See below. Also see comment re: the patch description.
-Ewan
On Wed, 2019-02-13 at 12:42 +0100, Hannes Reinecke wrote:
The change to use dma_set_mask_and_coherent() incorrectly made a second call with the 32 bit DMA mask value when the call with the 64 bit DMA mask value succeeded. This resulted in FC connections failing due
-------------------------------------------
to corrupted data buffers, and various other SCSI/FCP I/O errors.
-----------------------------------------------------------------
The last sentence should be removed from the patch description.
Fixes: a69b080025ea ("scsi: bfa: use dma_set_mask_and_coherent") Cc: stable@vger.kernel.org Suggested-by: Ewan D. Milne emilne@redhat.com Signed-off-by: Hannes Reinecke hare@suse.com
drivers/scsi/bfa/bfad.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/scsi/bfa/bfad.c b/drivers/scsi/bfa/bfad.c index 42a0caf6740d..2ffbe36f5860 100644 --- a/drivers/scsi/bfa/bfad.c +++ b/drivers/scsi/bfa/bfad.c @@ -727,7 +727,7 @@ bfad_init_timer(struct bfad_s *bfad) int bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad) {
- int rc = -ENODEV;
- int rc;
There are error paths from the calls to pci_enable_device() and pci_request_regions() that will return an undefined value if this initializer is removed. Leave it in place?
if (pci_enable_device(pdev)) { printk(KERN_ERR "pci_enable_device fail %p\n", pdev); @@ -739,11 +739,15 @@ bfad_pci_init(struct pci_dev *pdev, struct bfad_s *bfad) pci_set_master(pdev);
- if (dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64)) ||
dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32))) {
- rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
- if (rc)
rc = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32));- if (rc) { printk(KERN_ERR "dma_set_mask_and_coherent fail %p\n", pdev); goto out_release_region; }
- rc = -ENODEV;
/* Enable PCIE Advanced Error Recovery (AER) if kernel supports */ pci_enable_pcie_error_reporting(pdev);