Add RISC-V Profiles format #36

pz9115 · 2023-05-02T12:38:06Z

This is a subset work form #26.
Add RISC-V Profiles format only. The doc record in:
https://docs.google.com/document/d/1TZXRIgVfQHWQ6xrZflHXUCSav6xNmliojrW2bEsvPno/edit.

Add RISC-V Profiles format. Signed-off-by: Jiawei <[email protected]>

asb · 2023-05-03T10:06:29Z

README.mkd

+### Profile-based format
+
+Profiles should be recognized and used in the `-march=` option.
+The benefit is that it is easy for toolchain parsing the profiles string and expanding it into normal extensions combinations. It’s also easy to combine a profile with any extensions.


I'm not sure this is the right place to be advocating for why profiles make sense.

Regarding "It's also easy to combine a profile with any extensions" - this patch should give an example of that.

Fixed in new version, remove this descriptions, give an example behind.

kito-cheng · 2023-05-03T10:15:07Z

README.mkd

 If the 'C' (compressed) instruction set extension is targeted, the compiler 
 will generate compressed instructions where possible.


Those two lines should move before ### Profile-based format?

Thanks for your metion, moved~

kito-cheng · 2023-05-03T10:21:48Z

README.mkd

+which include all the mandatory extensions required by this profile.
+Here we are `not dealing with optional extensions` to profiles in the toolchain.
+
+Moreover, some extensions are still not ratified yet. So the toolchain only keeps their name strings only.


I would prefer drop that, RVA22 ratified, and means those ext. defined in RVA22 also ratified.

[1] https://github.com/riscv/riscv-profiles/releases/tag/v1.0

Droped these two lines.

README.mkd

Update profiles format descriptions, follow all comments in discuss. Thanks Alex Bradbury and Kito Cheng's review. Signed-off-by: Jiawei <[email protected]>

kito-cheng · 2023-06-30T07:42:40Z

README.mkd

+`-march` option is easy for toolchain parsing the profiles string and expanding
+it into normal extensions combinations. 
+
+Profiles format has the following form `-march=<profile-name>[+<option-ext>]+`.


I would suggest define that in BNF.

Updated, thanks for the suggestions.

preames · 2023-07-17T15:59:42Z

This got briefly brought up during the SIG toolchain meeting today, but got rapidly derailed into a process discussion. Since the root point seems important, I am repeating it here.

In addition to defining the syntax for the -march string for profiles, there is a separate set of task items required to define and implement support for each of the new extensions which were defined in the ratified profile document.

From the perspective of LLVM, my current intent is to make the support for all the various extensions a blocker for merging the -march profile parsing support. This may evolve, and I'm open to the arguments on why we should adopt a different posture, but that argument will need to be explicitly made.

pz9115 · 2023-07-17T16:15:20Z

This got briefly brought up during the SIG toolchain meeting today, but got rapidly derailed into a process discussion. Since the root point seems important, I am repeating it here.

In addition to defining the syntax for the -march string for profiles, there is a separate set of task items required to define and implement support for each of the new extensions which were defined in the ratified profile document.

From the perspective of LLVM, my current intent is to make the support for all the various extensions a blocker for merging the -march profile parsing support. This may evolve, and I'm open to the arguments on why we should adopt a different posture, but that argument will need to be explicitly made.

You are right, Christoph added a comment here riscv-admin/dev-partners#16 (comment). In fact, we need to nothing until those new extensions are actually ratified(more details are needed in extenisons specification to approach). So I think maybe is a good idea that only implements the new extensions' names in toolchain Profiles support.

topperc · 2023-07-18T00:22:30Z

README.mkd

+
+`-march = rv64imafdc_zicsr_ziccif_ziccrse_ziccamoa_zicclsm_za128rs_zba_zbb_zbc_zbs`
+
+and `-march = rva20u64` is an illegal profile input, it not use uppercase letters.


"it not use" -> "it does not use"

Updated, thanks for your checking.

Isn't this incorrect due to the spaces around the '=' character?

-march = rv64imafdc_zicsr_ziccif_ziccrse_ziccamoa_zicclsm_za128rs_zba_zbb_zbc_zbs

I think you are right, remove the incorrect spaces, thanks!

Update the syntax of description. Signed-off-by: Jiawei <[email protected]>

TommyMurphyTM1234 · 2023-07-18T13:09:46Z

README.mkd

+To distinguish between ordinary extension input and input with profiles,
+profiles are assumed to be entered `at the beginning of the -march option`, and
+then input other extensions. Profiles `should use uppercase letters` in the `-march`
+option.


Incorrect placement of code (backtick) markup?

I'm trying to emphasize here the rules for using, do you have any good solutions?

Suggested change

To distinguish between ordinary extension input and input with profiles,

profiles are assumed to be entered `at the beginning of the -march option`, and

then input other extensions. Profiles `should use uppercase letters` in the `-march`

option.

To distinguish between ordinary extension input and input with profiles,

profiles are assumed to be entered `at the beginning of the -march option`, and

then input other extensions. Profiles *should use uppercase letters* in the `-march`

option.

