From: jiawei@iscas.ac.cn
To: "Nelson Chu" <nelson@rivosinc.com>
Cc: binutils@sourceware.org, kito.cheng@sifive.com,
palmer@dabbelt.com, jbeulich@suse.com,
christoph.muellner@vrull.eu, jeremy.bennett@embecosm.com,
mary.bennett@embecosm.com, nandni.jamnadas@embecosm.com,
charlie.keaney@embecosm.com, simon.cook@embecosm.com,
tariq.kurd@codasip.com, gaofei@eswincomputing.com,
sinan.lin@linux.alibaba.com, wuwei2016@iscas.ac.cn,
shihua@iscas.ac.cn, shiyulong@iscas.ac.cn,
chenyixuan@iscas.ac.cn
Subject: Re: Re: [PATCH v2] RISC-V: Support Zca/f/d extensions.
Date: Thu, 13 Jul 2023 15:26:56 +0800 (GMT+08:00) [thread overview]
Message-ID: <72e31038.98e2.1894e25529c.Coremail.jiawei@iscas.ac.cn> (raw)
In-Reply-To: <CAPpQWtDhHm+r_453j_q33ECztUVLWAg0aWOJVykjOTQj5VtfwA@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 15742 bytes --]
>Thanks, just some minor issues as follows, ...
>
>On Tue, Jul 11, 2023 at 4:33 PM Jiawei <jiawei@iscas.ac.cn> wrote:
>
>> This patch add Zca/f/d extensions support, since all ZC*
>> extensions will imply Zca extension, just enabled compress
>> feature when Zca extension is available.
>>
>> V2: Merged patches and refined testcases as Nelson's suggestions.
>>
>> Co-Authored by: Charlie Keaney <charlie.keaney@embecosm.com>
>> Co-Authored by: Mary Bennett <mary.bennett@embecosm.com>
>> Co-Authored by: Nandni Jamnadas <nandni.jamnadas@embecosm.com>
>> Co-Authored by: Sinan Lin <sinan.lin@linux.alibaba.com>
>> Co-Authored by: Simon Cook <simon.cook@embecosm.com>
>> Co-Authored by: Shihua Liao <shihua@iscas.ac.cn>
>> Co-Authored by: Yulong Shi <yulong@iscas.ac.cn>
>>
>> bfd/ChangeLog:
>>
>> * elfxx-riscv.c (riscv_multi_subset_supports): New extensions.
>> (riscv_multi_subset_supports_ext): Ditto.
>>
>> gas/ChangeLog:
>>
>> * config/tc-riscv.c (riscv_set_arch): Extend compress check.
>> * testsuite/gas/riscv/zca.d: New test.
>> * testsuite/gas/riscv/zca.s: New test.
>> * testsuite/gas/riscv/zcd.d: New test.
>> * testsuite/gas/riscv/zcd.s: New test.
>> * testsuite/gas/riscv/zcf.d: New test.
>> * testsuite/gas/riscv/zcf.s: New test.
>>
>> ---
>> bfd/elfxx-riscv.c | 28 ++++++++++++------
>> gas/config/tc-riscv.c | 3 +-
>> gas/testsuite/gas/riscv/zca.d | 54 +++++++++++++++++++++++++++++++++++
>> gas/testsuite/gas/riscv/zca.s | 47 ++++++++++++++++++++++++++++++
>> gas/testsuite/gas/riscv/zcd.d | 16 +++++++++++
>> gas/testsuite/gas/riscv/zcd.s | 10 +++++++
>> gas/testsuite/gas/riscv/zcf.d | 16 +++++++++++
>> gas/testsuite/gas/riscv/zcf.s | 10 +++++++
>> 8 files changed, 174 insertions(+), 10 deletions(-)
>> create mode 100644 gas/testsuite/gas/riscv/zca.d
>> create mode 100644 gas/testsuite/gas/riscv/zca.s
>> create mode 100644 gas/testsuite/gas/riscv/zcd.d
>> create mode 100644 gas/testsuite/gas/riscv/zcd.s
>> create mode 100644 gas/testsuite/gas/riscv/zcf.d
>> create mode 100644 gas/testsuite/gas/riscv/zcf.s
>>
>> diff --git a/bfd/elfxx-riscv.c b/bfd/elfxx-riscv.c
>> index bdfc0ef01f4..25864a8cf89 100644
>> --- a/bfd/elfxx-riscv.c
>> +++ b/bfd/elfxx-riscv.c
>> @@ -1171,6 +1171,8 @@ static struct riscv_implicit_subset
>> riscv_implicit_subsets[] =
>> {"zvksg", "zvkg", check_implicit_always},
>> {"zvksc", "zvks", check_implicit_always},
>> {"zvksc", "zvbc", check_implicit_always},
>> + {"zcf", "zca", check_implicit_always},
>> + {"zcd", "zca", check_implicit_always},
>
>
>1. Should we also imply f for zcf and d for zcd?
