This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] aarch64 sim FP to unsigned integer conversion fix
- From: Nick Clifton <nickc at redhat dot com>
- To: Jim Wilson <jim dot wilson at linaro dot org>, gdb-patches at sourceware dot org
- Date: Tue, 3 Jan 2017 14:36:49 +0000
- Subject: Re: [PATCH] aarch64 sim FP to unsigned integer conversion fix
- Authentication-results: sourceware.org; auth=none
- References: <CABXYE2W4SHYN5kS_9DLbTV6D6ancjTgnAGgS6vxKjavtNW4mmw@mail.gmail.com>
Hi Jim,
> FP to unsigned integer conversions are being truncated to the signed
> INT_MAX and LONG_MAX when they should be truncated to UINT_MAX and
> ULONG_MAX. This requires adding a bunch of new variables, as this
> truncation happens inside a macro that uses token pasting to create
> variable names.
>
> The testcase fails without the patch and works with the patch. GCC C
> testsuite failures drop from 2568 to 2538 with the patch.
Approved - please apply.
Cheers
Nick