[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CA+wH296d8p06E6LU4HUxcGZEf6ZkcUS_oEUY4k_zcC-Y6+-1+Q@mail.gmail.com>
Date: Thu, 12 May 2022 09:35:03 +0100
From: Alex Bradbury <asb@...radbury.org>
To: Changbin Du <changbin.du@...wei.com>
Cc: Craig Topper <craig.topper@...ive.com>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>,
Albert Ou <aou@...s.berkeley.edu>,
Steven Rostedt <rostedt@...dmis.org>, hw.huiwang@...wei.com,
linux-riscv@...ts.infradead.org, linux-kernel@...r.kernel.org,
llvm@...ts.linux.dev, codegen-riscv@...course.llvm.org,
llvmproject@...course.llvm.org
Subject: Re: riscv: llvm-compiler: calling convention violation: temporary
register $t2 is used to pass the ninth function parameter
On Thu, 12 May 2022 at 09:30, Changbin Du <changbin.du@...wei.com> wrote:
>
> On Wed, May 11, 2022 at 01:07:14PM -0700, Craig Topper wrote:
> > I’m guessing that because the function is static, the calling convention was changed to fastcall which allows us to ignore the ABI.
> >
> I think so. But the mcount function assumes the ABI is not changed.
It feels like this is a problem that would be encountered on other
architectures with clangbuiltlinux as well, due to dropping the ABI
requirements for local functions. Have you seen something similar
elsewhere Nick?
Best,
Alex
Powered by blists - more mailing lists