RFC: ESR_I/DLK processing

arch/powerpc/kernel/head_fsl_booke.S has the following code:

	/* Data Storage Interrupt */
	START_EXCEPTION(DataStorage)
	NORMAL_EXCEPTION_PROLOG
	mfspr	r5,SPRN_ESR		/* Grab the ESR, save it, pass =

arg3 */

	stw	r5,_ESR(r11)
	mfspr	r4,SPRN_DEAR		/* Grab the DEAR, save it, pass =

arg2 */

	andis.	r10,r5,(ESR_ILK|ESR_DLK)@h
	bne	1f
	EXC_XFER_EE_LITE(0x0300, handle_page_fault)
1:
	addi	r3,r1,STACK_FRAME_OVERHEAD
	EXC_XFER_EE_LITE(0x0300, CacheLockingException)

I need something similar for A2 (and all book3e) and was wondering, why =
this isn't just:

unsigned long address,

 	}
 #endif
=20
+#ifdef CONFIG_PPC_BOOK3E
+	if (error_code & (ESR_DLK|ESR_ILK)) {
+		/* detect that this is a privileged op and SIGILL */
+		_exception(SIGILL, regs, ILL_PRVOPC, regs->nip);
+		return 0;
+	}
+#endif
+
 	if (notify_page_fault(regs))
 		return 0;

Its not like this need to be fast or anything.
I'd be happy to submit a patch that adds to fault.c and removed the =
I/DLK processing from head_fsl_booke.S

Thoughts?
-jx

On Oct 28, 2011, at 3:43 PM, Jimi Xenidis wrote:

arch/powerpc/kernel/head_fsl_booke.S has the following code:

quoted

	/* Data Storage Interrupt */
	START_EXCEPTION(DataStorage)
	NORMAL_EXCEPTION_PROLOG
	mfspr	r5,SPRN_ESR		/* Grab the ESR, save it, pass =

arg3 */

quoted

	stw	r5,_ESR(r11)
	mfspr	r4,SPRN_DEAR		/* Grab the DEAR, save it, pass =

arg2 */

quoted

	andis.	r10,r5,(ESR_ILK|ESR_DLK)@h
	bne	1f
	EXC_XFER_EE_LITE(0x0300, handle_page_fault)
1:
	addi	r3,r1,STACK_FRAME_OVERHEAD
	EXC_XFER_EE_LITE(0x0300, CacheLockingException)

=20
=20
I need something similar for A2 (and all book3e) and was wondering, =

why this isn't just:

=20

quoted

diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
index 88abe70..8451822 100644
--- a/arch/powerpc/mm/fault.c
+++ b/arch/powerpc/mm/fault.c

@@ -159,6 +159,14 @@ int __kprobes do_page_fault(struct pt_regs =

*regs, unsigned long address,

quoted

	}
#endif
=20
+#ifdef CONFIG_PPC_BOOK3E
+	if (error_code & (ESR_DLK|ESR_ILK)) {
+		/* detect that this is a privileged op and SIGILL */
+		_exception(SIGILL, regs, ILL_PRVOPC, regs->nip);
+		return 0;
+	}
+#endif
+
	if (notify_page_fault(regs))
		return 0;

=20
Its not like this need to be fast or anything.
I'd be happy to submit a patch that adds to fault.c and removed the =

I/DLK processing from head_fsl_booke.S

=20
Thoughts?
-jx

Probably because at one point in time DSI had a fast path handling for =
us.  I've got no issues w/the proposed patch, just remember to mixup the =
CONFIG_PPC_BOOK3E as we don't define that in 32-bit fsl-booke=20

- k=