-
Huacai Chen authored
This and the next patch resolve memory corruption problems while CPU hotplug. Without these patches, memory corruption can triggered easily as below: On a quad-core MIPS platform, use "spawn" of UnixBench-5.1.3 (http:// code.google.com/p/byte-unixbench/) and a CPU hotplug script like this (hotplug.sh): while true; do echo 0 >/sys/devices/system/cpu/cpu1/online echo 0 >/sys/devices/system/cpu/cpu2/online echo 0 >/sys/devices/system/cpu/cpu3/online sleep 1 echo 1 >/sys/devices/system/cpu/cpu1/online echo 1 >/sys/devices/system/cpu/cpu2/online echo 1 >/sys/devices/system/cpu/cpu3/online sleep 1 done Run "hotplug.sh" and then run "spawn 10000", spawn will get segfault after a few minutes. This patch: Currently, clear_page()/copy_page() are generated by Micro-assembler dynamically. But they are unavailable until uasm_resolve_relocs() has finished because jump labels are illegal before that. Since these functions are shared by every CPU, we only call build_clear_page()/ build_copy_page() only once at boot time. Without this patch, programs will get random memory corruption (segmentation fault, bus error, etc.) while CPU Hotplug (e.g. one CPU is using clear_page() while another is generating it in cpu_cache_init()). For similar reasons we modify build_tlb_refill_handler()'s invocation. V2: 1, Rework the code to make CPU#0 can be online/offline. 2, Introduce cpu_has_local_ebase feature since some types of MIPS CPU need a per-CPU tlb_refill_handler(). Signed-off-by:
Huacai Chen <chenhc@lemote.com> Signed-off-by:
Hongbing Hu <huhb@lemote.com> Acked-by:
David Daney <david.daney@cavium.com> Patchwork: http://patchwork.linux-mips.org/patch/4994/Acked-by:
John Crispin <blogic@openwrt.org>
8759934e