Skip to contentSkip to conversation

memremap: Check for size parameter

PATCH In review Andy Shevchenko 3 participants 9 messages Nov 15, 2019 – Apr 15, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/5 memremap: Check for size parameter

There is no use of memremap() to be called with size = 0. Simple return NULL pointer and allow callers to drop this check.

Fixes 92281dee825f ("arch: introduce memremap()")
Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Marek Szyprowski <[email protected]>
Cc Robin Murphy <[email protected]>
Cc Jaroslav Kysela <[email protected]>
Cc Takashi Iwai <[email protected]>
Cc Pierre-Louis Bossart <[email protected]>
Cc Ard Biesheuvel <[email protected]>
Cc Alex Williamson <[email protected]>
Cc Cornelia Huck <[email protected]>
---
 kernel/iomem.c | 3 +++
 1 file changed, 3 insertions(+)
kernel/iomem.c +3 -0
index 62c92e43aa0d..bf7a5fc32760 100644
@@ -74,6 +74,9 @@ void *memremap(resource_size_t offset, size_t size, unsigned long flags)
 				       IORESOURCE_SYSTEM_RAM, IORES_DESC_NONE);
 	void *addr = NULL;
 
+	if (!size)
+		return NULL;
+
 	if (!flags)
 		return NULL;
 
Signature · 1 line
2.24.0
#1 lore raw

PATCH 2/5 dma-mapping: Drop duplicate check for size parameter of memremap()

Since memremap() returns NULL pointer for size = 0, there is no need to duplicate this check in the callers.

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Marek Szyprowski <[email protected]>
Cc Robin Murphy <[email protected]>
---
 kernel/dma/coherent.c | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)
kernel/dma/coherent.c +1 -6
index 551b0eb7028a..7e669c083324 100644
@@ -42,16 +42,11 @@ static int dma_init_coherent_memory(phys_addr_t phys_addr,
 		struct dma_coherent_mem **mem)
 {
 	struct dma_coherent_mem *dma_mem = NULL;
-	void *mem_base = NULL;
+	void *mem_base;
 	int pages = size >> PAGE_SHIFT;
 	int bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
 	int ret;
 
-	if (!size) {
-		ret = -EINVAL;
-		goto out;
-	}
-
 	mem_base = memremap(phys_addr, size, MEMREMAP_WC);
 	if (!mem_base) {
 		ret = -EINVAL;
Signature · 1 line
2.24.0
#3 lore raw

PATCH 3/5 ALSA: hda: Drop duplicate check for size parameter of memremap()

In reply to Andy Shevchenko

Since memremap() returns NULL pointer for size = 0, there is no need to duplicate this check in the callers.

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Jaroslav Kysela <[email protected]>
Cc Takashi Iwai <[email protected]>
Cc Pierre-Louis Bossart <[email protected]>
---
 sound/hda/intel-nhlt.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
sound/hda/intel-nhlt.c +2 -3
index 097ff6c10099..5fc10f1b244c 100644
@@ -16,7 +16,7 @@ struct nhlt_acpi_table *intel_nhlt_init(struct device *dev)
 	acpi_handle handle;
 	union acpi_object *obj;
 	struct nhlt_resource_desc *nhlt_ptr;
-	struct nhlt_acpi_table *nhlt_table = NULL;
+	struct nhlt_acpi_table *nhlt_table;
 
 	handle = ACPI_HANDLE(dev);
 	if (!handle) {
@@ -36,8 +36,7 @@ struct nhlt_acpi_table *intel_nhlt_init(struct device *dev)
 	}
 
 	nhlt_ptr = (struct nhlt_resource_desc  *)obj->buffer.pointer;
-	if (nhlt_ptr->length)
-		nhlt_table = (struct nhlt_acpi_table *)
+	nhlt_table = (struct nhlt_acpi_table *)
 			memremap(nhlt_ptr->min_addr, nhlt_ptr->length,
 				 MEMREMAP_WB);
 	ACPI_FREE(obj);
Signature · 1 line
2.24.0
#4 lore raw

PATCH 4/5 efi/esrt: Drop duplicate check for size parameter of memremap()

In reply to Andy Shevchenko

Since memremap() returns NULL pointer for size = 0, there is no need to duplicate this check in the callers.

