Skip to contentSkip to conversation

staging: Android: vsoc: Create wc kernel mapping for region shm.

PATCH In review Alistair Strachan 2 participants 2 messages May 3, 2018 Staging

lore
Share & tools
lore
#0 lore raw

PATCH 1/3 staging: Android: vsoc: Create wc kernel mapping for region shm.

Map the region shm as write-combining instead of uncachable.

Cc Greg Kroah-Hartman <[email protected]>
Cc Arve Hjønnevåg <[email protected]>
Cc Todd Kjos <[email protected]>
Cc Martijn Coenen <[email protected]>
Cc Greg Hartman <[email protected]>
---
 drivers/staging/android/TODO   | 1 -
 drivers/staging/android/vsoc.c | 4 +---
 2 files changed, 1 insertion(+), 4 deletions(-)
drivers/staging/android/TODO +0 -1
index 2ea6f97b8f0f..ebd6ba3ae02e 100644
@@ -18,7 +18,6 @@ vsoc.c, uapi/vsoc_shm.h
    waiting threads. We should eventually use multiple queues and select the
    queue based on the region.
  - Add debugfs support for examining the permissions of regions.
- - Use ioremap_wc instead of ioremap_nocache.
  - Remove VSOC_WAIT_FOR_INCOMING_INTERRUPT ioctl. This functionality has been
    superseded by the futex and is there for legacy reasons.
 
drivers/staging/android/vsoc.c +1 -3
index 587c66d709b9..794137b7751f 100644
@@ -802,9 +802,7 @@ static int vsoc_probe_device(struct pci_dev *pdev,
 
 	dev_info(&pdev->dev, "shared memory @ DMA %p size=0x%zx\n",
 		 (void *)vsoc_dev.shm_phys_start, vsoc_dev.shm_size);
-	/* TODO(ghartman): ioremap_wc should work here */
-	vsoc_dev.kernel_mapped_shm = ioremap_nocache(
-			vsoc_dev.shm_phys_start, vsoc_dev.shm_size);
+	vsoc_dev.kernel_mapped_shm = pci_iomap_wc(pdev, SHARED_MEMORY_BAR, 0);
 	if (!vsoc_dev.kernel_mapped_shm) {
 		dev_err(&vsoc_dev.dev->dev, "cannot iomap region\n");
 		vsoc_remove_device(pdev);
#1 lore raw
Quoted 1 line · Alistair Strachan' via kernel-team

Map the region shm as write-combining instead of uncachable.

I think more commit message is needed. Why is this done, what does it fix, etc. Its hard to know what the improvement is without a cover-letter either.

I am assuming the improvement is in performance by using the write-combining mode.

thanks,

- Joel
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel