This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 0/2] RISC-V: Improve FP register support.
- From: Pedro Alves <palves at redhat dot com>
- To: Jim Wilson <jimw at sifive dot com>, gdb-patches at sourceware dot org
- Cc: Andrew Burgess <andrew dot burgess at embecosm dot com>
- Date: Tue, 23 Oct 2018 12:17:59 +0100
- Subject: Re: [PATCH 0/2] RISC-V: Improve FP register support.
- References: <CAFyWVabswzH+cwN6aBgPbP-ccUjUC2iya+HUyLjYzxLgUaM_Zg@mail.gmail.com>
On 10/19/2018 10:46 PM, Jim Wilson wrote:
> The first patch improves printing of values in FP registers. A 64-bit FP
> register can hold either float or double, and they are currently printed as
> only double. The patch changes this to print them as both float and
> double. Without the patch I see
> (gdb) print $fa0
> $1 = -nan(0xfffff3f828f5c)
> With the patch I see
> (gdb) print $fa0
> $1 = {float = 1.01999998, double = -nan(0xfffff3f828f5c)}
> Tested with a gdb testsuite run on a HiFive Unleashed running fedora
> core 29. The first patch has no effect on the results.
That screams "add some test then". :-)
Thanks,
Pedro Alves