Re: Last week: i686-pc-linux-gnu 'gdb.server' testsuite regressions
Kevin Buettner <[email protected]>
| Newsgroups | gmane.comp.gdb.devel |
|---|---|
| Message-ID | <[email protected]> |
On Thu, 2 May 2019 19:59:29 -0700 Kevin Buettner <[email protected]> wrote: > On Sat, 20 Apr 2019 11:30:38 +0200 > Thomas Schwinge <[email protected]> wrote: > > > > In > > > d970ee2bae1925bb9265d37adef0b92e2678d666..02e902e1a1ec7b74125f329b3faef1992efb6d51 > > > (that is, basically, last week) I'm too seeing a good number of > > > 'gdb.server' testsuite regressions in native i686-pc-linux-gnu testing: > > > > These i686-pc-linux-gnu gdbserver regressions are caused by the commit > > 3f52fdbcb599f76b4838020721ca6c9f1cc28f84 "Fix amd64->i386 linux syscall > > restart problem" changes to 'gdb/gdbserver/linux-x86-low.c'. > > I did a i686-pc-linux-gnu build (though still on a 64-bit host), but > was unable to reproduce these regressions. Do I need to do the build > on a 32-bit host? It seems that the answer is "yes". Or maybe there's some additional configuration that I needed to do when attempting to build on a 64-bit host. This is what I'm seeing in the log file for gdb.server/abspath.exp: 0xb7fd6130 in _start () from target:/lib/ld-linux.so.2 (gdb) break main Breakpoint 1 at 0x8049169: file /ironwood1/sourceware-git/f30-i386-syscallrestart/bld/../../worktree-syscallrestart/gdb/testsuite/gdb.server/normal.c, line 23. (gdb) continue Continuing. Program received signal SIGSEGV, Segmentation fault. 0xb7fd6eaa in _dl_start () from target:/lib/ld-linux.so.2 (gdb) FAIL: gdb.server/abspath.exp: continue to main Switching back to /mesquite2/sourceware-git/f30-i386-syscallrestart/bld/gdb/testsuite. Remote debugging from host ::1, port 44692 monitor exit Now that I've reproduced it, I'll try to figure out what's going on... Kevin