TommyMurphyTM1234 · 2023-07-18T13:10:27Z

README.mkd

+e.g. `-march = RVA20U64` is a legal profile input, it will be expanded into:
+
+`-march = rv64imafdc_zicsr_ziccif_ziccrse_ziccamoa_zicclsm_za128rs`,
+which include all the mandatory extensions required by this profile.
+
+`-march = RVA20U64+zba_zbb_zbc_zbs` is also a legal profile input, it will add
+four new extensions after expanded profile strings:
+
+`-march = rv64imafdc_zicsr_ziccif_ziccrse_ziccamoa_zicclsm_za128rs_zba_zbb_zbc_zbs`
+
+and `-march = rva20u64` is an illegal profile input, it does not use uppercase letters.


Spaces around '=' character are incorrect?

Removed, thank you!

Remove incorrect spaces using. Signed-off-by: Jiawei <[email protected]>

cmuellner · 2023-10-09T14:50:45Z

The status of the profiles support was briefly discussed in today's SIG toolchain call:

this was discussed at the GNU Cauldron with Kito and Palmer (I did not attend, therefore I can't say exactly what was discussed there)
patches are expected to get on the list this month (October 2023)

cmuellner · 2023-11-20T20:56:23Z

An initial patch was sent to the GCC mailing list today: https://gcc.gnu.org/pipermail/gcc-patches/2023-November/637490.html

jrtc27 · 2023-11-21T07:23:58Z

Uppercase strings being required in things like -march is very unusual; I would be much more in favour of keeping things lowercase

cmuellner · 2023-12-22T16:15:31Z

The architecture review committee has sent out the following text about how profiles support in toolchains should look like:

In response to some recent discussion in the Apps and Tools HC about how profiles should be represented in GCC/LLVM, the ARC provides this answer: compilers should use a single parameter for an ISA string. An ISA string begins with either a base ISA name (e.g. rv64i) or a profile name (e.g. rva23u64) and is optionally followed by additional extensions (e.g. rv64imac_zicond or rva23u64_zfh_zicond). If the ISA string begins with a profile name, it is equivalent to replacing the profile name with its mandatory base ISA and its mandatory extensions; any optional extensions in a profile must be explicitly named if their inclusion is desired. ISAs are sets, and concatenating strings takes the union, so redundancy is legal (e.g. rva23u64, rva23u64_zicsr, and rva23u64_zicsr_zicsr are all valid and equivalent).

https://lists.riscv.org/g/tech-unprivileged/message/611

wangpc-pp · 2023-12-25T11:08:39Z

Draft LLVM implementation: llvm/llvm-project#76357.

This PR implements the draft riscv-non-isa/riscv-toolchain-conventions#36. Currently, we replace specified profile in `-march` with standard arch string. We may need to pass it to backend so that we can emit an ELF attr proposed by riscv-non-isa/riscv-elf-psabi-doc#409.

wangpc-pp · 2024-02-07T07:16:14Z

Reverse ping. Is there any progress?

cmuellner · 2024-02-07T07:59:12Z

The GCC implementation is tracked here: riscv-admin/dev-partners#16.

There was a request to wait on the GCC patch regarding the adoption of profiles in toolchain components: https://gcc.gnu.org/pipermail/gcc-patches/2023-December/641025.html.

So, at this point, we have clear guidelines from RVI regarding the adoption. We have this PR, and we have patches for GCC and LLVM. This can all move forward if upstream maintainers are fine with it.

Given the relatively low impact on toolchains, I'm fine with waiting until the GCC 15 branch opens up.
LLVM might move on earlier.

This PR implements the draft riscv-non-isa/riscv-toolchain-conventions#36. Currently, we replace specified profile in `-march` with standard arch string.

This PR implements the draft riscv-non-isa/riscv-toolchain-conventions#36. Currently, we replace specified profile in `-march` with standard arch string. This is recommitted as 66f88de was reverted because of failures caused by lacking `--target` option.

This PR implements the draft riscv-non-isa/riscv-toolchain-conventions#36. Currently, we replace specified profile in `-march` with standard arch string.

This PR implements the draft riscv-non-isa/riscv-toolchain-conventions#36. Currently, we replace specified profile in `-march` with standard arch string. This is recommitted as 66f88de was reverted because of failures caused by lacking `--target` option.

cmuellner · 2024-03-26T11:35:27Z

LLVM support landed upstream a few days ago (see also llvm/llvm-project#76357).

kepstin · 2024-10-20T15:07:14Z

The text in this PR currently differs from the ARC decision mentioned in #36 (comment) and the implementation in LLVM in llvm/llvm-project#76357 in two significant ways:

Profile names are specified in lowercase, not uppercase
Additional extensions are specified with _ as the separator, not +

I worry that this inconsistency might cause compatibility issues.

wangpc-pp · 2024-10-22T04:17:12Z

@pz9115 Hi Jiawei, can you please update the RFC according to the comment?

