[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <10231b81-ea42-26d0-4c11-92851229e658@iscas.ac.cn>
Date: Thu, 27 Jul 2023 03:34:16 +0800
From: Mingzheng Xing <xingmingzheng@...as.ac.cn>
To: Conor Dooley <conor@...nel.org>
Cc: Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>,
Nathan Chancellor <nathan@...nel.org>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Tom Rix <trix@...hat.com>, Bin Meng <bmeng@...ylab.org>,
linux-riscv@...ts.infradead.org, linux-kernel@...r.kernel.org,
llvm@...ts.linux.dev, stable@...r.kernel.org
Subject: Re: [PATCH v2] riscv: Handle zicsr/zifencei issue between gcc and
binutils
On 7/27/23 02:02, Conor Dooley wrote:
> On Thu, Jul 27, 2023 at 01:45:24AM +0800, Mingzheng Xing wrote:
>> Binutils-2.38 and GCC-12.1.0 bump[0] default ISA spec to newer version
>> 20191213 which moves some instructions from the I extension to the
>> Zicsr and Zifencei extensions. So if one of the binutils and GCC exceeds
>> that version, we should turn on TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI
>> to cope with the new changes.
>>
>> The case of clang is special[1][2], where older clang versions (<17) need
>> to be rolled back to old ISA spec to fix it. And the less common case,
>> since older GCC versions (<11.1.0) did not support zicsr and zifencei
> Can you provide a link to explain why this is 11.1.0 in particular?
Okay, I can add it in commit message. gcc-11.1.0 is particular
because it add support zicsr and zifencei extension for -march[1].
Link:
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=b03be74bad08c382da47e048007a78fa3fb4ef49
[1]
>> extension for -march, also requires a fallback to cope with it.
>>
>> For more information, please refer to:
>> commit 6df2a016c0c8 ("riscv: fix build with binutils 2.38")
>> commit e89c2e815e76 ("riscv: Handle zicsr/zifencei issues between clang and binutils")
>> Link: https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=98416dbb0a62579d4a7a4a76bab51b5b52fec2cd [0]
>> Link: https://lore.kernel.org/all/20230308220842.1231003-1-conor@kernel.org [1]
>> Link: https://lore.kernel.org/all/20230223220546.52879-1-conor@kernel.org [2]
>> Link: https://lore.kernel.org/all/20230725170405.251011-1-xingmingzheng@iscas.ac.cn
> This shouldn't be here, you don't link to your old patches.
My bad, I will fix it.
>
>> Signed-off-by: Mingzheng Xing <xingmingzheng@...as.ac.cn>
> This is still broken for:
> CONFIG_CLANG_VERSION=0
> CONFIG_AS_IS_GNU=y
> CONFIG_AS_VERSION=23500
> CONFIG_LD_IS_BFD=y
> CONFIG_LD_VERSION=23500
Do you mean that these CONFIG_* will cause kernel
compilation errors when paired with certain versions of GCC?
Or perhaps I misunderstood your meaning.
>
> Please don't post a v2 while there is still ongoing discussion on the
> v1. I'll try to reply here tomorrow with a diff you can fold in to fix
> the problem.
Okay, thanks for your review.
> Thanks,
> Conor.
>
>> ---
>>
>> v2:
>> - Update the Kconfig help text and commit message.
>> - Add considerations for low version gcc case.
>>
>> Sorry for the formatting error on my mailing list reply.
>>
>> arch/riscv/Kconfig | 23 +++++++++++++----------
>> 1 file changed, 13 insertions(+), 10 deletions(-)
>>
>> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> index 4c07b9189c86..08afd47de157 100644
>> --- a/arch/riscv/Kconfig
>> +++ b/arch/riscv/Kconfig
>> @@ -570,24 +570,27 @@ config TOOLCHAIN_HAS_ZIHINTPAUSE
>> config TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI
>> def_bool y
>> # https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=aed44286efa8ae8717a77d94b51ac3614e2ca6dc
>> - depends on AS_IS_GNU && AS_VERSION >= 23800
>> + # https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=98416dbb0a62579d4a7a4a76bab51b5b52fec2cd
>> + depends on GCC_VERSION >= 120100 || (AS_IS_GNU && AS_VERSION >= 23800)
>> help
>> - Newer binutils versions default to ISA spec version 20191213 which
>> - moves some instructions from the I extension to the Zicsr and Zifencei
>> - extensions.
>> + Binutils-2.38 and GCC-12.1.0 bump default ISA spec to newer version
>> + 20191213 which moves some instructions from the I extension to the
>> + Zicsr and Zifencei extensions.
>>
>> config TOOLCHAIN_NEEDS_OLD_ISA_SPEC
>> def_bool y
>> depends on TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI
>> # https://github.com/llvm/llvm-project/commit/22e199e6afb1263c943c0c0d4498694e15bf8a16
>> - depends on CC_IS_CLANG && CLANG_VERSION < 170000
>> + # https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=b03be74bad08c382da47e048007a78fa3fb4ef49
>> + depends on (CC_IS_CLANG && CLANG_VERSION < 170000) || \
>> + (CC_IS_GCC && GCC_VERSION < 110100)
>> help
>> - Certain versions of clang do not support zicsr and zifencei via -march
>> - but newer versions of binutils require it for the reasons noted in the
>> - help text of CONFIG_TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI. This
>> + Certain versions of clang (or GCC) do not support zicsr and zifencei via
>> + -march but newer versions of binutils require it for the reasons noted
>> + in the help text of CONFIG_TOOLCHAIN_NEEDS_EXPLICIT_ZICSR_ZIFENCEI. This
>> option causes an older ISA spec compatible with these older versions
>> - of clang to be passed to GAS, which has the same result as passing zicsr
>> - and zifencei to -march.
>> + of clang (or GCC) to be passed to GAS, which has the same result as
>> + passing zicsr and zifencei to -march.
>>
>> config FPU
>> bool "FPU support"
>> --
>> 2.34.1
>>
>>
>> _______________________________________________
>> linux-riscv mailing list
>> linux-riscv@...ts.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-riscv
Powered by blists - more mailing lists