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Ard Biesheuvel <[email protected]>
---
 drivers/firmware/efi/esrt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/firmware/efi/esrt.c +1 -1
index 2762e0662bf4..dce56a4c4011 100644
@@ -381,7 +381,7 @@ static int __init esrt_sysfs_init(void)
 	int error;
 
 	pr_debug("esrt-sysfs: loading.\n");
-	if (!esrt_data || !esrt_data_size)
+	if (!esrt_data)
 		return -ENOSYS;
 
 	esrt = memremap(esrt_data, esrt_data_size, MEMREMAP_WB);
Signature · 1 line
2.24.0
#5 lore raw

PATCH 5/5 vfio/pci: Drop duplicate check for size parameter of memremap()

In reply to Andy Shevchenko

Since memremap() returns NULL pointer for size = 0, there is no need to duplicate this check in the callers.

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Alex Williamson <[email protected]>
Cc Cornelia Huck <[email protected]>
---
 drivers/vfio/pci/vfio_pci_igd.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/vfio/pci/vfio_pci_igd.c +1 -7
index 53d97f459252..3088a33af271 100644
@@ -75,13 +75,7 @@ static int vfio_pci_igd_opregion_init(struct vfio_pci_device *vdev)
 		return -EINVAL;
 	}
 
-	size = le32_to_cpu(*(__le32 *)(base + 16));
-	if (!size) {
-		memunmap(base);
-		return -EINVAL;
-	}
-
-	size *= 1024; /* In KB */
+	size = le32_to_cpu(*(__le32 *)(base + 16)) * 1024; /* In KB */
 
 	if (size != OPREGION_SIZE) {
 		memunmap(base);
Signature · 1 line
2.24.0
#6 lore raw

Re: [PATCH v1 5/5] vfio/pci: Drop duplicate check for size parameter of memremap()

Quoted 21 lines · On Fri, 15 Nov 2019 20:00:44 +0200 Andy Shevchenko…

Since memremap() returns NULL pointer for size = 0, there is no need to duplicate this check in the callers.

Cc Christoph Hellwig <[email protected]>
Cc Dan Williams <[email protected]>
Cc Alex Williamson <[email protected]>
Cc Cornelia Huck <[email protected]>
---
 drivers/vfio/pci/vfio_pci_igd.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/vfio/pci/vfio_pci_igd.c +0 -1
index 53d97f459252..3088a33af271 100644
@@ -75,13 +75,7 @@ static int vfio_pci_igd_opregion_init(struct vfio_pci_device *vdev)
 		return -EINVAL;
 	}
 
-	size = le32_to_cpu(*(__le32 *)(base + 16));

drivers/vfio/pci/vfio_pci_igd.c static int vfio_pci_igd_opregion_init(struct vfio_pci_device *vdev) · line 81

Quoted 10 lines
drivers/vfio/pci/vfio_pci_igd.c +1 -6
-	if (!size) {
-		memunmap(base);
-		return -EINVAL;
-	}
-
-	size *= 1024; /* In KB */
+	size = le32_to_cpu(*(__le32 *)(base + 16)) * 1024; /* In KB */
 
 	if (size != OPREGION_SIZE) {
 		memunmap(base);

This seems convoluted, patch 1/5 states "[t]here is no use of memremap() to be called with size = 0", which we weren't doing thanks to the check removed above. So now we are potentially calling it with zero, apparently only to take advantage of this new behavior, and we lose the error granularity that previously such a condition failed with an -EINVAL and now we fail with an -ENONMEM and cannot distinguish whether the OpRegion table size was empty or we just weren't able to memremap() it. I don't see how this is an improvement. Thanks,

Alex

#7 lore raw
In reply to Andy Shevchenko
Quoted 2 lines · Andy Shevchenko

There is no use of memremap() to be called with size = 0. Simple return NULL pointer and allow callers to drop this check.

Given that this really is an error condition, maybe a WARN_ON_ONCE would fit here?

#8 lore raw
Quoted 6 lines · Christoph Hellwig
Quoted 2 lines · Andy Shevchenko

There is no use of memremap() to be called with size = 0. Simple return NULL pointer and allow callers to drop this check.

Given that this really is an error condition, maybe a WARN_ON_ONCE would fit here?

It appears some users are using defensive programming and rely simple on error code. I dunno if they are really expect to have size == 0 in some (non-fatal) cases.

Signature · 2 lines
With Best Regards,
Andy Shevchenko