[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200207213857.GL1663@brightrain.aerifal.cx>
Date: Fri, 7 Feb 2020 16:38:57 -0500
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: Considering x86-64 fenv.s to C
On Sat, Feb 08, 2020 at 08:25:43AM +1100, Damian McGuckin wrote:
>
> In considering an i386 architecture with XMM capability, there is an issue
> with fegetenv()/feholdexcept() and fesetenv()/feupdateenv().
>
> There is no __mxcsr field in the fenv_t structure.
>
> So, there is no way to return the MXCSR using the first pair.
>
> Do I just bury any MXCSR exceptions into the __status_word?
>
> Regards - Damian
There's plenty of unused space to tuck it away somewhere if we wanted,
but I don't know if it's actually useful to do so. I'm fine with just
merging the states like the other fenv functions do.
Rich
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
