On 03/11/15 12:49, Leif Lindholm wrote: > On 11/03/15 11:28, Alexander Graf wrote: >>>>>> Note that qemu commit aa685789 [1] mentioned in patches #2 and #4 is >>>>>> pending a pull [2]. The commit hash is expected to stay stable [3]. >>>> >>>> (Meanwhile the qemu patch in question has been merged, indeed with the >>>> same commit hash.) >>>> >>>>> Will enabling XHCI cause any issues on earlier qemu versions, or will >>>>> they simply fall back to EHCI? >>>> >>>> Are you okay with the answers from Alex and myself? >>>> >>>> http://thread.gmane.org/gmane.comp.bios.tianocore.devel/13160 >>>> http://thread.gmane.org/gmane.comp.bios.tianocore.devel/13161 >>> >>> Sorry, yes. >>> Breaking in a small window of commits is not a problem. >>> >>>> Ard asked for more details in the commit message of 1/4, and I'll write >>>> them up. Would you like to see any changes beyond that? >>> >>> I know this is a bit tedious, but if it is feasible, could we list the >>> range of affected QEMU commits in the commit message? >>> >>> I would quite like to see people start making proper use of the PCI >>> support, before 2.3 is out. >> >> 2.3-rc0 should be out any day, so how about we just declare PCI support >> usable as of that one? > > Works for me. > But can we have the "breaks with" commit mentioned?
I'm certainly willing to name any qemu commits in the edk2 commit message, but I'm a bit confused. The qemu commit range (well, it's not really a "range", because the qemu commit history is -- clearly -- not linear, it contains merges) would be "we have PCI, but not the XHCI fix". That would be from: 4ab29b82 arm: Add PCIe host bridge in virt machine (inclusive) until aa685789 xhci: generate a Transfer Event for each Transfer TRB with the IOC bit set (exclusive); ie. git log --no-merges --oneline --reverse 4ab29b82^...aa685789^ (Note the carets on both sides; git revision ranges are usually expressed with the precisely reversed inclusive/exclusive sense -- the LHS is usually meant as exclusive, and the RHS is usually meant as inclusive.) That git-log command returns a quite messy commit log; it contains 16 merges. On the other hand, as Alex mentioned earlier, both of these patches (or, more precisely for Alex's PCIe patch: its containing series) were merged between 2.2 and 2.3. So 2.2 has neither PCIe-on-arm-virt nor the xHCI fix, and 2.3 will have both. Maybe this makes mentioning the exact PCIe introduction commit superfluous? (I can also drop the mention of aa685789, and just say "this is supposed to work in qemu 2.3".) Thanks Laszlo ------------------------------------------------------------------------------ Dive into the World of Parallel Programming The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ edk2-devel mailing list edk2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/edk2-devel