Hi Joerg, There is no easy way to check whether this BIOS bug is present, so we have to list them explicitly. I think we can do just like what platform drivers(ex. asus_quirks[1]) do, check the system vendor and product name to identify those machines, and give them a working ioapic value.
1. drivers/platform/x86/asus-nb-wmi.c Best regards, AceLan Kao. 2015-05-04 19:08 GMT+08:00 Joerg Roedel <[email protected]>: > Hi Acelan, > > On Thu, Apr 23, 2015 at 10:15:33AM +0800, AceLan Kao wrote: >> We can confirm that the issue comes from the resources conflict, >> adding this boot option can solve the issue. >> ivrs_ioapic[5]=00:00.1 >> We've tried to negotiate with the OEM BIOS team, but they refused to >> modify the BIOS code, since there is no issue on MS Windows and it'll >> require to re-start a full test if the BIOS code is changed, so they >> don't want to do that. >> >> I'm wondering if we can add a quirk to list those machine and assign >> them the correct ivrs_ioapic value in amd_iommu driver? > > It is probably hard to match a specific machine which has the bug. What > do you think would be a good indicator to check whether this BIOS bug is > present? > > > Joerg > _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
