This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: [BUG] syscall.unlink no longer works after upgrading kernel to 3.7.3
- From: Josh Stone <jistone at redhat dot com>
- To: Zheng Da <zhengda1936 at gmail dot com>
- Cc: Mark Wielaard <mjw at redhat dot com>, agentzh <agentzh at gmail dot com>, "systemtap at sourceware dot org" <systemtap at sourceware dot org>
- Date: Tue, 28 May 2013 13:22:21 -0700
- Subject: Re: [BUG] syscall.unlink no longer works after upgrading kernel to 3.7.3
- References: <CAB4Tn6PdW3GOa09z_tfjQs=F+7XLOqMr5+c5GourX5e0v8FMeQ at mail dot gmail dot com> <1360054656 dot 3837 dot 13 dot camel at bordewijk dot wildebeest dot org> <51114188 dot 60400 at redhat dot com> <CAFLer83DQhCQg7Y3NKR0EUYePzp+fETDTeYEthUXKarAySM0_g at mail dot gmail dot com> <20130528191449 dot GA31042 at toonder dot wildebeest dot org> <CAFLer81NE1bocCbufPTtLLZ-pZz2kVA5r3rKoCgjmc_6w+fwng at mail dot gmail dot com>
On 05/28/2013 12:56 PM, Zheng Da wrote:
> Sorry, I forgot to mention the version of systemtap I'm using.
> I'm using v2.2.1. I checked its source code and the patch you
> mentioned has been applied.
> Does the patch only work with gcc 4.7?
I'm not sure if the -mfentry bug afflicted earlier gcc, but if so that
patch should work just as well.
But there are other reasons for "not accessible" errors, as described in
"man error::dwarf". Just optimization or inlining can be enough to make
variables unavailable from certain addresses.
Are you able to look at the DWARF dump to see where that variable can be
reached? The specific symptom of -mfentry is that the locations don't
start until exactly 5 bytes after the start of the function.
It would also help if you gave us your script code snippet that led to
the error. By $sdev, I'm guessing this is a scsi probe - but I can't
find that exact line in our scsi tapset, so I guess this is your code.
Josh