From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Hellstrom Date: Fri, 01 Apr 2011 08:17:37 +0000 Subject: Re: [PATCH] sparc32,leon: need VIRQ:IRQ 1:1, don't mask/ack IRQ Controller Message-Id: <4D958A21.1020902@gaisler.com> List-Id: References: <1300290864-20678-1-git-send-email-daniel@gaisler.com> In-Reply-To: <1300290864-20678-1-git-send-email-daniel@gaisler.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: sparclinux@vger.kernel.org David Miller wrote: >From: Daniel Hellstrom >Date: Thu, 31 Mar 2011 17:13:35 +0200 > > > >>This is sad, as we have designed all LEON architecture to be Plug&Play >>as far as possible, and it will be software that limits it in the >>end. >> >> > >Never is this the case. > >For any future device you can provide properly formed firmware >device tree nodes, with all IRQ entries included. > > That is the problem, some cores need tweeking in bootloader or SPARC/LEON low IRQ layers. One can think of it as the LEON has a read-only OpenBoot device tree in hardware, but only provides the first IRQ. >Stop fighting this, you made a mistake in the past, and the set of >broken device nodes that need fixing will not expand if you start >creating them properly starting now. > >You made a mistake, add handlers to cope with that mistake, and >take this opportunity to fix things so that future LEON device >nodes found in firmware are properly formed and contain multiple >IRQ entries when necessary. > > You are correct. The hardware should be redesigned to contain the full information, and in the meantime we must fix this in software. I will try fixing this in the bootloader at first and see how things goes. Thanks, Daniel