Latest working kernel version: Earliest failing kernel version: Distribution: Debian Hardware Environment: Intel ClassicR+ with Intel EtherExpress 16 Software Environment: Debian testing (lenny) Problem Description: I was unable to access a computer containing an Intel EtherExpress 16 network card using IPv6. I traced this to failure of neighbour discovery. When I used an "ip -6 neigh add" command, on the computer attempting access, to insert a binding between the IPv6 address of the computer with the Intel EtherExpress 16 network card and the card's ethernet address, I was able to access that computer using IPv6. Neighbour discovery requires working multicast. The driver sources file eexpress.c contains an approximately 30 line function eexp_setup_filter used when loading multicast addresses. I found 3 problems in this function 1) It wrote the number of multicast addresses to the card instead of the number of bytes in the multicast addresses. 2) When loading multiple multicast addresses it loaded the first one provided multiple times instead of loading each one once. 3) The setting of pointer 'data' from 'dmi->dmi_addr' occured before the test for the error situation of 'dmi' being NULL. Correcting these problems allows the computer with the Intel EtherExpress 16 network card to found by IPv6 neighbour discovery. p.s. There is some information on the Intel EtherExpress 16 at http://www.intel.com/support/etherexpress/vintage/sb/cs-013500.htm Datasheet for the Intel 82586 ethernet controller used by the card http://www.datasheetcatalog.com/datasheets_pdf/8/2/5/8/82586.shtml --- eexpress.c.000 2008-01-24 22:58:37.000000000 +0000 +++ eexpress.c 2008-04-30 00:17:51.000000000 +0100 @@ -201,9 +201,9 @@ 0x0000,Cmd_MCast, 0x0076, /* link to next command */ #define CONF_NR_MULTICAST 0x44 - 0x0000, /* number of multicast addresses */ + 0x0000, /* number of bytes in multicast address(es) */ #define CONF_MULTICAST 0x46 0x0000, 0x0000, 0x0000, /* some addresses */ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, @@ -1568,9 +1568,9 @@ } static void eexp_setup_filter(struct net_device *dev) { - struct dev_mc_list *dmi = dev->mc_list; + struct dev_mc_list *dmi; unsigned short ioaddr = dev->base_addr; int count = dev->mc_count; int i; if (count > 8) { @@ -1579,19 +1579,20 @@ count = 8; } outw(CONF_NR_MULTICAST & ~31, ioaddr+SM_PTR); - outw(count, ioaddr+SHADOW(CONF_NR_MULTICAST)); - for (i = 0; i < count; i++) { - unsigned short *data = (unsigned short *)dmi->dmi_addr; + outw(6*count, ioaddr+SHADOW(CONF_NR_MULTICAST)); + for (i = 0, dmi = dev->mc_list; i < count; i++, dmi = dmi->next) { + unsigned short *data; if (!dmi) { printk(KERN_INFO "%s: too few multicast addresses\n", de v->name); break; } if (dmi->dmi_addrlen != ETH_ALEN) { printk(KERN_INFO "%s: invalid multicast address length g iven.\n", dev->name); continue; } + data = (unsigned short *)dmi->dmi_addr; outw((CONF_MULTICAST+(6*i)) & ~31, ioaddr+SM_PTR); outw(data[0], ioaddr+SHADOW(CONF_MULTICAST+(6*i))); outw((CONF_MULTICAST+(6*i)+2) & ~31, ioaddr+SM_PTR); outw(data[1], ioaddr+SHADOW(CONF_MULTICAST+(6*i)+2));
Created attachment 15985 [details] patch to file eexpress.c Here's the patch as an attached file
Reply-To: akpm@linux-foundation.org thanks. Please do submit patches via mail in future - merging them from bugzilla is a hassle. Please email a signed-off-by: for this patch as per Documentation/SubmittingPatches?
The patch has been included in 2.6.26-rc2. http://kernel.org/pub/linux/kernel/v2.6/testing/ChangeLog-2.6.26-rc2