Hi,
The following patch implements support for instantiation of 8MB D-TLB
entries for the kernel direct virtual mapping on 8xx, thus reducing TLB
space consumed for the kernel.
Test used: writing 40MB from /dev/zero to file in ext2fs over
RAMDISK.
$ time dd if=/dev/zero of=file bs=4k count=10000
VANILLA 8MB kernel data pages
real 0m11.485s real 0m11.267s
user 0m0.218s user 0m0.250s
sys 0m8.939s sys 0m9.108s
real 0m11.518s real 0m10.978s
user 0m0.203s user 0m0.222s
sys 0m9.585s sys 0m9.138s
real 0m11.554s real 0m10.967s
user 0m0.228s user 0m0.222s
sys 0m9.497s sys 0m9.127s
real 0m11.633s real 0m11.286s
user 0m0.214s user 0m0.196s
sys 0m9.529s sys 0m9.134s
and averages for both:
real 11.54750 real 11.12450
Which is a 3.6% improvement in execution time. More improvement is
expected for loads with larger kernel data footprint (real workloads).
Dan, could you please review the code.
diff --git a/arch/ppc/kernel/head_8xx.S b/arch/ppc/kernel/head_8xx.S
index de09787..fe25f3f 100644
--- a/arch/ppc/kernel/head_8xx.S
+++ b/arch/ppc/kernel/head_8xx.S
@@ -375,6 +375,14 @@ DataStoreTLBMiss:
lis r11, swapper_pg_dir at h
ori r11, r11, swapper_pg_dir at l
rlwimi r10, r11, 0, 2, 19
+ stw r12, 16(r0)
+ mflr r12
+ stw r12, 20(r0) /* save LR */
+ lis r3, LoadLargeDTLB at h
+ ori r3, r3, LoadLargeDTLB at l
+ tophys (r3, r3)
+ mtlr r3
+ blr
3:
lwz r11, 0(r10) /* Get the level 1 entry */
rlwinm. r10, r11,0,0,19 /* Extract page descriptor page address */
@@ -430,6 +438,83 @@ DataStoreTLBMiss:
InstructionTLBError:
b InstructionAccess
+LoadLargeDTLB:
+ li r12, 0
+ lwz r11, 0(r10) /* Get the level 1 entry */
+ rlwinm. r10, r11,0,0,19 /* Extract page descriptor page address */
+ beq 3f /* If zero, don't try to find a pte */
+
+ /* We have a pte table, so load fetch the pte from the table.
+ */
+ ori r11, r11, 1 /* Set valid bit in physical L2 page */
+ DO_8xx_CPU6(0x3b80, r3)
+ mtspr SPRN_MD_TWC, r11 /* Load pte table base address */
+ mfspr r10, SPRN_MD_TWC /* ....and get the pte address */
+ lwz r10, 0(r10) /* Get the pte */
+
+ /* Insert the Guarded flag into the TWC from the Linux PTE.
+ * It is bit 27 of both the Linux PTE and the TWC (at least
+ * I got that right :-). It will be better when we can put
+ * this into the Linux pgd/pmd and load it in the operation
+ * above.
+ */
+ rlwimi r11, r10, 0, 27, 27
+
+ rlwimi r12, r10, 0, 0, 9 /* extract phys. addr */
+ mfspr r3, SPRN_MD_EPN
+ rlwinm r3, r3, 0, 0, 9 /* extract virtual address */
+ tophys(r3, r3)
+ cmpw r3, r12 /* only use 8M page if it is a direct
+ kernel mapping */
+ bne 1f
+ ori r11, r11, MD_PS8MEG
+ li r12, 1
+ b 2f
+1:
+ li r12, 0 /* can't use 8MB TLB, so zero r12. */
+2:
+ DO_8xx_CPU6(0x3b80, r3)
+ mtspr SPRN_MD_TWC, r11
+
+ /* The Linux PTE won't go exactly into the MMU TLB.
+ * Software indicator bits 21, 22 and 28 must be clear.
+ * Software indicator bits 24, 25, 26, and 27 must be
+ * set. All other Linux PTE bits control the behavior
+ * of the MMU.
+ */
+3: li r11, 0x00f0
+ rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */
+ cmpwi r12, 1
+ bne 4f
+ ori r10, r10, 0x8
+
+ mfspr r12, SPRN_MD_EPN
+ lis r3, 0xff80 /* 10-19 must be clear for 8MB TLB */
+ ori r3, r3, 0x0fff
+ and r12, r3, r12
+ DO_8xx_CPU6(0x3780, r3)
+ mtspr SPRN_MD_EPN, r12
+
+ lis r3, 0xff80 /* 10-19 must be clear for 8MB TLB */
+ ori r3, r3, 0x0fff
+ and r10, r3, r10
+4:
+ DO_8xx_CPU6(0x3d80, r3)
+ mtspr SPRN_MD_RPN, r10 /* Update TLB entry */
+
+ mfspr r10, SPRN_M_TW /* Restore registers */
+ lwz r11, 0(r0)
+ mtcr r11
+ lwz r11, 4(r0)
+
+ lwz r12, 20(r0)
+ mtlr r12 /* Restore LR */
+ lwz r12, 16(r0)
+#ifdef CONFIG_8xx_CPU6
+ lwz r3, 8(r0)
+#endif
+ rfi
+
/* This is the data TLB error on the MPC8xx. This could be due to
* many reasons, including a dirty update to a pte. We can catch that
* one here, but anything else is an error. First, we track down the
