> Date: Fri, 15 Aug 2025 14:06:06 +0200
> From: Martin Pieuchot <[email protected]>
> 
> On 15/08/25(Fri) 13:17, Mark Kettenis wrote:
> > > Date: Fri, 15 Aug 2025 11:51:18 +0200
> > > From: Martin Pieuchot <[email protected]>
> > > 
> > > On 15/08/25(Fri) 09:39, Miod Vallat wrote:
> > > > > Please don't.  Keeping that page read-only is important for security.
> > > > > Maybe if nobody cares about the amd64 and i386 pmaps we should just
> > > > > delete those architectures?
> > > > 
> > > > But remember, because the end argument was wrong (sz instead of va +
> > > > sz), this call did *nothing*.
> > > > 
> > > > At least the commented out code will be correct now.
> > > 
> > > Exactly.  Since you committed this code Mark it does nothing.  That's
> > > what I said in the original report it's dead code.
> > 
> > The original code (before I "broke" it) did an unmap of the memory.
> > BTW that means we need to fix the comment.
> 
> Indeed. I'd be glad if you could do that when enable the uvm_map_protect()
> call.

ok?


Index: kern/kern_exec.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_exec.c,v
retrieving revision 1.265
diff -u -p -r1.265 kern_exec.c
--- kern/kern_exec.c    4 Aug 2025 04:59:31 -0000       1.265
+++ kern/kern_exec.c    15 Aug 2025 13:43:29 -0000
@@ -871,8 +871,10 @@ exec_sigcode_map(struct process *pr)
         * memory) that we keep a permanent reference to and that we map
         * in all processes that need this sigcode. The creation is simple,
         * we create an object, add a permanent reference to it, map it in
-        * kernel space, copy out the sigcode to it and unmap it.  Then we map
-        * it with PROT_EXEC into the process just the way sys_mmap would map 
it.
+        * kernel space, copy out the sigcode to it and map it PROT_READ
+        * such that the coredump code can write it out into core dumps.
+        * Then we map it with PROT_EXEC into the process just the way
+        * sys_mmap would map it.
         */
        if (sigobject == NULL) {
                extern int sigfillsiz;
@@ -899,7 +901,7 @@ exec_sigcode_map(struct process *pr)
                }
                memcpy((caddr_t)va, sigcode, sz);
 
-               (void) uvm_map_protect(kernel_map, va, round_page(sz),
+               (void) uvm_map_protect(kernel_map, va, round_page(va + sz),
                    PROT_READ, 0, FALSE, FALSE);
                sigcode_va = va;
                sigcode_sz = round_page(sz);

Reply via email to