[Bug runtime/31215] New: @__compat_task misbehaves on RHEL9 x86_64

[Bug runtime/31215] New: @__compat_task misbehaves on RHEL9 x86_64

[mcermak at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=31215

            Bug ID: 31215
           Summary: @__compat_task misbehaves on RHEL9 x86_64
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
          Assignee: systemtap at sourceware dot org
          Reporter: mcermak at redhat dot com
  Target Milestone: ---

Apparently the @__compat_task macro misbehaves on RHEL9.  See how
_stp_is_compat_task() returns 0 although it should clearly return 1 (the last
column in systemtap output below). GCC was invoked with -m32:

(Note that the 943211848ea34e4d5d10f0094dc84b47667d84ed version of
sched_rr_get_interval.c had to be slightly updated to avoid SEGV by commenting
out the latest subtest)




el9 x86_64 # gcc 
/root/systemtap/testsuite/systemtap.syscall/sched_rr_get_interval.c   -m32 -lrt
-Wno-stringop-overflow  -lm  -o  /tmp/sched_rr_get_interval -g
el9 x86_64 # stap --poison-cache -ge 'probe syscall.mmap2 {printdln(", ", name,
_stp_syscall_nr(), syscall_name(_stp_syscall_nr()), execname(), probefunc(),
%{_stp_is_compat_task()%})}' -c /tmp/sched_rr_get_interval 
/tmp/stap06kHAR/stap_ad583362bc2821360e315a0e4befae13_166569_src.o: warning:
objtool: _stp_vsprint_memory+0x220: call to __get_user_nocheck_1() with UACCESS
enabled
/tmp/stap06kHAR/stap_ad583362bc2821360e315a0e4befae13_166569_src.o: warning:
objtool: .altinstr_replacement+0x36: call to copy_user_generic_string() with
UACCESS enabled
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 192, lgetxattr, sched_rr_get_in, 0xffffffffabf83e40, 0
mmap2, 9, mmap, stapio, 0xffffffffabc32720, 0
mmap2, 9, mmap, stapio, 0xffffffffabc32720, 0
mmap2, 9, mmap, stapio, 0xffffffffabc32720, 0
el9 x86_64 #




Now, here is how this works fine on rhel8:




el8 x86_64 # gcc 
/root/systemtap/testsuite/systemtap.syscall/sched_rr_get_interval.c   -m32 -lrt
-Wno-stringop-overflow  -lm  -o  /tmp/sched_rr_get_interval -g
el8 x86_64 # stap --poison-cache -ge 'probe syscall.mmap2 {printdln(", ", name,
_stp_syscall_nr(), syscall_name(_stp_syscall_nr()), execname(), probefunc(),
%{_stp_is_compat_task()%})}' -c /tmp/sched_rr_get_interval
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
el8 x86_64 # stap --poison-cache -ge 'probe syscall.mmap2 {printdln(", ", name,
_stp_syscall_nr(), syscall_name(_stp_syscall_nr()), execname(), probefunc(),
%{_stp_is_compat_task()%})}' -c /tmp/sched_rr_get_interval
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 192, mmap2, sched_rr_get_in, 0xffffffff8c0e9d50, 1
mmap2, 9, mmap, stapio, 0xffffffff8be2e730, 0
mmap2, 9, mmap, stapio, 0xffffffff8be2e730, 0
mmap2, 9, mmap, stapio, 0xffffffff8be2e730, 0
el8 x86_64 #

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug runtime/31215] @__compat_task misbehaves on RHEL9 x86_64

[mcermak at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=31215

Martin Cermak <mcermak at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|systemtap at sourceware dot org    |mcermak at redhat dot com

--- Comment #1 from Martin Cermak <mcermak at redhat dot com> ---
I have a patch for this here:
https://sourceware.org/bugzilla/show_bug.cgi?id=31117#c6

-- 
You are receiving this mail because:
You are the assignee for the bug.