[PATCH] [RFC][V3] bluegene: use MMU feature flag to conditionalize L1 writethrough code

From: Eric Van Hensbergen
Date: Tue Jun 07 2011 - 17:36:38 EST

Next message: Benjamin Herrenschmidt: "Re: [PATCHES v3] pci/of: Generic PCI <-> OF matching"
Previous message: Luck, Tony: "RE: [PATCH v2 3/3] efi: Add support for using efivars as a pstorebackend"
Next in thread: Benjamin Herrenschmidt: "Re: [PATCH] [RFC][V3] bluegene: use MMU feature flag toconditionalize L1 writethrough code"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

BG/P nodes need to be configured for writethrough to work in SMP
configurations. This patch adds the right hooks in the MMU code
to make sure BGP_L1_WRITETHROUGH configurations are setup for BG/P.

RFC note: this essentially just changes the ifdefs to use the
BEGIN_MMU_FTR_SECTION macros. A couple of things that I really didn't
like about this:
a) we introduced at least one extra op that isn't needed to get around
otherwise having multiple labels
b) we are introducting a bunch of no-ops in places that could be critical
paths and jimix says this may not be the best thing for multiple reasons
including having no-ops around the DCBZs is a bad thing
c) the ELSE_MMU_FTR_SECTION stuff appears to be broken (or I don't know
how to use it, it gave me the error:
Error: non-constant expression in ".if" statement
so I switched out the else clauses with redundant FTR_SECTIONS
(one for IFSET and on for IFCLR). Please someone throw me a clue
as to what I was doing wrong. I'm running gcc 4.3.2 from crosstools-ng
if it has some sort of impact.

Jimix has thrown me some code to try and do a better job by branching
to stub code inside of the MMU_FTRs so I don't have so many no-ops. I'm
open to alternatives. jimix also suggested changing NEED_L1_WRITETHROUGH
to DCBZ_BROKEN, which I'm open to if you think appropriate, or maybe
DCBZ_BROKEN_DAMNIT would be more apt.

Thanks for any help.

Signed-off-by: Eric Van Hensbergen <ericvh@xxxxxxxxx>
---
arch/powerpc/include/asm/mmu-44x.h | 2 ++
arch/powerpc/kernel/head_44x.S | 27 +++++++++++++++++++++++++--
arch/powerpc/kernel/misc_32.S | 24 +++++++++++++++++++++++-
arch/powerpc/lib/copy_32.S | 13 +++++++++++--
arch/powerpc/mm/44x_mmu.c | 7 +++++--
5 files changed, 66 insertions(+), 7 deletions(-)

diff --git a/arch/powerpc/include/asm/mmu-44x.h b/arch/powerpc/include/asm/mmu-44x.h
index bf52d70..ca1b90c 100644
--- a/arch/powerpc/include/asm/mmu-44x.h
+++ b/arch/powerpc/include/asm/mmu-44x.h
@@ -8,6 +8,7 @@

#define PPC44x_MMUCR_TID 0x000000ff
#define PPC44x_MMUCR_STS 0x00010000
+#define PPC44x_MMUCR_U2 0x00200000

#define PPC44x_TLB_PAGEID 0
#define PPC44x_TLB_XLAT 1
@@ -32,6 +33,7 @@

/* Storage attribute and access control fields */
#define PPC44x_TLB_ATTR_MASK 0x0000ff80
+#define PPC44x_TLB_WL1 0x00100000 /* Write-through L1 */
#define PPC44x_TLB_U0 0x00008000 /* User 0 */
#define PPC44x_TLB_U1 0x00004000 /* User 1 */
#define PPC44x_TLB_U2 0x00002000 /* User 2 */
diff --git a/arch/powerpc/kernel/head_44x.S b/arch/powerpc/kernel/head_44x.S
index 5e12b74..9a9a4ee 100644
--- a/arch/powerpc/kernel/head_44x.S
+++ b/arch/powerpc/kernel/head_44x.S
@@ -429,7 +429,17 @@ finish_tlb_load_44x:
andi. r10,r12,_PAGE_USER /* User page ? */
beq 1f /* nope, leave U bits empty */
rlwimi r11,r11,3,26,28 /* yes, copy S bits to U */
-1: tlbwe r11,r13,PPC44x_TLB_ATTRIB /* Write ATTRIB */
+1:
+BEGIN_MMU_FTR_SECTION
+ andi. r10, r11, PPC44x_TLB_I
+ bne 2f
+ oris r11,r11,PPC44x_TLB_WL1@h /* Add coherency for */
+ /* non-inhibited */
+ ori r11,r11,PPC44x_TLB_U2|PPC44x_TLB_M
+END_MMU_FTR_SECTION_IFSET(MMU_FTR_NEED_L1_WRITETHROUGH)
+
+2:
+ tlbwe r11,r13,PPC44x_TLB_ATTRIB /* Write ATTRIB */

/* Done...restore registers and get out of here.
*/
@@ -799,7 +809,12 @@ skpinv: addi r4,r4,1 /* Increment */
sync

/* Initialize MMUCR */
+BEGIN_MMU_FTR_SECTION
+ lis r5, PPC44x_MMUCR_U2@h
+END_MMU_FTR_SECTION_IFSET(MMU_FTR_NEED_L1_WRITETHROUGH)
+BEGIN_MMU_FTR_SECTION
li r5,0
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
mtspr SPRN_MMUCR,r5
sync

