On 12/9/21 23:59, Andy Shevchenko wrote: > If 64-bit mask attempt fails, the 32-bit will fail by the very same reason. > Don't even try the latter. It's a continuation of the changes that contains, > e.g. dcc02c19cc06 ("sata_sil24: use dma_set_mask_and_coherent"). > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> > --- > drivers/ata/libahci_platform.c | 9 ++------- > 1 file changed, 2 insertions(+), 7 deletions(-) > > diff --git a/drivers/ata/libahci_platform.c b/drivers/ata/libahci_platform.c > index 1af642c84e7b..972f5ec86a27 100644 > --- a/drivers/ata/libahci_platform.c > +++ b/drivers/ata/libahci_platform.c > @@ -637,13 +637,8 @@ int ahci_platform_init_host(struct platform_device *pdev, > if (hpriv->cap & HOST_CAP_64) { > rc = dma_coerce_mask_and_coherent(dev, DMA_BIT_MASK(64)); > if (rc) { > - rc = dma_coerce_mask_and_coherent(dev, > - DMA_BIT_MASK(32)); > - if (rc) { > - dev_err(dev, "Failed to enable 64-bit DMA.\n"); > - return rc; > - } > - dev_warn(dev, "Enable 32-bit DMA instead of 64-bit.\n"); > + dev_err(dev, "Failed to enable 64-bit DMA.\n"); > + return rc; > } > } > > Applied to for-5.17. For patch 1/2, waiting for you v2 restoring the irq == 0 check. -- Damien Le Moal Western Digital Research