Date: Sat, 5 Apr 2014 20:09:50 +0000 From: =?utf-8?B?V2Vpw58sIErDvHJnZW4=?= <weiss@uni-mainz.de> To: 'Andrew Turner' <andrew@fubar.geek.nz> Cc: "'freebsd-arm@freebsd.org'" <freebsd-arm@freebsd.org> Subject: RE: sleep command with armv6hf Message-ID: <22ab1b9199024cae9c8c9ff6f445c631@e15be-01.zdv.Uni-Mainz.DE> In-Reply-To: <20140405183958.65834b35@bender.Home> References: <79020061b2a74d8e8d56d47361ca3f4d@e15be-01.zdv.Uni-Mainz.DE> <20140405183958.65834b35@bender.Home>
index | next in thread | previous in thread | raw e-mail
Thanks very much for the patch. Now the function expects its arguments in the r0/r1 registers. The sleep command works as well. Regards Juergen > -----Original Message----- > From: Andrew Turner [mailto:andrew@fubar.geek.nz] > Sent: Saturday, April 05, 2014 7:40 PM > To: Weiß, Jürgen > Cc: 'freebsd-arm@freebsd.org' > Subject: Re: sleep command with armv6hf > > On Sat, 5 Apr 2014 16:25:08 +0000 > "Weiß, Jürgen" <weiss@uni-mainz.de> wrote: > > > under armv6hf the sleep command gives > > > > sleep: nanosleep: Invalid argument > > > > Further investigations seem to indicate the following problem: > > > > sleep calls the function __aeabi_d2lz > > > > vmov r0, r1, d8 > > bl __aeabi_d2lz > > mov r4, r0 > > mov r5, r1 > > > > with the argument in the r0/r1 registers. But the function > > in libgcc.a expects the argument in d0 (seems to be > > compiled with hardfloat as well). > > Can you try the attached patch. It should fix the function to use the > slightly odd calling convention of the __aeabi_* functions. > > Andrewhelp
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?22ab1b9199024cae9c8c9ff6f445c631>