The architecture review committee has sent out the following text about how profiles support in toolchains should look like:

In response to some recent discussion in the Apps and Tools HC about how profiles should be represented in GCC/LLVM, the ARC provides this answer: compilers should use a single parameter for an ISA string. An ISA string begins with either a base ISA name (e.g. rv64i) or a profile name (e.g. rva23u64) and is optionally followed by additional extensions (e.g. rv64imac_zicond or rva23u64_zfh_zicond). If the ISA string begins with a profile name, it is equivalent to replacing the profile name with its mandatory base ISA and its mandatory extensions; any optional extensions in a profile must be explicitly named if their inclusion is desired. ISAs are sets, and concatenating strings takes the union, so redundancy is legal (e.g. rva23u64, rva23u64_zicsr, and rva23u64_zicsr_zicsr are all valid and equivalent).

https://lists.riscv.org/g/tech-unprivileged/message/611

Add RISC-V Profiles format

ec70e4c

Add RISC-V Profiles format. Signed-off-by: Jiawei <[email protected]>

pz9115 mentioned this pull request May 2, 2023

Profiles Compilers riscv-admin/dev-partners#16

Open

13 tasks

asb reviewed May 3, 2023

View reviewed changes

kito-cheng reviewed May 3, 2023

View reviewed changes

Update profiles format descriptions

47389da

Update profiles format descriptions, follow all comments in discuss. Thanks Alex Bradbury and Kito Cheng's review. Signed-off-by: Jiawei <[email protected]>

pz9115 requested review from asb and kito-cheng June 29, 2023 04:47

kito-cheng reviewed Jun 30, 2023

View reviewed changes

Update Profile name defination with BNF.

11a65d4

pz9115 force-pushed the master branch from 9ff1eb6 to 11a65d4 Compare June 30, 2023 13:14

pz9115 requested a review from kito-cheng July 12, 2023 02:19

topperc reviewed Jul 18, 2023

View reviewed changes

Update the syntax.

378d5eb

Update the syntax of description. Signed-off-by: Jiawei <[email protected]>

pz9115 requested a review from topperc July 18, 2023 07:11

TommyMurphyTM1234 reviewed Jul 18, 2023

View reviewed changes

Remove incorrect spaces using.

f6bd873

Remove incorrect spaces using. Signed-off-by: Jiawei <[email protected]>

pz9115 requested a review from TommyMurphyTM1234 July 19, 2023 08:25

kito-cheng mentioned this pull request Dec 7, 2023

[RFC] Defining an ELF attribute for RISC-V target profiles riscv-non-isa/riscv-elf-psabi-doc#409

Closed

wangpc-pp mentioned this pull request Dec 25, 2023

[RISCV] Support RISC-V Profiles in -march option llvm/llvm-project#76357

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add RISC-V Profiles format #36

Add RISC-V Profiles format #36

pz9115 commented May 2, 2023

asb May 3, 2023

pz9115 Jun 29, 2023

kito-cheng May 3, 2023

pz9115 Jun 29, 2023

kito-cheng May 3, 2023

pz9115 Jun 29, 2023

kito-cheng Jun 30, 2023

pz9115 Jun 30, 2023

preames commented Jul 17, 2023

pz9115 commented Jul 17, 2023 •

edited

Loading

topperc Jul 18, 2023

pz9115 Jul 18, 2023

TommyMurphyTM1234 Jul 18, 2023

pz9115 Jul 19, 2023

TommyMurphyTM1234 Jul 18, 2023

pz9115 Jul 19, 2023

pthariensflame Mar 26, 2024

TommyMurphyTM1234 Jul 18, 2023

pz9115 Jul 19, 2023

cmuellner commented Oct 9, 2023

cmuellner commented Nov 20, 2023

jrtc27 commented Nov 21, 2023

cmuellner commented Dec 22, 2023

wangpc-pp commented Dec 25, 2023

wangpc-pp commented Feb 7, 2024

cmuellner commented Feb 7, 2024

cmuellner commented Mar 26, 2024

kepstin commented Oct 20, 2024

wangpc-pp commented Oct 22, 2024

		If the 'C' (compressed) instruction set extension is targeted, the compiler
		will generate compressed instructions where possible.


		`-march = rv64imafdc_zicsr_ziccif_ziccrse_ziccamoa_zicclsm_za128rs_zba_zbb_zbc_zbs`

		and `-march = rva20u64` is an illegal profile input, it not use uppercase letters.

Add RISC-V Profiles format #36

Are you sure you want to change the base?

Add RISC-V Profiles format #36

Conversation

pz9115 commented May 2, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

preames commented Jul 17, 2023

pz9115 commented Jul 17, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cmuellner commented Oct 9, 2023

cmuellner commented Nov 20, 2023

jrtc27 commented Nov 21, 2023

cmuellner commented Dec 22, 2023

wangpc-pp commented Dec 25, 2023

wangpc-pp commented Feb 7, 2024

cmuellner commented Feb 7, 2024

cmuellner commented Mar 26, 2024

kepstin commented Oct 20, 2024

wangpc-pp commented Oct 22, 2024

pz9115 commented Jul 17, 2023 •

edited

Loading