Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux

iommu/vt-d: add quirk for broken interrupt remapping on 55XX chipsets

A few years back intel published a spec update:
http://www.intel.com/content/dam/doc/specification-update/5520-and-5500-chipset-ioh-specification-update.pdf

For the 5520 and 5500 chipsets which contained an errata (specificially errata
53), which noted that these chipsets can't properly do interrupt remapping, and
as a result the recommend that interrupt remapping be disabled in bios. While
many vendors have a bios update to do exactly that, not all do, and of course
not all users update their bios to a level that corrects the problem. As a
result, occasionally interrupts can arrive at a cpu even after affinity for that
interrupt has be moved, leading to lost or spurrious interrupts (usually
characterized by the message:
kernel: do_IRQ: 7.71 No irq handler for vector (irq -1)

There have been several incidents recently of people seeing this error, and
investigation has shown that they have system for which their BIOS level is such
that this feature was not properly turned off. As such, it would be good to
give them a reminder that their systems are vulnurable to this problem. For
details of those that reported the problem, please see:
https://bugzilla.redhat.com/show_bug.cgi?id=887006

[ Joerg: Removed CONFIG_IRQ_REMAP ifdef from early-quirks.c ]

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Prarit Bhargava <prarit@redhat.com>
CC: Don Zickus <dzickus@redhat.com>
CC: Don Dutile <ddutile@redhat.com>
CC: Bjorn Helgaas <bhelgaas@google.com>
CC: Asit Mallick <asit.k.mallick@intel.com>
CC: David Woodhouse <dwmw2@infradead.org>
CC: linux-pci@vger.kernel.org
CC: Joerg Roedel <joro@8bytes.org>
CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
CC: Arkadiusz Miśkiewicz <arekm@maven.pl>
Signed-off-by: Joerg Roedel <joro@8bytes.org>

authored by

Neil Horman and committed by
Joerg Roedel
03bbcb2e bd5cdad0

+40
+2
arch/x86/include/asm/irq_remapping.h
··· 28 28 29 29 extern void setup_irq_remapping_ops(void); 30 30 extern int irq_remapping_supported(void); 31 + extern void set_irq_remapping_broken(void); 31 32 extern int irq_remapping_prepare(void); 32 33 extern int irq_remapping_enable(void); 33 34 extern void irq_remapping_disable(void); ··· 55 54 56 55 static inline void setup_irq_remapping_ops(void) { } 57 56 static inline int irq_remapping_supported(void) { return 0; } 57 + static inline void set_irq_remapping_broken(void) { } 58 58 static inline int irq_remapping_prepare(void) { return -ENODEV; } 59 59 static inline int irq_remapping_enable(void) { return -ENODEV; } 60 60 static inline void irq_remapping_disable(void) { }
+20
arch/x86/kernel/early-quirks.c
··· 18 18 #include <asm/apic.h> 19 19 #include <asm/iommu.h> 20 20 #include <asm/gart.h> 21 + #include <asm/irq_remapping.h> 21 22 22 23 static void __init fix_hypertransport_config(int num, int slot, int func) 23 24 { ··· 193 192 } 194 193 #endif 195 194 195 + static void __init intel_remapping_check(int num, int slot, int func) 196 + { 197 + u8 revision; 198 + 199 + revision = read_pci_config_byte(num, slot, func, PCI_REVISION_ID); 200 + 201 + /* 202 + * Revision 0x13 of this chipset supports irq remapping 203 + * but has an erratum that breaks its behavior, flag it as such 204 + */ 205 + if (revision == 0x13) 206 + set_irq_remapping_broken(); 207 + 208 + } 209 + 196 210 #define QFLAG_APPLY_ONCE 0x1 197 211 #define QFLAG_APPLIED 0x2 198 212 #define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED) ··· 237 221 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs }, 238 222 { PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS, 239 223 PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd }, 224 + { PCI_VENDOR_ID_INTEL, 0x3403, PCI_CLASS_BRIDGE_HOST, 225 + PCI_BASE_CLASS_BRIDGE, 0, intel_remapping_check }, 226 + { PCI_VENDOR_ID_INTEL, 0x3406, PCI_CLASS_BRIDGE_HOST, 227 + PCI_BASE_CLASS_BRIDGE, 0, intel_remapping_check }, 240 228 {} 241 229 }; 242 230
+10
drivers/iommu/intel_irq_remapping.c
··· 524 524 525 525 if (disable_irq_remap) 526 526 return 0; 527 + if (irq_remap_broken) { 528 + WARN_TAINT(1, TAINT_FIRMWARE_WORKAROUND, 529 + "This system BIOS has enabled interrupt remapping\n" 530 + "on a chipset that contains an erratum making that\n" 531 + "feature unstable. To maintain system stability\n" 532 + "interrupt remapping is being disabled. Please\n" 533 + "contact your BIOS vendor for an update\n"); 534 + disable_irq_remap = 1; 535 + return 0; 536 + } 527 537 528 538 if (!dmar_ir_support()) 529 539 return 0;
+6
drivers/iommu/irq_remapping.c
··· 19 19 int irq_remapping_enabled; 20 20 21 21 int disable_irq_remap; 22 + int irq_remap_broken; 22 23 int disable_sourceid_checking; 23 24 int no_x2apic_optout; 24 25 ··· 210 209 if (amd_iommu_irq_ops.prepare() == 0) 211 210 remap_ops = &amd_iommu_irq_ops; 212 211 #endif 212 + } 213 + 214 + void set_irq_remapping_broken(void) 215 + { 216 + irq_remap_broken = 1; 213 217 } 214 218 215 219 int irq_remapping_supported(void)
+2
drivers/iommu/irq_remapping.h
··· 32 32 struct msi_msg; 33 33 34 34 extern int disable_irq_remap; 35 + extern int irq_remap_broken; 35 36 extern int disable_sourceid_checking; 36 37 extern int no_x2apic_optout; 37 38 extern int irq_remapping_enabled; ··· 90 89 91 90 #define irq_remapping_enabled 0 92 91 #define disable_irq_remap 1 92 + #define irq_remap_broken 0 93 93 94 94 #endif /* CONFIG_IRQ_REMAP */ 95 95