From mboxrd@z Thu Jan 1 00:00:00 1970 From: Grant Grundler Subject: Re: Yet another inline asm worry: mtsp() macro (and may be other)? Date: Sun, 22 Jun 2008 11:15:42 -0600 Message-ID: <20080622171542.GA31925@colo.lackof.org> References: <20080619161151.GB6049@colo.lackof.org> <485ABAFE.2010504@scarlet.be> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Grant Grundler , linux-parisc , Helge Deller , kyle To: Joel Soete Return-path: In-Reply-To: <485ABAFE.2010504@scarlet.be> List-ID: List-Id: linux-parisc.vger.kernel.org On Thu, Jun 19, 2008 at 08:01:02PM +0000, Joel Soete wrote: ... >> We could. But I thought there are cases were we might want to DMA to/from >> user space directly and that's why Space ID is a parameter. > > Ok but it was a BUG if temporary space reg is not 0, so??? It was a BUG in 2.4 kernels since 2.4 didn't have the VM support to do DMA to/from user space. I thought 2.6 kernels could DMA to/from user space (related to Direct IO). > btw system seems to boot too with asm volatile("mtsp > %%r0,%%sr1":::"memory") > may be the cpp can catch in the mtsp(gr, cr) macro if gr==0 then I can use > this insn if not the original macro??? Yes. I like willy's proposed patch for this. thanks, grant