Messages in this thread | | | Date | Mon, 17 Mar 2014 11:39:05 +0100 | From | Jiri Olsa <> | Subject | Re: [PATCH 1/5] perf tests: Add tip/pid mmap automated tests |
| |
On Mon, Mar 17, 2014 at 01:50:12PM +0900, Namhyung Kim wrote: > Hi Jiri, > > On Fri, 14 Mar 2014 15:00:02 +0100, Jiri Olsa wrote: > > +static int thread_init(struct thread_data *td) > > +{ > > + void *map; > > + > > + map = mmap(NULL, page_size, PROT_READ|PROT_WRITE, > > + MAP_SHARED|MAP_ANONYMOUS, -1, 0); > > Shouldn't it be an executable mapping to be found by MAP__FUNCTION?
yea.. looks like my arch implies PROT_EXEC via PROT_READ, man mmap:
On some hardware architectures (e.g., i386), PROT_WRITE implies PROT_READ. It is architecture dependent whether PROT_READ implies PROT_EXEC or not. Portable programs should always set PROT_EXEC if they intend to execute code in the new mapping.
mm/mmap.c:
/* * Does the application expect PROT_READ to imply PROT_EXEC? * * (the exception is when the underlying filesystem is noexec * mounted, in which case we dont add PROT_EXEC.) */ if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC)) if (!(file && (file->f_path.mnt->mnt_flags & MNT_NOEXEC))) prot |= PROT_EXEC;
I'll set the PROT_EXEC as the man page says.
thanks, jirka
| |