This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: Role of svc_entry macro with 64 bytes in kprobes/systemtap on ARM
- From: Josh Stone <jistone at redhat dot com>
- To: systemtap at sourceware dot org
- Date: Thu, 12 Dec 2013 12:21:26 -0800
- Subject: Re: Role of svc_entry macro with 64 bytes in kprobes/systemtap on ARM
- Authentication-results: sourceware.org; auth=none
- References: <CAAKGpwR=BAKGDDP3S6eKmWgvZc2f+iALDn4SBVF0Koq-m-bujA at mail dot gmail dot com>
On 12/11/2013 07:10 PM, Jey Jay wrote:
> Hi,
>
> Currently Im working on "systemtap" and "kprobes".
>
> My current work is to convert the kprobes in to loadable kernel
> module(.ko) on arm architecture.
>
> Converted module is able to run the examples in
> "http://lxr.free-electrons.com/source/samples/kprobes/?a=arm".
>
> But few systemtap scripts like
>
> 1. https://sourceware.org/systemtap/SystemTap_Beginners_Guide/traceiosect.html
> 2. https://sourceware.org/systemtap/SystemTap_Beginners_Guide/mainsect-disk.html#disktop
>
> producing kernel panics.
In case you aren't aware, stap has the runtime option -DSTP_ALIBI to
exculpate itself from such panics. It basically compiles out the probe
handlers, leaving just an atomic counter to track probe hits.
Sharing the details of your kernel panics would also help.
But I'm sorry, I don't know about svc_entry myself...
Josh