mmap /proc/<pid>/mem ... alternatives?

From: Zinx Verituse (zinx@epicsol.org)
Date: Fri Jun 07 2002 - 02:38:51 EST

Next message: Dag Nygren: "Re: Devfs strangeness in 2.4.18"
Previous message: Erik Andersen: "Re: [PATCH] initcall dependency solution."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

I have a black-box process (of which I can disassemble enough to
insert debugging traps). I need to read a large number of sparse
structures that are a few bytes to a few hundred kbyte each. In
Linux 2.2, I was simply mmap()ing the process's memory, but that
feature seems removed in Linux 2.4, so.. I need alternatives..

Here's a list of things I'm considering:

provide mmap() patch for /proc/<pid>/mem
Definitely out of my league :)

mmap() of /dev/mem
needs root (I can live with that)

need to know where the process's memory is --
I don't know how to determine this.

shared memory
need to bootstrap some code at the beginning (doable)

I don't know how to re-map the data, bss, and stack to
shared memory properly.

use read/write on /proc/<pid>/mem
write() doesn't work (doesn't write to the process).

complicates the code in my program quite a bit.

        hybrid read, ptrace(PTRACE_POKEDATA, ...)
                works, but I would like to avoid all the extra copies,
                  and this complicates the code quite a bit.

downgrade to 2.2
Only as an absolute last resort, if nothing else works.

-- 
Zinx Verituse
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Next message: Dag Nygren: "Re: Devfs strangeness in 2.4.18"
Previous message: Erik Andersen: "Re: [PATCH] initcall dependency solution."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

This archive was generated by hypermail 2b29 : Fri Jun 07 2002 - 22:00:30 EST