Issue from Arch_test.h

[JAYANTH-IITM]

Hi @UmerShahidengr @Abdulwadoodd ,

When running ACT tests for Zcmp and Zcmt , I am getting error in RVTEST_CODE_BEGIN and in RVTEST_SIGBASE
if I comment out .option rvc and .option norvc , which wasn't previously , the test is passing and I am getting expected coverage , the error is attched below for reference .

[allenjbaum]

For which files are you commenting out .option rvc/norvc? These are compressed tests, so the tests (and not CODE_BEGIN/SIGBASE) should explicitly perform .push;.option rvc at the beginning, and a .pop at the end, if I remember the syntax correctly. Initialization code defaults to .norvc, if I recall correctly.

…

On Wed, Oct 23, 2024 at 7:59 AM JAYANTH-IITM ***@***.***> wrote: Hi @UmerShahidengr <https://github.com/UmerShahidengr> @Abdulwadoodd <https://github.com/Abdulwadoodd> , When running ACT tests for Zcmp and Zcmt , I am getting error in RVTEST_CODE_BEGIN and in RVTEST_SIGBASE if I comment out .option rvc and .option norvc , which wasn't previously , the test is passing and I am getting expected coverage , the error is attched below for reference . Screenshot.from.2024-10-23.20-22-24.png (view on web) <https://github.com/user-attachments/assets/a5d4c2c7-2b71-4eb6-a672-b2ea67ec29e4> — Reply to this email directly, view it on GitHub <#543>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHPXVJV4L3PU5GHXYIDAOLTZ462TRAVCNFSM6AAAAABQPBCJOSVHI2DSMVQWIX3LMV43ASLTON2WKOZSGYYDQOJTGE3DKMQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[JAYANTH-IITM]

Sir , I have commented out .option rvc/norvc for my zcmp and zcmt tests

[allenjbaum]

please send a URL to the test code; I don't see it in the dev repo

…

On Wed, Oct 23, 2024 at 10:47 PM JAYANTH-IITM ***@***.***> wrote: Sir , I have commented out .option rvc/norvc for my zcmp and zcmt tests — Reply to this email directly, view it on GitHub <#543 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHPXVJQWLGKUPEYLKHP3C6TZ5CCWNAVCNFSM6AAAAABQPBCJOSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZUGM2TINRTG4> . You are receiving this because you commented.Message ID: ***@***.***>

[JAYANTH-IITM]

arch_test.zip line number 1949 sir

[JAYANTH-IITM]

@allenjbaum @UmerShahidengr can you please help me here

[allenjbaum]

I'm stumped. This temporarily allows generation of compressed ops, aligns the code to a 32B address, then disallows it again. If this code is entered at a halfword aligned location, this enables the tools to generate a c.noop to get it aligned to a 4B boundary, enabling it to get aligned to a 32B boundary with regular noops. If the code is already aligned to a 4B boundary, it should have no effect. I really don't see how this should affect your test at all. Note that these are compressed ops, and your tests start with RVTEST_ISA("RV64I_Zca_Zcmt"). Maybe the problem is that you should be using RVTEST_ISA("RV64IC_Zca_Zcmt"). without that, it may be the case that your tests are not assuming compressed ops can be generated?

…

On Thu, Oct 24, 2024 at 10:31 PM JAYANTH-IITM ***@***.***> wrote: @allenjbaum <https://github.com/allenjbaum> @UmerShahidengr <https://github.com/UmerShahidengr> can you please help me here — Reply to this email directly, view it on GitHub <#543 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHPXVJUHGTBGIZJIDKZB2T3Z5HJTJAVCNFSM6AAAAABQPBCJOSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMZWHA3DQOJXGY> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[allenjbaum]

The only thing left that I can think of is you're using a version of the toolchain (including Sail) which doesn't support those extensions. If that doesn't work I am not sure how to proceed.

…

On Sun, Oct 27, 2024 at 9:48 PM JAYANTH-IITM ***@***.***> wrote: Include 'C' in my ISA (RV32IC_Zca_Zcmp) , yet the error exists . Screenshot.from.2024-10-28.10-17-24.png (view on web) <https://github.com/user-attachments/assets/32eecad6-bb2b-43fd-a55e-f5ac6aa41ad4> — Reply to this email directly, view it on GitHub <#543 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHPXVJQUNWBQZ7ZQUB2GMA3Z5W62XAVCNFSM6AAAAABQPBCJOSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINBQGU2DIOJQGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[UmerShahidengr]

I think this issue has been discussed in today DevPartners Meeting, @pz9115 has replied to this issue here. @JAYANTH-IITM are you satisfied with the response?

[JAYANTH-IITM]

yes umer , I will use the updated toolchain and redo the test . Thanks !