[LITMUS^RT] The Newest Litmus-rt Installation Problem

Bob Wu just4thin at gmail.com
Sat Sep 14 17:00:28 CEST 2013


Hi Glenn,
Thanks for your guidance, the new kernel finally works fine after I comment the printk(). I tried this on physical machine, and the crash happened as well, so it seems nothing wrong with vmware.
Unfortunately, I met another problem when I moved forward. I've tried an older litmus-rt version 2010.2, everything went fine during installation. However, the syscall.c cannot compile when I tried to install liblitmus. All syscall is not defined except _NR_gettid. So I searched the source code, and found those definition like _NR_set_rt_task_param were in litmus/unistd_32.h and litmus/unistd_64.h. Then I manually added the unistd_64.h to syscall.c, well, another symbol is not defined, _NR_LITMUS. I found it was defined in arch/arm/include/uapi/asm/unistd.h. It's very strange that the patch file of version 2010.2 patched this definition into x86 arch. Is that a mistake or I went wrong?
P.S. Before the installation of liblitmus, I copied litmus include folder to /usr/include, otherwise files like rt_param.h cannot be found. Am I doing right?

Regards.
-Bob


在 2013-9-13,19:53,Glenn Elliott <gelliott at cs.unc.edu> 写道:

> Interesting.  I wonder if vmware exposes an unusual.  Does the crash persist if you comment out the printf()?
> 
> -Glenn
> 
> 
> On Sep 13, 2013, at 12:54 AM, Lily <just4thin at gmail.com> wrote:
> 
>> Hi Glenn,
>> 
>> I've followed your guidance, and got this result:
>> 
>> 0xffffffff81234216 is in init_topology (litmus/affinity.c:28).
>> 23	} else {
>> 24		/* size = num bits in mask */
>> 25		neigh_info[cpu].size[i] =
>> 26			cpumask_weight((struct cpumask *)&neigh_info[cpu].neighbors[i]);
>> 27	}
>> 28	printk("CPU %d has %d neighbors at level %d. (mask = %lx)\n",
>> 29			cpu, neigh_info[cpu].size[i], i,
>> 30			*cpumask_bits(neigh_info[cpu].neighbors[i]));
>> 31	}
>> 32
>> 
>> Thx, regards.
>> -Bob
>> 
>> 在 2013年9月12日星期四,Glenn Elliott 写道:
>> Hi Bob,
>> 
>> It looks like you've compiled with CONFIG_DEBUG_INFO, right?  Let's do the following:
>> 
>> > cd litmus-rt
>> > gdb vmlinux
>> > (gdb) l *0xffffffff81234216  // that's a lower-case L, followed by the address of the line where the crash occurred (it immediately follows "RIP" in the panic message)
>> 
>> This should print out a line number of where the crash happened.
>> 
>> -Glenn
>> 
>> 
>> On Sep 12, 2013, at 5:38 AM, Lily <just4thin at gmail.com> wrote:
>> 
>> > Dear litmus-rt developers:
>> >
>> > I'm new to Litmus-rt and I am just beginning to get started, but I encountered some problems during installing the litmus-rt. I downloaded the most recently version of litmus-rt, 2013.1, and followed the installation instruction provided by wiki web. Somehow after I finished the first step, and reboot to get into the new kernel, it failed.
>> >
>> > My approach:
>> > #cd /usr/src
>> > #tar jxvf linux-3.10.5.tar.bz2
>> > #mv linux-3.10.5 litmus-rt
>> > #cd litmus-rt
>> > #patch -p1 <../litmus-rt-2013.1.patch
>> > #make menuconfig   ///where I chose all litmus options
>> > #make -j8
>> > #make modules_install install
>> >
>> > My OS is RHEL 6.0, and I run it in vmware. I've tried to build kernel without litmus-rt, and it worked successfully. The screen snap is in the attachment, I was wondering whether this is caused by function init_topology(), so I posted my CPU information as well.
>> >
>> > Hope for your grateful help.
>> >
>> > Regards.
>> > Bob Wu
>> >
>> > <cpuinfo.txt><Other 64-bit-2013-09-12-15-45-38.png>_______________________________________________
>> > litmus-dev mailing list
>> > litmus-dev at lists.litmus-rt.org
>> > https://lists.litmus-rt.org/listinfo/litmus-dev
>> 
>> 
>> _______________________________________________
>> litmus-dev mailing list
>> litmus-dev at lists.litmus-rt.org
>> https://lists.litmus-rt.org/listinfo/litmus-dev
>> _______________________________________________
>> litmus-dev mailing list
>> litmus-dev at lists.litmus-rt.org
>> https://lists.litmus-rt.org/listinfo/litmus-dev
> 
> _______________________________________________
> litmus-dev mailing list
> litmus-dev at lists.litmus-rt.org
> https://lists.litmus-rt.org/listinfo/litmus-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20130914/b59e51b7/attachment.html>


More information about the litmus-dev mailing list