s390 architecture support improvements
Quoted 16 lines · Heiko Carstens
Quoted 9 lines · Eugene Syromyatnikov
Hello.
This patch set introduces compat personality support on s390x along with decoders of s390-specific system calls.
Unfortunately, I am no s390 expert by any means, so I likely have missed something. I also have little idea regarding applications where these s390-specific system calls are used, so the feedback regarding quality of the decoders is also welcome.
Looks all good to me. However the
s390_sthyitest case fails:(quoted diff) +1
-1-s390_sthyi(STHYI_FC_CP_IFL_CAP, {/* header */ {infhflg1=00, ...}, /* machine */ {infmval1=0xf0, infmscps=2, infmdcps=0, infmsifl=90, infmdifl=2, ...}, /* partition */ {infpflg1=0x80, infpval1=0xfc, infppnum=63, infpscps=0, infpdcps=0, infpsifl=32, infpdifl=0, infppnam="\xe2\xf8\xf3\xd3\xd7\xf6\xf0\x40" /* "S83LP60 " */, ...}, /* hypervisor 1 */ {infyflg1=0xc0, infycpt=1, infyiflt=2, infysyid="\xc2\xd6\xc5\xe2\xf8\xf3\xf6\xf0" /* "BOES8360" */, infysifl=32, ...}, /* guest 1 */ {infgflg1=0x4, infgusid="\xe2\xf8\xf3\xf6\xf0\xf0\xf1\xf6" /* "S8360016" */, infgsifl=6, infgpflg=00, ...}}, [0], 0) = 0 +s390_sthyi(STHYI_FC_CP_IFL_CAP, {/* header */ {infhflg1=00, ...}, /* machine */ {infmval1=0xf0, infmscps=2, infmdcps=0, infmsifl=90, infmdifl=2, ...}, /* partition */ {infpflg1=0x80, infpval1=0xfc, infppnum=63, infpscps=0, infpdcps=0, infpsifl=32, infpdifl=0, infppnam="\xe2\xf8\xf3\xd3\xd7\xf6\xf0\x40" /* "S83LP60 " */, ...}, /* hypervisor 1 */ {infyflg1=0xc0, infysyid="\xc2\xd6\xc5\xe2\xf8\xf3\xf6\xf0" /* "BOES8360" */, infysifl=32, ...}, /* guest 1 */ {infgflg1=0x4, infgusid="\xe2\xf8\xf3\xf6\xf0\xf0\xf1\xf6" /* "S8360016" */, infgsifl=6, infgpflg=00, ...}}, [0], 0) = 0 +++ exited with 0 +++
Quoted 1 line
s390_sthyi.gen.test: failed test: ../../strace -e trace=s390_sthyi -a47 ../s390_sthyi output mismatch
Thank you for the report! Unfortunately, when I decided that infycpt/infyiflt aren't that important for abbreviated output, I forgot to update the test, and the lack of MT-enabled hardware didn't allow me to catch it early. I'll resend the patch with the fixed test.