From: Chris L. <ch...@ex...> - 2005-03-23 00:00:02
|
I'm observing strace(1) behaving oddly under recent 2.4.x versions. Under 2.4.27 with uml-patch-2.4.27-1 and strace 4.5.8, we get: # strace /bin/echo execve("/bin/echo", ["/bin/echo"], [/* 13 vars */]) = 0 exit_group(0) = ? which isn't very useful. Under 2.4.26 + uml-patch-2.4.26-3, the output is as normal: # strace /bin/echo execve("/bin/echo", ["/bin/echo"], [/* 13 vars */]) = 0 uname({sys="Linux", node="vs1", ...}) = 0 brk(0) = 0x804c32c # etc. etc. etc. write(1, "\n", 1 ) = 1 munmap(0x40018000, 4096) = 0 exit_group(0) = ? Later versions (e.g. 2.4.29 + uml-patch-2.4.27-1) show the same behaviour as 2.4.27 above. I haven't looked at the -bb patches exhaustively yet but certainly the same problem occurs with uml-2.4.27-bs2-pre6. Even on the kernels where strace is broken, gdb seems to work (at least to the extent of being able to set breakpoints and break on them -- I haven't tested it extensively). Has anyone else observed this and looked into it in more detail? -- ``Because that's where the money is.'' (Willie Sutton, explaining why he robbed banks) |