On 07/14/15 23:15, Paolo Bonzini wrote: >> The long delay that Alex reported (for the case when all guest memory >> was set to UC up-front) is due to the fact that the SEC phase of OVMF >> decompresses an approximately 1712 KB sized, LZMA-compressed blob, to >> approx. 896 KB worth of PEI drivers and 8192 KB worth of DXE and UEFI >> drivers -- and this decompression is extremely memory-intensive. >> >> (When Jordan implemented that reset vector first, we saw similar >> performance degradation on AMD hosts (albeit not due to MTRR but due to >> page attributes). See >> <https://github.com/tianocore/edk2/commit/98f378a7>. I'm only mentioning >> it here because it makes me appreciate the current problem report.) >> >> Anyway, the reset vector's page table building is implemented in >> "OvmfPkg/ResetVector/Ia32/PageTables64.asm". The decompression in SEC >> can be found in "OvmfPkg/Sec/SecMain.c", function DecompressMemFvs(). > > Perhaps the OVMF reset vector should initialize the MTRRs for the BSP? That's an idea, yes, if someone feels sufficiently drawn to writing assembly. Complications: - the reset vector is specific to OvmfPkg only in the OvmfPkgX64.dsc build - it needs to be determined what memory to cover. > I think SEC doesn't do any MMIO, so it should be enough to enable MTRRs > and set the default type to writeback. Seems safe to me, off the top of my head (and testing could confirm / disprove quickly). > In any case we're going to have to quirk it, because of the broken > guests in the wild. Thanks. Laszlo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