I think it's really a good idea to add this imply relationship.
It's very convenient for users,I will add it in next version.
>2. Is c totally the same as zca for now? Or c includes all rvc
>instructions, like zca, zcf, zcd and zcb, ... It seems the latter makes
>more sense.
You are right,it means the latter one. However, all zc* extension will
defaultly implying zca. It changes the relationship, zca is the public part
when use other zc* extensions.
Here is an example, when we want to zcb extension, it should generate compress
code and set the align with 16 bits, what we really use is (zca + zcb), we can
choice zca or zcb to enable the compress feature as c extension did. If case change
to zcmp(zcmp = zca + zcmp), we choice form zca or zcmp again. So I just set zca
to do this work as c extension did.
>
>
>> {"smaia", "ssaia", check_implicit_always},
>> {"smstateen", "ssstateen", check_implicit_always},
>> {"smepmp", "zicsr", check_implicit_always},
>> @@ -1304,6 +1306,9 @@ static struct riscv_supported_ext
>> riscv_supported_std_z_ext[] =
>> {"zvl32768b", ISA_SPEC_CLASS_DRAFT, 1, 0, 0 },
>> {"zvl65536b", ISA_SPEC_CLASS_DRAFT, 1, 0, 0 },
>> {"ztso", ISA_SPEC_CLASS_DRAFT, 0, 1, 0 },
>> + {"zca", ISA_SPEC_CLASS_DRAFT, 1, 0, 0 },
>> + {"zcf", ISA_SPEC_CLASS_DRAFT, 1, 0, 0 },
>> + {"zcd", ISA_SPEC_CLASS_DRAFT, 1, 0, 0 },
>> {NULL, 0, 0, 0, 0}
>> };
>>
>> @@ -2384,13 +2389,16 @@ riscv_multi_subset_supports (riscv_parse_subset_t
>> *rps,
>> case INSN_CLASS_Q:
>> return riscv_subset_supports (rps, "q");
>> case INSN_CLASS_C:
>>
>
>Renamed to INSN_CLASS_ZCA?
I'm a little concerned if this change will cause any compatibility issues, if not,
it's a good choice.
>
>
>> - return riscv_subset_supports (rps, "c");
>> + return riscv_subset_supports (rps, "c")
>> + || riscv_subset_supports (rps, "zca");
>> case INSN_CLASS_F_AND_C:
>>
>
>Renamed to INSN_CLASS_F_AND_ZCF or INSN_CLASS_ZCF when implying f?
>
>
>> return (riscv_subset_supports (rps, "f")
>> - && riscv_subset_supports (rps, "c"));
>> + && (riscv_subset_supports (rps, "c")
>> + || riscv_subset_supports (rps, "zcf")));
>
> case INSN_CLASS_D_AND_C:
>>
>
>Renamed to INSN_CLASS_F_AND_ZCD or INSN_CLASS_ZCD when implying d?
>
>
>> return (riscv_subset_supports (rps, "d")
>> - && riscv_subset_supports (rps, "c"));
>> + && (riscv_subset_supports (rps, "c")
>> + || riscv_subset_supports (rps, "zcd")));
>> case INSN_CLASS_F_INX:
>> return (riscv_subset_supports (rps, "f")
>> || riscv_subset_supports (rps, "zfinx"));
>> @@ -2566,20 +2574,22 @@ riscv_multi_subset_supports_ext
>> (riscv_parse_subset_t *rps,
>> return "c";
>> case INSN_CLASS_F_AND_C:
>> if (!riscv_subset_supports (rps, "f")
>> - && !riscv_subset_supports (rps, "c"))
>> - return _("f' and `c");
>> + && !(riscv_subset_supports (rps, "c")
>> + || riscv_subset_supports (rps, "zcf")))
>> + return _("f' and `c', or `f' and `zcf");
>> else if (!riscv_subset_supports (rps, "f"))
>> return "f";
>> else
>> - return "c";
>> + return _("c' or `zcf");
>
>
>case INSN_CLASS_F_AND_ZCF:
> if (!riscv_subset_supports (rps, "f")
> {
> if (!riscv_subset_supports (rps, "c") && !riscv_subset_supports (rps,
>"zcf"))
> return _("f' and `c', or `f' and `zcf");
> else
> return "f";
> }
> else
> return _("c' or `zcf");
>
>Or if we imply f for zcf, then just,
>
>case INSN_CLASS_ZCF:
> return _("c' or `zcf");
>
Thanks for your demo.
