On 5/11/2026 10:47 AM, Philippe Mathieu-Daudé wrote: > On 11/5/26 18:36, Cédric Le Goater wrote: >> On 5/11/26 17:58, Eric Farman wrote: >>> On Mon, 2026-05-11 at 17:48 +0200, Cédric Le Goater wrote: >>>> Hello, >>>> >>>> On 5/6/26 15:55, Philippe Mathieu-Daudé wrote: >>>>> Introduce a source set common to system / user. Start it >>>>> with the files built in both sets: 'cpu_models_user.c' >>>>> and 'gdbstub.c' No logical change intended. >>>>> >>>>> Signed-off-by: Philippe Mathieu-Daudé <[email protected]> >>>>> Reviewed-by: Ilya Leoshkevich <[email protected]> >>>>> Message-Id: <[email protected]> >>>> >>>> This change introduced a regression with PCI passthrough which >>>> stopped working. Guest kernel reports : >>>> >>>> [ 0.156501] zpci: PCI is not supported because CPU facilities >>>> 69 or 71 are not available >>>> >>>> and other devices (ap, ccw) are impacted too I think. >>>> >>>>> --- >>>>> target/s390x/meson.build | 25 +++++++++++++------------ >>>>> 1 file changed, 13 insertions(+), 12 deletions(-) >>>>> >>>>> diff --git a/target/s390x/meson.build b/target/s390x/meson.build >>>>> index 6f98ce34d73..44f58ac2919 100644 >>>>> --- a/target/s390x/meson.build >>>>> +++ b/target/s390x/meson.build >>>>> @@ -1,4 +1,9 @@ >>>>> s390x_ss = ss.source_set() >>>>> +s390x_common_ss = ss.source_set() >>>>> +s390x_common_system_ss = ss.source_set() >>>>> +s390x_system_ss = ss.source_set() >>>>> +s390x_user_ss = ss.source_set() >>>>> + >>>>> s390x_ss.add(files( >>>>> 'cpu.c', >>>>> 'cpu_features.c', >>>>> @@ -15,15 +20,16 @@ gen_features_h = custom_target('gen-features.h', >>>>> capture: true, >>>>> command: gen_features) >>>>> -s390x_ss.add(gen_features_h) >>>>> +s390x_common_ss.add(gen_features_h) >>>>> +s390x_common_ss.add(files( >>>>> + 'cpu_models_user.c', >>>>> + 'gdbstub.c', >>>>> +)) >>>>> -s390x_system_ss = ss.source_set() >>>>> s390x_system_ss.add(files( >>>>> 'ioinst.c', >>>>> )) >>>>> -s390x_common_system_ss = ss.source_set() >>>>> -s390x_common_system_ss.add(gen_features_h) >>>>> s390x_common_system_ss.add(files( >>>>> 'helper.c', >>>>> 'arch_dump.c', >>>>> @@ -32,19 +38,14 @@ s390x_common_system_ss.add(files( >>>>> 'mmu_helper.c', >>>>> 'sigp.c', >>>>> 'cpu-system.c', >>>>> - 'cpu_models_system.c', >>>> >>>> This file is compiled anymore and the CPU features and facilities are >>>> not propagated to the guest anymore. >>>> >>>> Without this patch, cpuinfo reports : >>>> >>>> features : esan3 zarch stfle msa ldisp eimm dfp edat etf3eh >>>> highgprs te vx vxd vxe gs vxe2 vxp sort dflt vxp2 nnpa >>>> facilities : 0 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 >>>> 19 20 21 22 23 24 25 26 27 28 30 31 32 33 34 35 36 37 38 40 41 42 43 >>>> 44 45 47 48 49 50 51 52 53 54 57 58 59 60 61 64 65 69 71 72 73 74 75 >>>> 76 77 78 80 81 82 129 130 131 133 134 135 138 139 140 146 147 148 >>>> 150 151 152 155 156 165 192 193 194 196 197 >>>> >>>> with : >>>> >>>> features : esan3 zarch stfle msa ldisp eimm dfp edat etf3eh >>>> highgprs te vx vxd vxe gs vxe2 vxp sort dflt vxp2 >>>> acilities : 0 1 2 3 4 6 7 8 9 10 11 13 14 16 17 18 19 20 21 >>>> 22 23 24 25 26 27 28 30 31 32 33 34 35 36 37 38 40 41 42 43 44 45 46 >>>> 47 48 49 50 51 52 53 54 55 57 58 59 60 61 64 72 73 74 75 76 77 78 80 >>>> 81 82 129 130 131 133 134 135 138 139 146 147 148 150 151 152 155 192 >>>> >>> >>> Thanks, Cedric... I wonder if the above is addressed by this patch? >>> (I haven't checked yet, but will >>> do that now.) >>> >>> https://lore.kernel.org/qemu-devel/20260510204640.2178926-1- >>> [email protected]/ >> >> Something like below seems to work. >> >> C. >> >> >> @@ -22,7 +22,6 @@ gen_features_h = custom_target('gen-feat >> >> s390x_common_ss.add(gen_features_h) >> s390x_common_ss.add(files( >> - 'cpu_models_user.c', >> 'gdbstub.c', >> )) >> >> @@ -38,8 +37,11 @@ s390x_common_system_ss.add(files( >> 'mmu_helper.c', >> 'sigp.c', >> 'cpu-system.c', >> + 'cpu_models_system.c', >> )) >> >> +s390x_user_ss.add(files('cpu_models_user.c')) >> + >> subdir('tcg') >> subdir('kvm') > > Right. I wonder how I could screw that change... > >
The only possible thing is that code becoming common skipped some ifdef that were here before per target. So, something that should have been poisoned has not been, and we definitely need to identify what it is to prevent other regressions like this. Regards, Pierrick
