Michał, On Fri, Mar 04, 2016 at 08:17:11PM +0100, Michał Zegan wrote: > Done, will test. Just for reference, was path0/5 truncated, or it could > be another problem?
With git, patch 0/X is typically the "cover letter", describing the rest of the patches (1/X through X/X). As such, there's usually nothing in there for the 'patch' utility to do, and if you're using 'git am path/to/folder/*.patch', you'll want to git am --skip over the 0/X cover letter. Not sure that's the "standard" way to do it, it's just a habit I developed :) Cheers, --Gabriel > > W dniu 04.03.2016 o 20:06, Marcel Apfelbaum pisze: > > On 03/04/2016 08:56 PM, Michał Zegan wrote: > >> How are you supposed to apply those patches? git am says the first one > >> is empty. > > > > I suggest cloning Laszlo's public branch including this series: > > > > https://github.com/lersek/edk2/commits/pcie_issue32_v1 > > > > Thanks, > > Marcel > > > >> > >> W dniu 04.03.2016 o 15:46, Laszlo Ersek pisze: > >>> The first two patches should fix > >>> <https://github.com/tianocore/edk2/issues/32>, based on Marcel's > >>> analysis in that item. (Thanks a lot for that again!) These patches are > >>> (and should be) independent of -D USE_OLD_PCI_HOST; that is, they are > >>> active regardless of the fallback that we're temporarily retaining. > >>> > >>> Michał, can you please test if the series solves your issue? > >>> > >>> Then, based on the first two patches, the other three expose ECAM to the > >>> firmware modules proper (DXE and later). These do depend on > >>> USE_OLD_PCI_HOST==FALSE (which is the default now, and the only > >>> supported build in the mid term). > >>> > >>> I tested & regression tested the series in a bunch of scenarios; the > >>> most interesting is probably the usual physical GPU assignment. On Q35, > >>> the core PciHostBridgeDxe now permits access to the full extended config > >>> space, and it works all right with the PCIe GTX750. > >>> > >>> Gabriel, can you please check if your Q35 OSX guest(s) continue working > >>> with this? > >>> > >>> Public branch: <https://github.com/lersek/edk2/commits/pcie_issue32_v1>. > >>> > >>> Cc: Gabriel Somlo <so...@cmu.edu> > >>> Cc: Jordan Justen <jordan.l.jus...@intel.com> > >>> Cc: Marcel Apfelbaum <mar...@redhat.com> > >>> Cc: Michał Zegan <webczat_...@poczta.onet.pl> > >>> Cc: Ruiyu Ni <ruiyu...@intel.com> > >>> > >>> Thanks! > >>> Laszlo > >>> > >>> Laszlo Ersek (5): > >>> OvmfPkg: IndustryStandard/Q35MchIch9.h: add PCIEXBAR macros > >>> OvmfPkg: PlatformPei: enable PCIEXBAR (aka MMCONFIG / ECAM) on Q35 > >>> OvmfPkg: add DxePciLibI440FxQ35 > >>> OvmfPkg: match PCI config access to machine type (if not > >>> USE_OLD_PCI_HOST) > >>> OvmfPkg: PciHostBridgeLib: permit access to the full extended config > >>> space > >>> > >>> > >>> OvmfPkg/Include/IndustryStandard/Q35MchIch9.h > >>> > >>> | 8 + > >>> > >>> OvmfPkg/Library/DxePciLibI440FxQ35/DxePciLibI440FxQ35.inf > >>> > >>> | 47 ++++++ > >>> {MdePkg/Library/BasePciLibCf8 => > >>> OvmfPkg/Library/DxePciLibI440FxQ35}/PciLib.c | 161 +++++++++++++++----- > >>> > >>> OvmfPkg/Library/PciHostBridgeLib/PciHostBridgeLib.c > >>> > >>> | 4 +- > >>> > >>> OvmfPkg/Library/PciHostBridgeLib/PciHostBridgeLib.inf > >>> > >>> | 1 + > >>> > >>> OvmfPkg/OvmfPkgIa32.dsc > >>> > >>> | 29 ++++ > >>> > >>> OvmfPkg/OvmfPkgIa32X64.dsc > >>> > >>> | 29 ++++ > >>> > >>> OvmfPkg/OvmfPkgX64.dsc > >>> > >>> | 29 ++++ > >>> > >>> OvmfPkg/PlatformPei/Platform.c > >>> > >>> | 81 ++++++++++ > >>> > >>> OvmfPkg/PlatformPei/PlatformPei.inf > >>> > >>> | 3 + > >>> 10 files changed, 356 insertions(+), 36 deletions(-) > >>> create mode 100644 > >>> OvmfPkg/Library/DxePciLibI440FxQ35/DxePciLibI440FxQ35.inf > >>> copy {MdePkg/Library/BasePciLibCf8 => > >>> OvmfPkg/Library/DxePciLibI440FxQ35}/PciLib.c (85%) > >>> > > _______________________________________________ edk2-devel mailing list edk2-devel@lists.01.org https://lists.01.org/mailman/listinfo/edk2-devel
