[Top] [All Lists]

Re: 2.6.14 - problem with malta

To: "Kevin D. Kissell" <>
Subject: Re: 2.6.14 - problem with malta
From: "Kishore K" <>
Date: Mon, 27 Mar 2006 20:12:05 +0600
Cc: "Ralf Baechle" <>,
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta;; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=qGglGeqXRI/VUJXkDfE9eg5iGJhH1DChHo8QD15A000MEhhQ97koPyZHk43vTngyqkuB1svBCiLBcjG/d1o/DSKSVy7QvDvQfLIBhCouoySXTcouYuKpPHvSBmFeJmp0X5ZcWvl9bUnyfr/fuGcr1CY6m+q5Uniy8e+CFpxCIKc=
In-reply-to: <005901c6519f$b226e060$10eca8c0@grendel>
Original-recipient: rfc822;
References: <> <> <> <000d01c65022$90d758a0$10eca8c0@grendel> <> <005901c6519f$b226e060$10eca8c0@grendel>
On 3/27/06, Kevin D. Kissell <> wrote:
My own Malta kernels are built from rather old 2.6 sources, so it's not
clear that all the same options are required/possible.  The questions that
come to mind for me are:
1) You're booting a big-endian kernel.  The Malta board can be set up
    for either endianness with a DIP switch, but I believe that it ships
    from the factory in little-endian mode.  Have you verified that the
    endianness setup?

After reset, I get the message saying that, board is in Big endian mode. So,  I assume that there is no problem here.

2)  You're configuring for hot-pluggable CPUs.  That isn't fully supported
     in the old source base I've got.  It may be fixed in 2.6.14, but do you
     actually need it on a Malta?
3) You've got support for MIPS MT (the multithreading extension) enabled,
    with the slave-virtual-processor boot loader, etc.  This stuff only works
    on a 34K core, and could cause severe problems if invoked on a 4Kc.
    At least you don't have SMP/SMTC multi-virtual-processor support
    enabled, which would *definitely* cause a prompt crash on a 4Kc,
    but please turn MIPS_MT off.  If MIPS MT + VPE loader are really
    selected options in the default malta config file, it needs to be fixed!
            Kevin K.

I reset the above options, but the same problem persists. If I compiled 2.6.10 with default configuration (endian is changed to big endian), the board comes up without any problem.


<Prev in Thread] Current Thread [Next in Thread>