From: Andi Kleen Date: Tue, 23 Aug 2005 01:14:27 +0000 (+0200) Subject: [PATCH] x86_64: Don't oops at boot when empty Opteron node has IO X-Git-Tag: v2.6.13-rc7~5 X-Git-Url: https://git.openpandora.org/cgi-bin/gitweb.cgi?p=pandora-kernel.git;a=commitdiff_plain;h=d3813fcf105814d06b47fa586f6b61f3cff1cefc [PATCH] x86_64: Don't oops at boot when empty Opteron node has IO The code to detect IO links on Opteron would not check if the node had actually memory. This could lead to pci_bus_to_node returning an invalid node, which might cause crashes later when dma_alloc_coherent passes it to page_alloc_node(). The bug has been there forever but for some reason it is causing now crashes. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- diff --git a/arch/x86_64/pci/k8-bus.c b/arch/x86_64/pci/k8-bus.c index c2c38b579939..d80c323669e0 100644 --- a/arch/x86_64/pci/k8-bus.c +++ b/arch/x86_64/pci/k8-bus.c @@ -47,13 +47,22 @@ fill_mp_bus_to_cpumask(void) * if there are no busses hanging off of the current * ldt link then both the secondary and subordinate * bus number fields are set to 0. + * + * RED-PEN + * This is slightly broken because it assumes + * HT node IDs == Linux node ids, which is not always + * true. However it is probably mostly true. */ if (!(SECONDARY_LDT_BUS_NUMBER(ldtbus) == 0 && SUBORDINATE_LDT_BUS_NUMBER(ldtbus) == 0)) { for (j = SECONDARY_LDT_BUS_NUMBER(ldtbus); j <= SUBORDINATE_LDT_BUS_NUMBER(ldtbus); - j++) - pci_bus_to_node[j] = NODE_ID(nid); + j++) { + int node = NODE_ID(nid); + if (!node_online(node)) + node = 0; + pci_bus_to_node[j] = node; + } } } }