>
>> case INSN_CLASS_D_AND_C:
>> if (!riscv_subset_supports (rps, "d")
>> - && !riscv_subset_supports (rps, "c"))
>> - return _("d' and `c");
>> + && !(riscv_subset_supports (rps, "c")
>> + || riscv_subset_supports (rps, "zcd")))
>> + return _("d' and `c', or `d' and `zcd");
>> else if (!riscv_subset_supports (rps, "d"))
>> return "d";
>> else
>> - return "c";
>> + return _("c' or `zcd");
>>
>
>Likewise.
>
>
>> case INSN_CLASS_F_INX:
>> return _("f' or `zfinx");
>> case INSN_CLASS_D_INX:
>> diff --git a/gas/config/tc-riscv.c b/gas/config/tc-riscv.c
>> index 297bb9b2a81..2eb0d1877ee 100644
>> --- a/gas/config/tc-riscv.c
>> +++ b/gas/config/tc-riscv.c
>> @@ -337,7 +337,8 @@ riscv_set_arch (const char *s)
>> riscv_reset_subsets_list_arch_str ();
>>
>> riscv_set_rvc (false);
>> - if (riscv_subset_supports (&riscv_rps_as, "c"))
>> + if (riscv_subset_supports (&riscv_rps_as, "c")
>> + || riscv_subset_supports (&riscv_rps_as, "zca"))
>> riscv_set_rvc (true);
>>
>
>If c is just the same as zca, then it looks good. But if c includes not
>only zca, then this needs to be changed.
>
>
>> if (riscv_subset_supports (&riscv_rps_as, "ztso"))
>> diff --git a/gas/testsuite/gas/riscv/zca.d b/gas/testsuite/gas/riscv/zca.d
>> new file mode 100644
>> index 00000000000..86493985e70
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zca.d
>> @@ -0,0 +1,54 @@
>> +#as: -march=rv64i_zca
>> +#objdump: -d -Mno-aliases
>> +
>> +.*:[ ]+file format .*
>> +
>> +Disassembly of section .text:
>> +
>> +0+000 <target>:
>> +[ ]+[0-9a-f]+:[ ]+40fd[ ]+c.li[ ]+ra,31
>> +[ ]+[0-9a-f]+:[ ]+4101[ ]+c.li[ ]+sp,0
>> +[ ]+[0-9a-f]+:[ ]+6085[ ]+c.lui[ ]+ra,0x1
>> +[ ]+[0-9a-f]+:[ ]+61fd[ ]+c.lui[ ]+gp,0x1f
>> +[ ]+[0-9a-f]+:[ ]+4080[ ]+c.lw[ ]+s0,0\(s1\)
>> +[ ]+[0-9a-f]+:[ ]+5104[ ]+c.lw[ ]+s1,32\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+4502[ ]+c.lwsp[ ]+a0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+4082[ ]+c.lwsp[ ]+ra,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+6380[ ]+c.ld[ ]+s0,0\(a5\)
>> +[ ]+[0-9a-f]+:[ ]+6504[ ]+c.ld[ ]+s1,8\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+6502[ ]+c.ldsp[ ]+a0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+6082[ ]+c.ldsp[ ]+ra,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+c080[ ]+c.sw[ ]+s0,0\(s1\)
>> +[ ]+[0-9a-f]+:[ ]+d104[ ]+c.sw[ ]+s1,32\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+c02a[ ]+c.swsp[ ]+a0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+c006[ ]+c.swsp[ ]+ra,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+e380[ ]+c.sd[ ]+s0,0\(a5\)
>> +[ ]+[0-9a-f]+:[ ]+e504[ ]+c.sd[ ]+s1,8\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+e02a[ ]+c.sdsp[ ]+a0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+e006[ ]+c.sdsp[ ]+ra,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+0001[ ]+c.addi[ ]+zero,0
>> +[ ]+[0-9a-f]+:[ ]+0001[ ]+c.addi[ ]+zero,0
>> +[ ]+[0-9a-f]+:[ ]+007d[ ]+c.addi[ ]+zero,31
>> +[ ]+[0-9a-f]+:[ ]+908a[ ]+c.add[ ]+ra,sp
