2013-09-11 16:43:47

by Christophe Leroy

Subject: [PATCH] powerpc 8xx: Reverting commit e0908085fc2391c85b85fb814ae1df377c8e0dcb which has become useless

The commit e0908085fc2391c85b85fb814ae1df377c8e0dcb is not needed anymore.
The issue was because dcbst wrongly sets the store bit when causing a DTLB
error, but this is now fixed by commit 0a2ab51ffb8dfdf51402dcfb446629648c96bc78
which handles the buggy dcbx instructions on data page faults on the 8xx.

Signed-off-by: Christophe Leroy <[email protected]>

diff -ur linux-3.11.org/arch/powerpc/mm/pgtable.c linux-3.11/arch/powerpc/mm/pgtable.c
--- linux-3.11.org/arch/powerpc/mm/pgtable.c 2013-09-02 22:46:10.000000000 +0200
+++ linux-3.11/arch/powerpc/mm/pgtable.c 2013-09-09 11:25:57.000000000 +0200
@@ -32,8 +32,6 @@
#include <asm/tlbflush.h>
#include <asm/tlb.h>

-#include "mmu_decl.h"
-
static inline int is_exec_fault(void)
{
return current->thread.regs && TRAP(current->thread.regs) == 0x400;
@@ -72,7 +70,7 @@
* support falls into the same category.
*/

-static pte_t set_pte_filter(pte_t pte, unsigned long addr)
+static pte_t set_pte_filter(pte_t pte)
{
pte = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS);
if (pte_looks_normal(pte) && !(cpu_has_feature(CPU_FTR_COHERENT_ICACHE) ||
@@ -81,17 +79,6 @@
if (!pg)
return pte;
if (!test_bit(PG_arch_1, &pg->flags)) {
-#ifdef CONFIG_8xx
- /* On 8xx, cache control instructions (particularly
- * "dcbst" from flush_dcache_icache) fault as write
- * operation if there is an unpopulated TLB entry
- * for the address in question. To workaround that,
- * we invalidate the TLB here, thus avoiding dcbst
- * misbehaviour.
- */
- /* 8xx doesn't care about PID, size or ind args */
- _tlbil_va(addr, 0, 0, 0);
-#endif /* CONFIG_8xx */
flush_dcache_icache_page(pg);
set_bit(PG_arch_1, &pg->flags);
}
@@ -111,7 +98,7 @@
* as we don't have two bits to spare for _PAGE_EXEC and _PAGE_HWEXEC so
* instead we "filter out" the exec permission for non clean pages.
*/
-static pte_t set_pte_filter(pte_t pte, unsigned long addr)
+static pte_t set_pte_filter(pte_t pte)
{
struct page *pg;

@@ -193,7 +180,7 @@
* this context might not have been activated yet when this
* is called.
*/
- pte = set_pte_filter(pte, addr);
+ pte = set_pte_filter(pte);

/* Perform the setting of the PTE */
__set_pte_at(mm, addr, ptep, pte, 0);
Only in linux-3.11/arch/powerpc/mm/: pgtable.c.orig

2013-09-12 08:14:01

by Joakim Tjernlund

Subject: Re: [PATCH] powerpc 8xx: Reverting commit e0908085fc2391c85b85fb814ae1df377c8e0dcb which has become useless

Christophe Leroy <[email protected]> wrote on 2013/09/11 17:50:28:

> From: Christophe Leroy <[email protected]>
> To: Benjamin Herrenschmidt <[email protected]>, Paul Mackerras
<[email protected]>,
> Cc: [email protected], [email protected], Joakim
Tjernlund <[email protected]>
> Date: 2013/09/11 18:43
> Subject: [PATCH] powerpc 8xx: Reverting commit
e0908085fc2391c85b85fb814ae1df377c8e0dcb which has become useless
>
> The commit e0908085fc2391c85b85fb814ae1df377c8e0dcb is not needed
anymore.
> The issue was because dcbst wrongly sets the store bit when causing a
DTLB
> error, but this is now fixed by commit
0a2ab51ffb8dfdf51402dcfb446629648c96bc78
> which handles the buggy dcbx instructions on data page faults on the
8xx.
>
> Signed-off-by: Christophe Leroy <[email protected]>

Acked-by: Joakim Tjernlund <[email protected]>

Jocke