diff options
| author | Lichao Liu <liulichao@loongson.cn> | 2020-06-11 17:59:24 +0800 |
|---|---|---|
| committer | Thomas Bogendoerfer <tsbogend@alpha.franken.de> | 2020-06-22 17:35:31 +0200 |
| commit | 2984b3f8c24b87ba4d57e213d593b2f91a4a31dd (patch) | |
| tree | 45cc1b4362ac086fbfb4effe11fb999ebb7f5931 /arch/mips/loongson2ef | |
| parent | 9909bc43a2e218a5a71a6245c36d86e4f587c74b (diff) | |
| download | lwn-2984b3f8c24b87ba4d57e213d593b2f91a4a31dd.tar.gz lwn-2984b3f8c24b87ba4d57e213d593b2f91a4a31dd.zip | |
MIPS: Loongson-2EF: disable fix-loongson3-llsc in compiler
Firstly, Loongson-2EF support ll/sc instructions, but
doesn't need fix-loongson3-llsc compile option.
Secondly, fix-loongson3-llsc will cause kernel startup
fail at futex_init, because compiler will add 'sync' before
'll', which will affect __ex_table.
futex_init will pass NULL uaddr parameter to
futex_atomic_cmpxchg_inatomic.
futex_atomic_cmpxchg_inatomic will access uaddr directly,
which will cause page fault exception, the exception should be
handled by __ex_table's nextinsn if the exception insn exsit in
__ex_table. Because __ex_table is affected by compiler,
the exception can not be handled, and
futex_atomic_cmpxchg_inatomic will crash.
Error code as below:
__ex_table.insn = 1b, which is 'sync' compiled with
fix-loongson3-llsc, but the actual exception instrction is ll.
So, do_page_fault will not find the correct inst in __ex_table, and
can not handle this exception.
"1: "user_ll("%1", "%3")" \n"
" bne %1, %z4, 3f \n"
" .set pop \n"
" move $1, %z5 \n"
" .set "MIPS_ISA_ARCH_LEVEL" \n"
"2: "user_sc("$1", "%2")" \n"
" beqz $1, 1b \n"
"3: " __SYNC_ELSE(full, loongson3_war, __WEAK_LLSC_MB) "\n"
" .insn \n"
" .set pop \n"
" .section .fixup,\"ax\" \n"
"4: li %0, %6 \n"
" j 3b \n"
" .previous \n"
" .section __ex_table,\"a\" \n"
" "__UA_ADDR "\t1b, 4b \n"
" "__UA_ADDR "\t2b, 4b \n"
" .previous
Signed-off-by: Lichao Liu <liulichao@loongson.cn>
Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Diffstat (limited to 'arch/mips/loongson2ef')
| -rw-r--r-- | arch/mips/loongson2ef/Platform | 22 |
1 files changed, 22 insertions, 0 deletions
diff --git a/arch/mips/loongson2ef/Platform b/arch/mips/loongson2ef/Platform index cdad3c1a9a18..4ab55f1123a0 100644 --- a/arch/mips/loongson2ef/Platform +++ b/arch/mips/loongson2ef/Platform @@ -8,6 +8,28 @@ cflags-$(CONFIG_CPU_LOONGSON2E) += \ $(call cc-option,-march=loongson2e,-march=r4600) cflags-$(CONFIG_CPU_LOONGSON2F) += \ $(call cc-option,-march=loongson2f,-march=r4600) +# +# Some versions of binutils, not currently mainline as of 2019/02/04, support +# an -mfix-loongson3-llsc flag which emits a sync prior to each ll instruction +# to work around a CPU bug (see __SYNC_loongson3_war in asm/sync.h for a +# description). +# +# We disable this in order to prevent the assembler meddling with the +# instruction that labels refer to, ie. if we label an ll instruction: +# +# 1: ll v0, 0(a0) +# +# ...then with the assembler fix applied the label may actually point at a sync +# instruction inserted by the assembler, and if we were using the label in an +# exception table the table would no longer contain the address of the ll +# instruction. +# +# Avoid this by explicitly disabling that assembler behaviour. If upstream +# binutils does not merge support for the flag then we can revisit & remove +# this later - for now it ensures vendor toolchains don't cause problems. +# +cflags-$(CONFIG_CPU_LOONGSON2EF) += $(call as-option,-Wa$(comma)-mno-fix-loongson3-llsc,) + # Enable the workarounds for Loongson2f ifdef CONFIG_CPU_LOONGSON2F_WORKAROUNDS ifeq ($(call as-option,-Wa$(comma)-mfix-loongson2f-nop,),) |