aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorGary Hade <garyhade@us.ibm.com>2008-05-12 16:57:46 -0400
committerJesse Barnes <jbarnes@virtuousgeek.org>2008-06-10 13:59:50 -0400
commitbb71ad880204b79d60331d3384103976e086cb9f (patch)
tree612b14aed172db36ddbdae2287ffdd65e6baef00
parent5ca5c02f0e81c094c19d30dc0d13be4e929a994a (diff)
PCI: boot parameter to avoid expansion ROM memory allocation
Contention for scarce PCI memory resources has been growing due to an increasing number of PCI slots in large multi-node systems. The kernel currently attempts by default to allocate memory for all PCI expansion ROMs so there has also been an increasing number of PCI memory allocation failures seen on these systems. This occurs because the BIOS either (1) provides insufficient PCI memory resource for all the expansion ROMs or (2) provides adequate PCI memory resource for expansion ROMs but provides the space in kernel unexpected BIOS assigned P2P non-prefetch windows. The resulting PCI memory allocation failures may be benign when related to memory requests for expansion ROMs themselves but in some cases they can occur when attempting to allocate space for more critical BARs. This can happen when a successful expansion ROM allocation request consumes memory resource that was intended for a non-ROM BAR. We have seen this happen during PCI hotplug of an adapter that contains a P2P bridge where successful memory allocation for an expansion ROM BAR on device behind the bridge consumed memory that was intended for a non-ROM BAR on the P2P bridge. In all cases the allocation failure messages can be very confusing for users. This patch provides a new 'pci=norom' kernel boot parameter that can be used to disable the default PCI expansion ROM memory resource allocation. This provides a way to avoid the above described issues on systems that do not contain PCI devices for which drivers or user-level applications depend on the default PCI expansion ROM memory resource allocation behavior. Signed-off-by: Gary Hade <garyhade@us.ibm.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
-rw-r--r--Documentation/kernel-parameters.txt3
-rw-r--r--arch/x86/pci/common.c22
-rw-r--r--arch/x86/pci/pci.h1
3 files changed, 26 insertions, 0 deletions
diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index e07c432c731f..9cf7b34f2db0 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -1496,6 +1496,9 @@ and is between 256 and 4096 characters. It is defined in the file
1496 Use with caution as certain devices share 1496 Use with caution as certain devices share
1497 address decoders between ROMs and other 1497 address decoders between ROMs and other
1498 resources. 1498 resources.
1499 norom [X86-32,X86_64] Do not assign address space to
1500 expansion ROMs that do not already have
1501 BIOS assigned address ranges.
1499 irqmask=0xMMMM [X86-32] Set a bit mask of IRQs allowed to be 1502 irqmask=0xMMMM [X86-32] Set a bit mask of IRQs allowed to be
1500 assigned automatically to PCI devices. You can 1503 assigned automatically to PCI devices. You can
1501 make the kernel exclude IRQs of your ISA cards 1504 make the kernel exclude IRQs of your ISA cards
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 6e64aaf00d1d..3a5261bdff5d 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -121,6 +121,21 @@ void __init dmi_check_skip_isa_align(void)
121 dmi_check_system(can_skip_pciprobe_dmi_table); 121 dmi_check_system(can_skip_pciprobe_dmi_table);
122} 122}
123 123
124static void __devinit pcibios_fixup_device_resources(struct pci_dev *dev)
125{
126 struct resource *rom_r = &dev->resource[PCI_ROM_RESOURCE];
127
128 if (pci_probe & PCI_NOASSIGN_ROMS) {
129 if (rom_r->parent)
130 return;
131 if (rom_r->start) {
132 /* we deal with BIOS assigned ROM later */
133 return;
134 }
135 rom_r->start = rom_r->end = rom_r->flags = 0;
136 }
137}
138
124/* 139/*
125 * Called after each bus is probed, but before its children 140 * Called after each bus is probed, but before its children
126 * are examined. 141 * are examined.
@@ -128,7 +143,11 @@ void __init dmi_check_skip_isa_align(void)
128 143
129void __devinit pcibios_fixup_bus(struct pci_bus *b) 144void __devinit pcibios_fixup_bus(struct pci_bus *b)
130{ 145{
146 struct pci_dev *dev;
147
131 pci_read_bridge_bases(b); 148 pci_read_bridge_bases(b);
149 list_for_each_entry(dev, &b->devices, bus_list)
150 pcibios_fixup_device_resources(dev);
132} 151}
133 152
134/* 153/*
@@ -483,6 +502,9 @@ char * __devinit pcibios_setup(char *str)
483 else if (!strcmp(str, "rom")) { 502 else if (!strcmp(str, "rom")) {
484 pci_probe |= PCI_ASSIGN_ROMS; 503 pci_probe |= PCI_ASSIGN_ROMS;
485 return NULL; 504 return NULL;
505 } else if (!strcmp(str, "norom")) {
506 pci_probe |= PCI_NOASSIGN_ROMS;
507 return NULL;
486 } else if (!strcmp(str, "assign-busses")) { 508 } else if (!strcmp(str, "assign-busses")) {
487 pci_probe |= PCI_ASSIGN_ALL_BUSSES; 509 pci_probe |= PCI_ASSIGN_ALL_BUSSES;
488 return NULL; 510 return NULL;
diff --git a/arch/x86/pci/pci.h b/arch/x86/pci/pci.h
index 720c4c554534..291dafec07b7 100644
--- a/arch/x86/pci/pci.h
+++ b/arch/x86/pci/pci.h
@@ -27,6 +27,7 @@
27#define PCI_CAN_SKIP_ISA_ALIGN 0x8000 27#define PCI_CAN_SKIP_ISA_ALIGN 0x8000
28#define PCI_USE__CRS 0x10000 28#define PCI_USE__CRS 0x10000
29#define PCI_CHECK_ENABLE_AMD_MMCONF 0x20000 29#define PCI_CHECK_ENABLE_AMD_MMCONF 0x20000
30#define PCI_NOASSIGN_ROMS 0x40000
30 31
31extern unsigned int pci_probe; 32extern unsigned int pci_probe;
32extern unsigned long pirq_table_addr; 33extern unsigned long pirq_table_addr;