Skip to contentSkip to conversation

ata: pata_pxa: use devm_platform_ioremap_resource

PATCH Applied Rosen Penev 3 participants 3 messages Jul 15 – Jul 18 libata

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ata: pata_pxa: use devm_platform_ioremap_resource

Replace the open-coded platform_get_resource() plus devm_ioremap() sequences for the CMD and CTL ports with devm_platform_ioremap_resource(), which fetches the resource, requests the region and maps it in one call. Switch the error checks to IS_ERR()/PTR_ERR() accordingly and drop the now-unused cmd_res and ctl_res variables.

The DMA port keeps using devm_ioremap() as its resource is IORESOURCE_DMA. The CMD (res 0) and CTL (res 1) MEM resources are disjoint in the ATA task-file layout, so the added region reservation in devm_ioremap_resource() does not introduce overlap failures.

Built for ARM (drivers/ata/pata_pxa.o) with LLVM=1.

Assisted-by opencode:hy3-free
---
 drivers/ata/pata_pxa.c | 30 ++++++------------------------
 1 file changed, 6 insertions(+), 24 deletions(-)
drivers/ata/pata_pxa.c +6 -24
index 9f63bdfb8576..c816997dc0a9 100644
@@ -161,8 +161,6 @@ static int pxa_ata_probe(struct platform_device *pdev)
 	struct ata_host *host;
 	struct ata_port *ap;
 	struct pata_pxa_data *data;
-	struct resource *cmd_res;
-	struct resource *ctl_res;
 	struct resource *dma_res;
 	struct pata_pxa_pdata *pdata = dev_get_platdata(&pdev->dev);
 	struct dma_slave_config	config;
@@ -181,20 +179,6 @@ static int pxa_ata_probe(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
-	/*
-	 * CMD port base address
-	 */
-	cmd_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (unlikely(cmd_res == NULL))
-		return -EINVAL;
-
-	/*
-	 * CTL port base address
-	 */
-	ctl_res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
-	if (unlikely(ctl_res == NULL))
-		return -EINVAL;
-
 	/*
 	 * DMA port base address
 	 */
@@ -221,14 +205,12 @@ static int pxa_ata_probe(struct platform_device *pdev)
 	ap->pio_mask	= ATA_PIO4;
 	ap->mwdma_mask	= ATA_MWDMA2;
 
-	ap->ioaddr.cmd_addr	= devm_ioremap(&pdev->dev, cmd_res->start,
-						resource_size(cmd_res));
-	if (!ap->ioaddr.cmd_addr)
-		return -ENOMEM;
-	ap->ioaddr.ctl_addr	= devm_ioremap(&pdev->dev, ctl_res->start,
-						resource_size(ctl_res));
-	if (!ap->ioaddr.ctl_addr)
-		return -ENOMEM;
+	ap->ioaddr.cmd_addr	= devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(ap->ioaddr.cmd_addr))
+		return PTR_ERR(ap->ioaddr.cmd_addr);
+	ap->ioaddr.ctl_addr	= devm_platform_ioremap_resource(pdev, 1);
+	if (IS_ERR(ap->ioaddr.ctl_addr))
+		return PTR_ERR(ap->ioaddr.ctl_addr);
 	ap->ioaddr.bmdma_addr	= devm_ioremap(&pdev->dev, dma_res->start,
 						resource_size(dma_res));
 	if (!ap->ioaddr.bmdma_addr)
Signature · 1 line
2.55.0
#1 lore raw
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Teardown order bug in pxa_ata_remove() causes Use-After-Free --

--- Patch [1]: [PATCH] ata: pata_pxa: use devm_platform_ioremap_resource --- commit 5dcbe84075bf85bf6a553efa199d7ee5d2ff52be
Author: Rosen Penev <[email protected]>

ata: pata_pxa: use devm_platform_ioremap_resource

This commit replaces the open-coded platform_get_resource() and devm_ioremap() sequences for the CMD and CTL ports with devm_platform_ioremap_resource(). It simplifies the probe function by consolidating resource fetching and mapping.

drivers/ata/pata_pxa.c

Quoted 4 lines
drivers/ata/pata_pxa.c +0 -0
index 9f63bdfb8576e..c816997dc0a99 100644

[ ... ]

[Severity: High]
This is a pre-existing issue, but does pxa_ata_remove() have a use-after-free if an interrupt occurs during module unload?

Looking at drivers/ata/pata_pxa.c:pxa_ata_remove():

static void pxa_ata_remove(struct platform_device *pdev)

{
	struct ata_host *host = platform_get_drvdata(pdev);
	struct pata_pxa_data *data = host->ports[0]->private_data;

	dma_release_channel(data->dma_chan);

	ata_host_detach(host);
}

The DMA channel is explicitly freed via dma_release_channel() before calling ata_host_detach().

Because the ATA host is still active and its interrupt handler was registered via devres (which only unregisters the IRQ after the remove callback completes), an active device or a stray interrupt could trigger ata_sff_interrupt().

This interrupt handler calls pxa_bmdma_status(), which dereferences the already freed data->dma_chan.

Should ata_host_detach() be called before releasing the DMA channel to prevent this race?

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
maintainer Applied
#2 lore raw
In reply to Rosen Penev
Quoted 11 lines · Rosen Penev

Replace the open-coded platform_get_resource() plus devm_ioremap() sequences for the CMD and CTL ports with devm_platform_ioremap_resource(), which fetches the resource, requests the region and maps it in one call. Switch the error checks to IS_ERR()/PTR_ERR() accordingly and drop the now-unused cmd_res and ctl_res variables.

The DMA port keeps using devm_ioremap() as its resource is IORESOURCE_DMA. The CMD (res 0) and CTL (res 1) MEM resources are disjoint in the ATA task-file layout, so the added region reservation in devm_ioremap_resource() does not introduce overlap failures.

Quoted 4 lines

Built for ARM (drivers/ata/pata_pxa.o) with LLVM=1.

Assisted-by opencode:hy3-free

Applied to for-7.3. Thanks!

Signature · 2 lines
Damien Le Moal
Western Digital Research