Hello all,
This patch adds the check for 32-bit DMA capability for the tlan driver.
This is the first step for Documentation/DMA-mapping.txt compliance. As
for the preferred action if the driver fails on pci_set_dma_mask(), I plan
to add that in a future patch. Please review.
Regards,
Frank
--- drivers/net/tlan.c.old Sun May 26 14:37:35 2002
+++ drivers/net/tlan.c Sun Jun 23 22:15:30 2002
@@ -516,6 +516,10 @@
if (pdev && pci_enable_device(pdev))
return -EIO;
+ if(pci_set_dma_mask(pdev, 0xffffffff))
+ {
+ printk(KERN_WARNING "tlan : No suitable DMA available.\n");
+ }
dev = init_etherdev(NULL, sizeof(TLanPrivateInfo));
if (dev == NULL) {
Re,
Frank Davis <[email protected]> :
> Hello all,
> This patch adds the check for 32-bit DMA capability for the tlan driver.
> This is the first step for Documentation/DMA-mapping.txt compliance. As
> for the preferred action if the driver fails on pci_set_dma_mask(), I plan
> to add that in a future patch. Please review.
See comment regarding scsi drivers.
I've done some work for this one and rrunner around 2.4.7. Give me 24h to
dig the mail archive and see what the missing parts/problems were, ok ?
Btw, you may Cc:[email protected] as well as the adequate maintainer
if there's one.
--
Ueimor
Greetings,
[remaining DMA-mapping conversion in drivers/net]
- tlan plays isa/pci and does as well some interesting things with copy
buffers. The patch is attached but I'll feel safer if somebody with
the adequate hardware at hand appears.
- Imho rrunner wouldn't suffer from compat code removal + pci driver init
style + DMA mapping conversion. I'd like to know what the maintainer
thinks before resurrecting the patch. Jes ?
--
Ueimor