); > > > @@ -712,7 +713,7 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > > if (!chan->inb_ring_virt_addr) { > > > dev_err(dev, "No dma memory for inb_ring_virt_addr\n"); > > > ret = -ENOMEM; > > > - goto err_free; > > > + goto err_free_irq; > > > } > > > > > > chan->oub_ring_virt_addr = dma_pool_alloc(chan->re_dev->hw_desc_pool, > > > @@ -744,8 +745,6 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > > out_be32(&chan->jrregs->jr_config_1, > > > FSL_RE_CFG1_CBSI | FSL_RE_CFG1_CBS0 | status); > > > > > > - platform_set_drvdata(chan_ofdev, chan); > > > - > > > /* Enable RE/CHAN */ > > > out_be32(&chan->jrregs->jr_command, FSL_RE_ENABLE); > > > > > > @@ -754,7 +753,12 @@ static int fsl_re_chan_probe(struct platform_device *ofdev, > > > err_free_1: > > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > > > chan->inb_phys_addr); > > > +err_free_irq: > > > + free_irq(chan->irq, chan_ofdev); > > > +err_free_tasklet: > > > + tasklet_kill(&chan->irqtask); > > > err_free: > > > + of_platform_device_destroy(chan_ofdev); > > > return ret; > > > } > > > > > > @@ -892,6 +896,8 @@ static int fsl_re_probe(struct platform_device *ofdev) > > > > > > static void fsl_re_remove_chan(struct fsl_re_chan *chan) > > > { > > > + free_irq(chan->irq, chan->ofdev); > > > + > > > tasklet_kill(&chan->irqtask); > > > > > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->inb_ring_virt_addr, > > > @@ -899,6 +905,8 @@ static void fsl_re_remove_chan(struct fsl_re_chan *chan) > > > > > > dma_pool_free(chan->re_dev->hw_desc_pool, chan->oub_ring_virt_addr, > > > chan->oub_phys_addr); > > > + > > > + of_platform_device_destroy(chan->ofdev); > > > } > > > > > > static void fsl_re_remove(struct platform_device *ofdev) > > > diff --git a/drivers/dma/fsl_raid.h b/drivers/dma/fsl_raid.h > > > index adbfede330a7..34d362803610 100644 > > > --- a/drivers/dma/fsl_raid.h > > > +++ b/drivers/dma/fsl_raid.h > > > @@ -270,6 +270,7 @@ struct fsl_re_chan { > > > struct list_head active_q; /* already issued on hw, not completed */ > > > struct list_head submit_q; > > > struct list_head free_q; /* alloc available queue */ > > > + struct platform_device *ofdev; > > > struct device *dev; > > > struct fsl_re_drv_private *re_dev; > > > struct dma_chan chan; > > > -- > > > 2.55.0 > > > ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2026-07-21 23:56 UTC | newest] Thread overview: 7+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-07-21 19:54 [PATCHv2 0/2] dmaengine: fsl_raid: fix resource leaks and simplify drvdata handling Rosen Penev 2026-07-21 19:54 ` [PATCHv2 1/2] dmaengine: fsl_raid: pass platform_device to drvdata Rosen Penev 2026-07-21 21:38 ` Frank Li 2026-07-21 19:54 ` [PATCHv2 2/2] dmaengine: fsl_raid: free resources in probe Rosen Penev 2026-07-21 21:43 ` Frank Li 2026-07-21 21:53 ` Rosen Penev 2026-07-21 23:56 ` Frank Li
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox