All of lore.kernel.org
 help / color / mirror / Atom feed
* [parisc-linux] Making the stack executable
@ 2000-08-08 18:30 David Huggins-Daines
  0 siblings, 0 replies; only message in thread
From: David Huggins-Daines @ 2000-08-08 18:30 UTC (permalink / raw)
  To: parisc-linux

Hi,

It turns out that my mysterious lockups in the rt_sigreturn path were
due to the fact that our stack pages are not executable.

They are being mapped with PAGE_COPY protection in exec.c, like every
other architecture does, but unlike every other architecture, our
PAGE_COPY did not include execute permissions.

Sorry folks but we do need trampolines.

What is more disturbing is that without _PAGE_EXEC set in PAGE_COPY,
attempting to execute code on the stack just hangs the machine solid!
(without printing out any error messages or register dumps)

The following program should demonstrate this problem:

	.LEVEL 1.1
	.text
	.align 4
	.globl main
main:
	.PROC
	.CALLINFO FRAME=64
	stw	%rp,-20(%sp)
	copy	%sp,%r19
	ldo	64(%sp),%sp
	ldil	L%0xebe0c002,%r20	/* bv,n %r0(%r31) */
	ldo	R%0xebe0c002(%r20),%r20
	stw	%r20,0(%r19)
	fdc	%r0(%r19)
	sync
	fic	%r0(%r19)
	sync
	ble	0(%sr7,%r19)
	nop
	ldo	-64(%sp),%sp
	bv,n	%r0(%rp)
	.ENTRY
	.EXIT
	.PROCEND

Is this some kind of issue with our ITLB miss handlers?

I'll check in the change to <asm/pgtable.h> shortly, but I'd like
someone to also look into the problem above (without the change).

-- 
dhd@linuxcare.com, http://www.linuxcare.com/
Linuxcare. Support for the revolution.

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2000-08-09  4:05 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2000-08-08 18:30 [parisc-linux] Making the stack executable David Huggins-Daines

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.