On Sat, Oct 24, 2009 at 01:43:48PM -0700, Jonathan Adams wrote: > > My next thought is to generate a bunch of calls to mdb -k from the > > shell and format everything together using awk, perl, or python. Won't > > be fast, will be kind of hacky, but probably will work. > > > > Is there a better way to approach this? > > One way would be to change ::lminfo to have a flag to prevent truncation, > or to just write a DCMD which gives the output in exactly the format > you want. > > You could even just write a module based on the ::lminfo implementation, > and do: > > echo "::load /path/to/my/module; ::mydcmd" | mdb -k > > Let me know if you need help going down this path.
I'm giving this a stab. It looks like the primary change I'll need to make is just to increase the size of LM_VNPATHLEN: http://arthur2.bludgeon.org/~rayvd/devel/mdb/raytest.c This doesn't _quite_ compile yet: cc -xstrconst -K pic -D_KERNEL -c ../common/raytest.c "../common/raytest.c", line 24: warning: implicit function declaration: mdb_pid2proc "../common/raytest.c", line 26: warning: implicit function declaration: mdb_vnode2path "../common/raytest.c", line 49: undefined symbol: MOD_API_VERSION "../common/raytest.c", line 49: non-constant initializer: op "NAME" It doesn't look like mdb_pid2proc or mdb_vnode2path are exposed by the mdb API. Should I recreate the headers for the functions and link against some library? Or completely redefine the functions locally? Perhaps I need to download a copy of the mdb source. I am using the API reference here: http://docs.sun.com/app/docs/doc/817-2543/api-5?a=view Thanks, Ray
