Message ID | 20200214213313.66622-1-sean.v.kelley@linux.intel.com (mailing list archive) |
---|---|
Headers | show |
Series | Add boot interrupt quirk mechanism for Xeon chipsets | expand |
Sean V Kelley <sean.v.kelley@linux.intel.com> writes: > When IRQ lines on secondary or higher IO-APICs are masked (e.g., > Real-Time threaded interrupts), many chipsets redirect IRQs on > this line to the legacy PCH and in turn the base IO-APIC in the > system. The unhandled interrupts on the base IO-APIC will be > identified by the Linux kernel as Spurious Interrupts and can > lead to disabled IRQ lines. > > Disabling this legacy PCI interrupt routing is chipset-specific and > varies in mechanism between chipset vendors and across generations. > In some cases the mechanism is exposed to BIOS but not all BIOS > vendors choose to pick it up. With the increasing usage of RT as it > marches towards mainline, additional issues have been raised with > more recent Xeon chipsets. > > This patchset disables the boot interrupt on these Xeon chipsets where > this is possible with an additional mechanism. In addition, this > patchset includes documentation covering the background of this quirk. Well done! The documentation is really appreciated! Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Thomas Gleixner <tglx@linutronix.de> writes: > Sean V Kelley <sean.v.kelley@linux.intel.com> writes: >> When IRQ lines on secondary or higher IO-APICs are masked (e.g., >> Real-Time threaded interrupts), many chipsets redirect IRQs on >> this line to the legacy PCH and in turn the base IO-APIC in the >> system. The unhandled interrupts on the base IO-APIC will be >> identified by the Linux kernel as Spurious Interrupts and can >> lead to disabled IRQ lines. >> >> Disabling this legacy PCI interrupt routing is chipset-specific and >> varies in mechanism between chipset vendors and across generations. >> In some cases the mechanism is exposed to BIOS but not all BIOS >> vendors choose to pick it up. With the increasing usage of RT as it >> marches towards mainline, additional issues have been raised with >> more recent Xeon chipsets. >> >> This patchset disables the boot interrupt on these Xeon chipsets where >> this is possible with an additional mechanism. In addition, this >> patchset includes documentation covering the background of this quirk. > > Well done! The documentation is really appreciated! > > Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Bjorn, this should go into stable as well IMO. Thanks, tglx