openwrt/target/linux/generic/hack-4.19/300-MIPS-r4k_cache-use-more-efficient-cache-blast.patch
Stijn Tintel a37098a2d0 kernel: bump 4.19 to 4.19.16
Refresh patches.
Remove upstreamed patches:
- backport/096-mips-math-emu-Write-protect-delay-slot-emulation-pages.patch
- backport/096-v4.20-netfilter-ipv6-Preserve-link-scope-traffic-original-.patch
- backport/424-v4.20-net-dsa-fix-88e6060-roaming.patch
- hack/100-mtd-rawnand-qcom-fix-memory-corruption-that-causes-p.patch
- pending/510-f2fs-fix-sanity_check_raw_super-on-big-endian-machines.patch
Update patch that no longer applies:
- backport/343-netfilter-nft_flow_offload-handle-netdevice-events-f.patch

Compile-tested: mesongx
Runtime-tested: mesongx

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
2019-01-17 03:04:13 +02:00

67 lines
2.2 KiB
Diff

From: Ben Menchaca <ben.menchaca@qca.qualcomm.com>
Date: Fri, 7 Jun 2013 18:35:22 -0500
Subject: MIPS: r4k_cache: use more efficient cache blast
Optimize the compiler output for larger cache blast cases that are
common for DMA-based networking.
Signed-off-by: Ben Menchaca <ben.menchaca@qca.qualcomm.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
---
--- a/arch/mips/include/asm/r4kcache.h
+++ b/arch/mips/include/asm/r4kcache.h
@@ -683,16 +683,48 @@ static inline void prot##extra##blast_##
unsigned long end) \
{ \
unsigned long lsize = cpu_##desc##_line_size(); \
+ unsigned long lsize_2 = lsize * 2; \
+ unsigned long lsize_3 = lsize * 3; \
+ unsigned long lsize_4 = lsize * 4; \
+ unsigned long lsize_5 = lsize * 5; \
+ unsigned long lsize_6 = lsize * 6; \
+ unsigned long lsize_7 = lsize * 7; \
+ unsigned long lsize_8 = lsize * 8; \
unsigned long addr = start & ~(lsize - 1); \
- unsigned long aend = (end - 1) & ~(lsize - 1); \
+ unsigned long aend = (end + lsize - 1) & ~(lsize - 1); \
+ int lines = (aend - addr) / lsize; \
\
__##pfx##flush_prologue \
\
- while (1) { \
+ while (lines >= 8) { \
+ prot##cache_op(hitop, addr); \
+ prot##cache_op(hitop, addr + lsize); \
+ prot##cache_op(hitop, addr + lsize_2); \
+ prot##cache_op(hitop, addr + lsize_3); \
+ prot##cache_op(hitop, addr + lsize_4); \
+ prot##cache_op(hitop, addr + lsize_5); \
+ prot##cache_op(hitop, addr + lsize_6); \
+ prot##cache_op(hitop, addr + lsize_7); \
+ addr += lsize_8; \
+ lines -= 8; \
+ } \
+ \
+ if (lines & 0x4) { \
+ prot##cache_op(hitop, addr); \
+ prot##cache_op(hitop, addr + lsize); \
+ prot##cache_op(hitop, addr + lsize_2); \
+ prot##cache_op(hitop, addr + lsize_3); \
+ addr += lsize_4; \
+ } \
+ \
+ if (lines & 0x2) { \
+ prot##cache_op(hitop, addr); \
+ prot##cache_op(hitop, addr + lsize); \
+ addr += lsize_2; \
+ } \
+ \
+ if (lines & 0x1) { \
prot##cache_op(hitop, addr); \
- if (addr == aend) \
- break; \
- addr += lsize; \
} \
\
__##pfx##flush_epilogue \