>> +[ ]+[0-9a-f]+:[ ]+05fd[ ]+c.addi[ ]+a1,31
>> +[ ]+[0-9a-f]+:[ ]+0101[ ]+c.addi[ ]+sp,0
>> +[ ]+[0-9a-f]+:[ ]+25fd[ ]+c.addiw[ ]+a1,31
>> +[ ]+[0-9a-f]+:[ ]+2101[ ]+c.addiw[ ]+sp,0
>> +[ ]+[0-9a-f]+:[ ]+0040[ ]+c.addi4spn[ ]+s0,sp,4
>> +[ ]+[0-9a-f]+:[ ]+6105[ ]+c.addi16sp[ ]+sp,32
>> +[ ]+[0-9a-f]+:[ ]+9c25[ ]+c.addw[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+8c05[ ]+c.sub[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+9c05[ ]+c.subw[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+8c65[ ]+c.and[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+887d[ ]+c.andi[ ]+s0,31
>> +[ ]+[0-9a-f]+:[ ]+8c45[ ]+c.or[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+8c25[ ]+c.xor[ ]+s0,s1
>> +[ ]+[0-9a-f]+:[ ]+8006[ ]+c.mv[ ]+zero,ra
>> +[ ]+[0-9a-f]+:[ ]+0006[ ]+c.slli[ ]+zero,0x1
>> +[ ]+[0-9a-f]+:[ ]+0002[ ]+c.slli64[ ]+zero
>> +[ ]+[0-9a-f]+:[ ]+d845[ ]+c.beqz[ ]+s0,0[
>> ]+\<target\>
>> +[ ]+[0-9a-f]+:[ ]+f45d[ ]+c.bnez[ ]+s0,0[
>> ]+\<target\>
>> +[ ]+[0-9a-f]+:[ ]+b775[ ]+c.j[ ]+0[ ]+\<target\>
>> +[ ]+[0-9a-f]+:[ ]+8082[ ]+c.jr[ ]+ra
>> +[ ]+[0-9a-f]+:[ ]+9082[ ]+c.jalr[ ]+ra
>> +[ ]+[0-9a-f]+:[ ]+9002[ ]+c.ebreak
>> diff --git a/gas/testsuite/gas/riscv/zca.s b/gas/testsuite/gas/riscv/zca.s
>> new file mode 100644
>> index 00000000000..1ad4f09eefd
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zca.s
>> @@ -0,0 +1,47 @@
>> +target:
>> + c.li x1, 31
>> + c.li x2, 0
>> + c.lui x1, 1
>> + c.lui x3, 31
>> + c.lw x8, (x9)
>> + c.lw x9, 32(x10)
>> + lw a0, (sp)
>> + c.lwsp x1, (x2)
>> + c.ld x8, (x15)
>> + c.ld x9, 8(x10)
>> + ld a0,(sp)
>> + c.ldsp x1, (sp)
>> + c.sw x8, (x9)
>> + c.sw x9, 32(x10)
>> + sw a0, (sp)
>> + c.swsp x1, (x2)
>> + c.sd x8, (x15)
>> + c.sd x9, 8(x10)
>> + sd a0, (sp)
>> + c.sdsp x1, (sp)
>> + addi x0, x0, 0
>> + c.nop
>> + c.nop 31
>> + c.add x1, x2
>> + c.addi a1, 31
>> + c.addi x2, 0
>> + c.addiw a1, 31
>> + c.addiw x2, 0
>> + c.addi4spn x8, x2, 4
>> + c.addi16sp x2, 32
>> + c.addw x8, x9
>> + c.sub x8, x9
>> + c.subw x8, x9
>> + c.and x8, x9
>> + c.andi x8, 31
>> + c.or x8, x9
>> + c.xor x8, x9
>> + c.mv x0, x1
>> + c.slli x0, 1
>> + c.slli64 x0
>> + c.beqz x8, target
>> + c.bnez x8, target
>> + c.j target
>> + c.jr ra
>> + c.jalr ra
>> + c.ebreak
>> diff --git a/gas/testsuite/gas/riscv/zcd.d b/gas/testsuite/gas/riscv/zcd.d
>> new file mode 100644
>> index 00000000000..11723c4ee14
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zcd.d
>> @@ -0,0 +1,16 @@
>> +#as: -march=rv64id_zcd
>> +#objdump: -d -Mno-aliases
>> +
>> +.*:[ ]+file format .*
>> +
>> +Disassembly of section .text:
>> +
>> +0+000 <target>:
>> +[ ]+[0-9a-f]+:[ ]+2108[ ]+c.fld[ ]+fa0,0\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+200c[ ]+c.fld[ ]+fa1,0\(s0\)
>> +[ ]+[0-9a-f]+:[ ]+2502[ ]+c.fldsp[ ]+fa0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+2582[ ]+c.fldsp[ ]+fa1,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+a108[ ]+c.fsd[ ]+fa0,0\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+a00c[ ]+c.fsd[ ]+fa1,0\(s0\)
