| Message ID | 20100121114420.GA11892@kryten |
|---|---|
| State | Awaiting Upstream, archived |
| Delegated to: | David Miller |
| Headers | show |
On Thu, Jan 21, 2010 at 03:44, Anton Blanchard <anton@samba.org> wrote: > > Check for error return from pci_map_single/pci_map_page and clean up. > > With this and the previous patch the driver was able to handle a significant > percentage of errors (I set the fault injection rate to 10% and could still > download large files at a reasonable speed). > > Signed-off-by: Anton Blanchard <anton@samba.org> > --- > > I wasn't able to stress the jumbo frame path, so that part could do with some > more eyes. > > Index: linux.trees.git/drivers/net/e1000/e1000_main.c > =================================================================== > --- linux.trees.git.orig/drivers/net/e1000/e1000_main.c 2010-01-21 11:13:10.000000000 +1100 > +++ linux.trees.git/drivers/net/e1000/e1000_main.c 2010-01-21 11:15:54.000000000 +1100 > @@ -4014,11 +4014,21 @@ check_page: > } > } > > - if (!buffer_info->dma) > + if (!buffer_info->dma) { > buffer_info->dma = pci_map_page(pdev, > buffer_info->page, 0, > buffer_info->length, > PCI_DMA_FROMDEVICE); > + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { > + put_page(buffer_info->page); > + dev_kfree_skb(skb); > + buffer_info->page = NULL; > + buffer_info->skb = NULL; > + buffer_info->dma = 0; > + adapter->alloc_rx_buff_failed++; > + break; /* while !buffer_info->skb */ > + } > + } > > rx_desc = E1000_RX_DESC(*rx_ring, i); > rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); > @@ -4109,6 +4119,13 @@ map_skb: > skb->data, > buffer_info->length, > PCI_DMA_FROMDEVICE); > + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { > + dev_kfree_skb(skb); > + buffer_info->skb = NULL; > + buffer_info->dma = 0; > + adapter->alloc_rx_buff_failed++; > + break; /* while !buffer_info->skb */ > + } > > /* > * XXX if it was allocated cleanly it will never map to a I have added this patch to my queue of patches for review and testing. Thanks.
Index: linux.trees.git/drivers/net/e1000/e1000_main.c =================================================================== --- linux.trees.git.orig/drivers/net/e1000/e1000_main.c 2010-01-21 11:13:10.000000000 +1100 +++ linux.trees.git/drivers/net/e1000/e1000_main.c 2010-01-21 11:15:54.000000000 +1100 @@ -4014,11 +4014,21 @@ check_page: } } - if (!buffer_info->dma) + if (!buffer_info->dma) { buffer_info->dma = pci_map_page(pdev, buffer_info->page, 0, buffer_info->length, PCI_DMA_FROMDEVICE); + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { + put_page(buffer_info->page); + dev_kfree_skb(skb); + buffer_info->page = NULL; + buffer_info->skb = NULL; + buffer_info->dma = 0; + adapter->alloc_rx_buff_failed++; + break; /* while !buffer_info->skb */ + } + } rx_desc = E1000_RX_DESC(*rx_ring, i); rx_desc->buffer_addr = cpu_to_le64(buffer_info->dma); @@ -4109,6 +4119,13 @@ map_skb: skb->data, buffer_info->length, PCI_DMA_FROMDEVICE); + if (pci_dma_mapping_error(pdev, buffer_info->dma)) { + dev_kfree_skb(skb); + buffer_info->skb = NULL; + buffer_info->dma = 0; + adapter->alloc_rx_buff_failed++; + break; /* while !buffer_info->skb */ + } /* * XXX if it was allocated cleanly it will never map to a
Check for error return from pci_map_single/pci_map_page and clean up. With this and the previous patch the driver was able to handle a significant percentage of errors (I set the fault injection rate to 10% and could still download large files at a reasonable speed). Signed-off-by: Anton Blanchard <anton@samba.org> --- I wasn't able to stress the jumbo frame path, so that part could do with some more eyes. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html