[Bug 219153] head, stable/11, release/11.0.1: libkvm (& more?) not updated to handle powerpc/powerpc64 ET_DYN based vmcore.* 's and such

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=219153

--- Comment #14 from Mark Millard <markmi at dsl-only.net> ---
(In reply to John Baldwin from comment #8)

With the following hacks I've been able to get
an output for the debug.minidump=0 based
vmcore.2 (2 GiBYte RAM dumped) for
powerpc (32-bit) FreeBSD via:

# ps -aux -M /var/crash/vmcore.2 -N /usr/lib/debug/boot/kernel/kernel.debug 
USER    PID %CPU %MEM   VSZ   RSS TT  STAT STARTED    TIME COMMAND
root      0  0.0  0.0     0     0  -  DLs  31Dec69 0:00.11 [kernel]
root      1  0.0  0.0  5400   792  -  DLs  31Dec69 0:00.06 [init]
root      2  0.0  0.0     0     0  -  DL   31Dec69 0:00.00 [crypto]
root      3  0.0  0.0     0     0  -  DL   31Dec69 0:00.00 [crypto returns]
. . .
markmi 1086  0.0  0.2 14192  4168  -  D    31Dec69 0:00.00 [sshd]
markmi 1087  0.0  0.1  7048  2812  -  Ds   31Dec69 0:00.00 [sh]
root   1088  0.0  0.1  7900  2660  -  D    31Dec69 0:00.00 [su]
root   1089  0.0  0.1  7048  2800  -  D+   31Dec69 0:00.00 [sh]

(The STARTED column output is odd above.)

Be warned that I've not done much FreeBSD
coding and so am not familiar with the coding
standards and/or guidelines and just did
my own thing.

# svnlite diff /usr/src/lib/libkvm/kvm_private.c
Index: /usr/src/lib/libkvm/kvm_private.c
===================================================================
--- /usr/src/lib/libkvm/kvm_private.c   (revision 317820)
+++ /usr/src/lib/libkvm/kvm_private.c   (working copy)
@@ -128,7 +128,9 @@
 {

        return (kd->nlehdr.e_ident[EI_CLASS] == class &&
-           kd->nlehdr.e_type == ET_EXEC &&
+           (  kd->nlehdr.e_type == ET_EXEC ||
+              kd->nlehdr.e_type == ET_DYN
+           ) &&
            kd->nlehdr.e_machine == machine);
 }

Then in:

static int
_powerpc_kvatop(kvm_t *kd, kvaddr_t va, off_t *ofs)

I did:

# svnlite diff /usr/src/lib/libkvm/kvm_powerpc.c
Index: /usr/src/lib/libkvm/kvm_powerpc.c
===================================================================
--- /usr/src/lib/libkvm/kvm_powerpc.c   (revision 317820)
+++ /usr/src/lib/libkvm/kvm_powerpc.c   (working copy)
@@ -209,6 +209,53 @@
        if (be32toh(vm->ph->p_paddr) == 0xffffffff)
                return ((int)powerpc_va2off(kd, va, ofs));

+       // HACK in something for what I observe in
+       // a debug.minidump=0 vmcore.* for 32-bit powerpc
+       //
+       if (  be32toh(vm->ph->p_vaddr)  == 0xffffffff
+          && be32toh(vm->ph->p_paddr)  == 0
+          && be16toh(vm->eh->e_phnum)  == 1
+          ) {
+               // Presumes p_memsz is either unsigned
+               // 32-bit or is 64-bit, same for va .
+
+               if (be32toh(vm->ph->p_memsz) <= va)
+                       return 0; // Like powerpc_va2off
+
+               // If ofs was (signed) 32-bit there
+               // would be a problem for sufficiently
+               // large postive memsz's and va's
+               // near the end --because of p_offset
+               // and dmphdrsz causing overflow/wrapping
+               // for some large va values.
+               // Presumes 64-bit ofs for such cases.
+               // Also presumes dmphdrsz+p_offset
+               // is non-negative so that small
+               // non-negative va values have no
+               // problems with ofs going negative.
+
+               *ofs =    vm->dmphdrsz
+                       + be32toh(vm->ph->p_offset)
+                       + va;
+
+               // The normal return value overflows/wraps
+               // for p_memsz == 0x80000000u when va == 0 .
+               // Avoid this by depending on calling code's
+               // loop for sufficiently large cases.
+               // This code presumes p_memsz/2 <= MAX_INT .
+               // 32-bit powerpc FreeBSD does not allow
+               // using more than 2 GiBytes of RAM but
+               // does allow using 2 GiBytes on 64-bit
+               // hardware.
+               //
+               if (  (int)be32toh(vm->ph->p_memsz) < 0
+                  && va < be32toh(vm->ph->p_memsz)/2
+                  )
+                       return be32toh(vm->ph->p_memsz)/2;
+
+               return be32toh(vm->ph->p_memsz) - va;
+       }
+
        _kvm_err(kd, kd->program, "Raw corefile not supported");
        return (0);
 }

(Technically I combined this with my clang
targeting powerpc 32-bit testing by building
it as a clang based buildworld.)

I do not claim the code is appropriate for
FreeBSD use but it might get me closer to
investigating why production-style kernel
builds (gcc 4.2.1 based) panic on a pid 11
(idle process) thread every so often, even
when world was also gcc 4.2.1 based.

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