>> +[ ]+[0-9a-f]+:[ ]+a02a[ ]+c.fsdsp[ ]+fa0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+a02e[ ]+c.fsdsp[ ]+fa1,0\(sp\)
>> diff --git a/gas/testsuite/gas/riscv/zcd.s b/gas/testsuite/gas/riscv/zcd.s
>> new file mode 100644
>> index 00000000000..3c29d323c7b
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zcd.s
>> @@ -0,0 +1,10 @@
>> +target:
>> + # ZCD only compress double float instructions
>> + fld fa0, 0(a0)
>> + c.fld fa1, 0(s0)
>> + fld fa0, 0(sp)
>> + c.fldsp fa1, 0(sp)
>> + fsd fa0, 0(a0)
>> + c.fsd fa1, 0(s0)
>> + fsd fa0, 0(sp)
>> + c.fsdsp fa1, 0(sp)
>> diff --git a/gas/testsuite/gas/riscv/zcf.d b/gas/testsuite/gas/riscv/zcf.d
>> new file mode 100644
>> index 00000000000..44dee021233
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zcf.d
>> @@ -0,0 +1,16 @@
>> +#as: -march=rv32if_zcf
>> +#objdump: -d -Mno-aliases
>> +
>> +.*:[ ]+file format .*
>> +
>> +Disassembly of section .text:
>> +
>> +0+000 <target>:
>> +[ ]+[0-9a-f]+:[ ]+6108[ ]+c.flw[ ]+fa0,0\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+600c[ ]+c.flw[ ]+fa1,0\(s0\)
>> +[ ]+[0-9a-f]+:[ ]+6502[ ]+c.flwsp[ ]+fa0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+6582[ ]+c.flwsp[ ]+fa1,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+e108[ ]+c.fsw[ ]+fa0,0\(a0\)
>> +[ ]+[0-9a-f]+:[ ]+e00c[ ]+c.fsw[ ]+fa1,0\(s0\)
>> +[ ]+[0-9a-f]+:[ ]+e02a[ ]+c.fswsp[ ]+fa0,0\(sp\)
>> +[ ]+[0-9a-f]+:[ ]+e02e[ ]+c.fswsp[ ]+fa1,0\(sp\)
>> diff --git a/gas/testsuite/gas/riscv/zcf.s b/gas/testsuite/gas/riscv/zcf.s
>> new file mode 100644
>> index 00000000000..c98fedb3d6c
>> --- /dev/null
>> +++ b/gas/testsuite/gas/riscv/zcf.s
>> @@ -0,0 +1,10 @@
>> +target:
>> + # ZCF only compress single float instructions with RV32
>> + flw fa0, 0(a0)
>> + c.flw fa1, 0(s0)
>> + flw fa0, 0(sp)
>> + c.flwsp fa1, 0(sp)
>> + fsw fa0, 0(a0)
>> + c.fsw fa1, 0(s0)
>> + fsw fa0, 0(sp)
>> + c.fswsp fa1, 0(sp)
>> --
>> 2.25.1
>>
>>
>Thanks! The test cases look pretty good for now :)
>
>Nelson
next prev parent reply other threads:[~2023-07-13 7:27 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-11 8:32 Jiawei
2023-07-13 2:21 ` Nelson Chu
2023-07-13 7:26 ` jiawei [this message]
2023-07-18 2:54 ` Nelson Chu
2023-07-18 14:54 ` Palmer Dabbelt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=72e31038.98e2.1894e25529c.Coremail.jiawei@iscas.ac.cn \
--to=jiawei@iscas.ac.cn \
--cc=binutils@sourceware.org \
--cc=charlie.keaney@embecosm.com \
--cc=chenyixuan@iscas.ac.cn \
--cc=christoph.muellner@vrull.eu \
--cc=gaofei@eswincomputing.com \
--cc=jbeulich@suse.com \
--cc=jeremy.bennett@embecosm.com \
--cc=kito.cheng@sifive.com \
--cc=mary.bennett@embecosm.com \
--cc=nandni.jamnadas@embecosm.com \
--cc=nelson@rivosinc.com \
--cc=palmer@dabbelt.com \
--cc=shihua@iscas.ac.cn \
--cc=shiyulong@iscas.ac.cn \
--cc=simon.cook@embecosm.com \
--cc=sinan.lin@linux.alibaba.com \
--cc=tariq.kurd@codasip.com \
--cc=wuwei2016@iscas.ac.cn \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).