@@ -814,7 +829,15 @@ skpinv: addi r4,r4,1 /* Increment */
/* attrib fields */
/* Added guarded bit to protect against speculative loads/stores */
li r5,0
- ori r5,r5,(PPC44x_TLB_SW | PPC44x_TLB_SR | PPC44x_TLB_SX | PPC44x_TLB_G)
+BEGIN_MMU_FTR_SECTION
+ ori r5,r5,(PPC44x_TLB_SW | PPC44x_TLB_SR | PPC44x_TLB_SX | \
+ PPC44x_TLB_G | PPC44x_TLB_U2)
+ oris r5,r5,PPC44x_TLB_WL1@h
+END_MMU_FTR_SECTION_IFSET(MMU_FTR_NEED_L1_WRITETHROUGH)
+BEGIN_MMU_FTR_SECTION
+ ori r5,r5,(PPC44x_TLB_SW | PPC44x_TLB_SR | PPC44x_TLB_SX | \
+ PPC44x_TLB_G)
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)

li r0,63 /* TLB slot 63 */

diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index 998a100..b54e2e8 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -506,7 +506,27 @@ _GLOBAL(clear_pages)
li r0,PAGE_SIZE/L1_CACHE_BYTES
slw r0,r0,r4
mtctr r0
-1: dcbz 0,r3
+ li r4, 0
+1:
+BEGIN_MMU_FTR_SECTION
+ /* assuming 32 byte cacheline */
+ stw r4, 0(r3)
+ stw r4, 4(r3)
+ stw r4, 8(r3)
+ stw r4, 12(r3)
+ stw r4, 16(r3)
+ stw r4, 20(r3)
+ stw r4, 24(r3)
+ stw r4, 28(r3)
+END_MMU_FTR_SECTION_IFSET(MMU_FTR_NEED_L1_WRITETHROUGH)
+/*
+ * would have used an ELSE_MMU_FTR_SECTION here but it
+ * broke the code with Error: non-constant expression in ".if" statement
+ *
+ */
+BEGIN_MMU_FTR_SECTION
+ dcbz 0,r3
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
addi r3,r3,L1_CACHE_BYTES
bdnz 1b
blr
@@ -550,7 +570,9 @@ _GLOBAL(copy_page)
mtctr r0
1:
dcbt r11,r4
+BEGIN_MMU_FTR_SECTION
dcbz r5,r3
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
COPY_16_BYTES
#if L1_CACHE_BYTES >= 32
COPY_16_BYTES
diff --git a/arch/powerpc/lib/copy_32.S b/arch/powerpc/lib/copy_32.S
index 55f19f9..2646838 100644
--- a/arch/powerpc/lib/copy_32.S
+++ b/arch/powerpc/lib/copy_32.S
@@ -12,6 +12,7 @@
#include <asm/cache.h>
#include <asm/errno.h>
#include <asm/ppc_asm.h>
+#include <asm/mmu.h>

#define COPY_16_BYTES \
lwz r7,4(r4); \
@@ -98,7 +99,10 @@ _GLOBAL(cacheable_memzero)
bdnz 4b
3: mtctr r9
li r7,4
-10: dcbz r7,r6
+10:
+BEGIN_MMU_FTR_SECTION
+ dcbz r7,r6
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
addi r6,r6,CACHELINE_BYTES
bdnz 10b
clrlwi r5,r8,32-LG_CACHELINE_BYTES
@@ -187,7 +191,9 @@ _GLOBAL(cacheable_memcpy)
mtctr r0
beq 63f
53:
+BEGIN_MMU_FTR_SECTION
dcbz r11,r6
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
COPY_16_BYTES
#if L1_CACHE_BYTES >= 32
COPY_16_BYTES
@@ -368,7 +374,10 @@ _GLOBAL(__copy_tofrom_user)
mtctr r8

53: dcbt r3,r4
-54: dcbz r11,r6
+54:
+BEGIN_MMU_FTR_SECTION
+ dcbz r11,r6
+END_MMU_FTR_SECTION_IFCLR(MMU_FTR_NEED_L1_WRITETHROUGH)
.section __ex_table,"a"
.align 2
.long 54b,105f
diff --git a/arch/powerpc/mm/44x_mmu.c b/arch/powerpc/mm/44x_mmu.c
index 024acab..f5c60b3 100644
--- a/arch/powerpc/mm/44x_mmu.c
+++ b/arch/powerpc/mm/44x_mmu.c
@@ -80,9 +80,12 @@ static void __init ppc44x_pin_tlb(unsigned int virt, unsigned int phys)
:
#ifdef CONFIG_PPC47x
: "r" (PPC47x_TLB2_S_RWX),
-#else
+#elseif CONFIG_BGP_L1_WRITETHROUGH
+ : "r" (PPC44x_TLB_SW | PPC44x_TLB_SR | PPC44x_TLB_SX | PPC44x_TLB_WL1 \
+ | PPC44x_TLB_U2 | PPC44x_TLB_M),
+#else /* neither CONFIG_PPC47x or CONFIG_BGP_L1_WRITETHROUGH */
: "r" (PPC44x_TLB_SW | PPC44x_TLB_SR | PPC44x_TLB_SX | PPC44x_TLB_G),
-#endif
+#endif /* CONFIG_PPC47x */
"r" (phys),
"r" (virt | PPC44x_TLB_VALID | PPC44x_TLB_256M),
"r" (entry),
--
1.7.4.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Benjamin Herrenschmidt: "Re: [PATCHES v3] pci/of: Generic PCI <-> OF matching"
Previous message: Luck, Tony: "RE: [PATCH v2 3/3] efi: Add support for using efivars as a pstorebackend"
Next in thread: Benjamin Herrenschmidt: "Re: [PATCH] [RFC][V3] bluegene: use MMU feature flag toconditionalize L1 writethrough code"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]