| Message ID | 6fa81b4d-528d-5c33-50c5-a18396b4383a@suse.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | x86emul: further work | expand |
Paul,On 24.03.2020 13:26, Jan Beulich wrote: > Some of the later patches are still at least partly RFC, for > varying reasons (see there). I'd appreciate though if at least > some of the earlier ones could go in rather sooner than later. > > Patch 1 functionally (for the test harness) depends on > "libx86/CPUID: fix (not just) leaf 7 processing", while at > least patch 2 contextually depends on "x86emul: disable > FPU/MMX/SIMD insn emulation when !HVM". > > 1: x86emul: support AVX512_BF16 insns I should note that I also have a VP2INTERSECT patch ready, but the just released SDE segfaults when trying to test it. I'll be holding this back for some more time, I guess. > 2: x86emul: support MOVDIRI insn > 3: x86: determine HAVE_AS_* just once > 4: x86: move back clang no integrated assembler tests > 5: x86emul: support MOVDIR64B insn > 6: x86emul: support ENQCMD insn > 7: x86/HVM: scale MPERF values reported to guests (on AMD) > 8: x86emul: support RDPRU > 9: x86/HVM: don't needlessly intercept APERF/MPERF/TSC MSR reads > 10: x86emul: support MCOMMIT Paul, I should also note that I mistakenly Cc-ed your old Citrix address. I'd like to avoid re-posting the series - do you perhaps nevertheless get the xen-devel copies? Jan
> -----Original Message----- > From: Jan Beulich <jbeulich@suse.com> > Sent: 24 March 2020 12:43 > To: xen-devel@lists.xenproject.org > Cc: Andrew Cooper <andrew.cooper3@citrix.com>; Paul Durrant <paul@xen.org>; Wei Liu <wl@xen.org>; > Roger Pau Monne <roger.pau@citrix.com> > Subject: Re: [PATCH v5 00/10] x86emul: further work > > Paul,On 24.03.2020 13:26, Jan Beulich wrote: > > Some of the later patches are still at least partly RFC, for > > varying reasons (see there). I'd appreciate though if at least > > some of the earlier ones could go in rather sooner than later. > > > > Patch 1 functionally (for the test harness) depends on > > "libx86/CPUID: fix (not just) leaf 7 processing", while at > > least patch 2 contextually depends on "x86emul: disable > > FPU/MMX/SIMD insn emulation when !HVM". > > > > 1: x86emul: support AVX512_BF16 insns > > I should note that I also have a VP2INTERSECT patch ready, but the > just released SDE segfaults when trying to test it. I'll be holding > this back for some more time, I guess. > > > 2: x86emul: support MOVDIRI insn > > 3: x86: determine HAVE_AS_* just once > > 4: x86: move back clang no integrated assembler tests > > 5: x86emul: support MOVDIR64B insn > > 6: x86emul: support ENQCMD insn > > 7: x86/HVM: scale MPERF values reported to guests (on AMD) > > 8: x86emul: support RDPRU > > 9: x86/HVM: don't needlessly intercept APERF/MPERF/TSC MSR reads > > 10: x86emul: support MCOMMIT > > Paul, I should also note that I mistakenly Cc-ed your old Citrix > address. I'd like to avoid re-posting the series - do you perhaps > nevertheless get the xen-devel copies? > Yeah I have them. My filters just moved them into my general 'xen' mailbox but I got them. Paul > Jan
On 24/03/2020 12:26, Jan Beulich wrote: > Some of the later patches are still at least partly RFC, for > varying reasons (see there). I'd appreciate though if at least > some of the earlier ones could go in rather sooner than later. > > Patch 1 functionally (for the test harness) depends on > "libx86/CPUID: fix (not just) leaf 7 processing", while at > least patch 2 contextually depends on "x86emul: disable > FPU/MMX/SIMD insn emulation when !HVM". > > 1: x86emul: support AVX512_BF16 insns > 2: x86emul: support MOVDIRI insn > 3: x86: determine HAVE_AS_* just once I have (in this order when threaded): 02/10 Support AVX512_BF16 02/10 Support MOVDIRI 01/10 Support AVX512_BF16 03/10 Determine HAS_AS_* To a first approximation, the two AVX512_BF16 ones look identical, so I'm going to assume that the first one (chronologically) was the error. ~Andrew
On 25.03.2020 12:41, Andrew Cooper wrote: > On 24/03/2020 12:26, Jan Beulich wrote: >> Some of the later patches are still at least partly RFC, for >> varying reasons (see there). I'd appreciate though if at least >> some of the earlier ones could go in rather sooner than later. >> >> Patch 1 functionally (for the test harness) depends on >> "libx86/CPUID: fix (not just) leaf 7 processing", while at >> least patch 2 contextually depends on "x86emul: disable >> FPU/MMX/SIMD insn emulation when !HVM". >> >> 1: x86emul: support AVX512_BF16 insns >> 2: x86emul: support MOVDIRI insn >> 3: x86: determine HAVE_AS_* just once > > I have (in this order when threaded): > > 02/10 Support AVX512_BF16 > 02/10 Support MOVDIRI > 01/10 Support AVX512_BF16 > 03/10 Determine HAS_AS_* > > To a first approximation, the two AVX512_BF16 ones look identical, so > I'm going to assume that the first one (chronologically) was the error. Yes indeed, sorry. Jan
On 24.03.2020 13:26, Jan Beulich wrote: > Some of the later patches are still at least partly RFC, for > varying reasons (see there). I'd appreciate though if at least > some of the earlier ones could go in rather sooner than later. > > Patch 1 functionally (for the test harness) depends on > "libx86/CPUID: fix (not just) leaf 7 processing", while at > least patch 2 contextually depends on "x86emul: disable > FPU/MMX/SIMD insn emulation when !HVM". > > 1: x86emul: support AVX512_BF16 insns Thanks for the ack on patch 1, which I've just pushed, but may I also ask for an ack or otherwise on the (test harness) prereq named above, and ideally also the other patch named there? Thanks, Jan