[OpenAFS-devel] Crash on AIX 6.1 TL6 SP3

[Derrick Brashear]

wild guess: m_hdr changed size because of:

-        m_contig_t *mh_contig;          /* contig mpool support struct */
+        struct mclContig *mh_contig;            /* contig mpool support st=
ruct
*/

where
-typedef struct m_contig
-{
-        caddr_t va;             /* addr of contiguous block */
-        uint    elemsize;       /* size of the elements of the block */
-        uint    elemcnt;        /* # of elements in the block */
-        ulong   freecnt;        /* how many of these have been freed */
-        memreg_t memreg;        /* contains the mem reg corr for block */
-} m_contig_t;

+typedef struct mclContig {
+        caddr_t mclcCB;             /* addr of contiguous block */
+        uint32_t mclcElemSize;  /* size of the elements of the block */
+        uint32_t mclcElemCnt;   /* # of elements in the block */
+        memreg_t mclcMemReg;    /* contains the mem reg corr for block */
+} mclContig_t;

meaning struct mbuf shrank by 4 bytes

so the question is, what's the best way to get a mbuf provided to us
by the kernel that we can check out the size, and
then presumably we provide our own struct mbufs, so we can make
m_next, m_len, m_flags, m_ext, m_pkthdr and whatever
else be instead macros which dtrt given what we booted on... and screw
the whole "build twice" plan.

On Tue, Feb 22, 2011 at 5:24 AM, Niklas Edmundsson
<Niklas.Edmundsson@hpc2n.umu.se> wrote:
>
> Just a small heads up to let you know that the OpenAFS client on AIX 6.1
> Technology Level 6 crashes the machine on first usage.
>
> I've seen this with both an old 1.4.x client and the recent 1.4.14.
>
> The same client binaries works fine on AIX 6.1 TL4, but upgrading the
> machine to TL6 it dumps on first AFS usage:
>
> ------------8<-------------------
> CRASH INFORMATION:
> CPU 0 CSA F00000002FF47600 at time of crash, error code for LEDs: 7000000=
0
> pvthread+008800 STACK:
> [0001BF00]abend_trap+000000 ()
> [00450128]m_freem+0003C8 (??)
> [041CB884].hea_tx+000C44 ()
> [041D1FDC]hea_tx_atomic+00015C (??, ??, ??, ??, ??, ??)
> [041CDA98]hea_output+0003D8 (??, ??)
> [047B3038]047B3038 ()
> [045E7F50]045E7F50 ()
> [045E98F4]ip_output+0000D4 (??, ??, ??, ??, ??, ??)
> [0461D6CC]udp_output+0008AC (??, ??, ??, ??)
> [0461CB4C]udp_usrreq+00092C (??, ??, ??, ??, ??)
> [F1000000C034D2AC]osi_NetSend+00078C (F1000E00002BC408, F00000002FF45198,
> =A0 F1000A0021C60050, 0000000200000002, 0000000000000030, 000000000000000=
0)
> ------------8<-------------------
>
> I don't have much time to spend on this right now (it's just a test
> machine), suggestions on what might have changed are welcome.
>
> /Nikke
> --
> -=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D=
-
> =A0Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se =A0 =A0 | =A0 =A0nikke@h=
pc2n.umu.se
> -------------------------------------------------------------------------=
--
> =A0You're good, kid, but as long as I'm around.....
> =3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D=
-=3D
> _______________________________________________
> OpenAFS-devel mailing list
> OpenAFS-devel@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-devel
>



--=20
Derrick