> > static void > get_netdump_regs_arm(struct bt_info *bt, ulong *eip, ulong *esp) > { > machdep->get_stack_frame(bt, eip, esp); > } > > If we're not going to use the ELF notes, why fail if they're not > there? I'm not sure what you mean about reading "directly from within the PT_LOAD segments", but I agree with you that the preliminary "if" section of get_netdump_regs_arm() which sets bt->machdep is a no-op -- given that arm_get_dumpfile_stack_frame() just ignores it, and resets bt->machdep from that ARM-specific ms->crash_task_regs area. Mika, you agree?

----- Original Message ----- > On Tue, Sep 20, 2011 at 10:41:51AM -0400, Dave Anderson wrote: > > > > > > static void > > > get_netdump_regs_arm(struct bt_info *bt, ulong *eip, ulong *esp) > > > { > > >   machdep->get_stack_frame(bt, eip, esp); > > > } > > > > > > If we're not going to use the ELF notes, why fail if they're not > > > there? > > > > I'm not sure what you mean about reading "directly from within the > > PT_LOAD segments",  but I agree with you that the preliminary "if" > > section of get_netdump_regs_arm() which sets bt->machdep is a > > no-op -- given that arm_get_dumpfile_stack_frame() just ignores > > it, and resets bt->machdep from that ARM-specific ms->crash_task_regs > > area. > > > > Mika, you agree? > > Absolutely - we already have similar for diskdump: > > static void > get_diskdump_regs_arm(struct bt_info *bt, ulong *eip, ulong *esp) > { >       machdep->get_stack_frame(bt, eip, esp); > } > > I tested Per's suggestion on my few kdump/netdump vmcores and it > works well. > > If you prefer to have a patch for this change, it is included below > (all credits goes to Per). > > diff --git a/netdump.c b/netdump.c > index 1e9960c..a194364 100644 > --- a/netdump.c > +++ b/netdump.c > @@ -2580,34 +2580,6 @@ get_netdump_regs_ppc64(struct bt_info *bt, > ulong *eip, ulong *esp) >  static void >  get_netdump_regs_arm(struct bt_info *bt, ulong *eip, ulong *esp) >  { > -     Elf64_Nhdr *note; > -     size_t len; > - > -     if ((bt->task == tt->panic_task) || > -             (is_task_active(bt->task) && nd->num_prstatus_notes > 1)) { > -             /* > -              * Registers are saved during the dump process for the > -              * panic task. Whereas in kdump, regs are captured for all > -              * CPUs if they responded to an IPI. > -              */ > -                if (nd->num_prstatus_notes > 1) { > -                     if (!nd->nt_prstatus_percpu[bt->tc->processor]) > -                             error(FATAL, > -                                   "cannot determine NT_PRSTATUS ELF note " > -                                   "for %s task: %lx\n", > -                                    (bt->task == tt->panic_task) ? > -                                    "panic" : "active", bt->task); > -                        note = (Elf64_Nhdr *) > - >                                nd->nt_prstatus_percpu[bt->tc->processor]; > -             } else > -                     note = (Elf64_Nhdr *)nd->nt_prstatus; > - > -             len = sizeof(Elf64_Nhdr); > -             len = roundup(len + note->n_namesz, 4); > -             bt->machdep = (void *)((char *)note + len + > -                     MEMBER_OFFSET("elf_prstatus", "pr_reg")); > -     } > - >       machdep->get_stack_frame(bt, eip, esp); >  } Queued for crash-5.1.9 -- thanks guys,  Dave -- Crash-utility mailing list Crash-utility(a)redhat.com https://www.redhat.com/mailman/listinfo/crash-utility