From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from rhirst.linuxcare.com (user-79-182.jakinternet.co.uk [194.88.79.182] (may be forged)) by puffin.external.hp.com (8.9.3/8.9.3) with ESMTP id LAA32016 for ; Mon, 11 Dec 2000 11:55:15 -0700 Received: by rhirst.linuxcare.com (Postfix, from userid 501) id 0DF1CB005; Mon, 11 Dec 2000 18:52:32 +0000 (GMT) Date: Mon, 11 Dec 2000 18:52:32 +0000 From: Richard Hirst To: Jes Sorensen Cc: Grant Grundler , parisc-linux@puffin.external.hp.com, debian-hppa@lists.debian.org Subject: Re: [parisc-linux] A500 status update Message-ID: <20001211185232.W7166@linuxcare.com> References: <200012070634.WAA07459@milano.cup.hp.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: ; from jes@linuxcare.com on Mon, Dec 11, 2000 at 04:37:16PM +0100 List-ID: On Mon, Dec 11, 2000 at 04:37:16PM +0100, Jes Sorensen wrote: > >>>>> "Grant" == Grant Grundler writes: > > Grant> o dd to > Grant> raw disk isn't working right. "No space left on device". See > Grant> the output below. > > Likely cause would be that the structs for stat don't match between > the kernel and glibc. I still think the best approach would be to go > for a clean 64 bit userland from the beginning as we have on the Alpha > and ia64 and provide backwards compat support afterwards. I don't like > to see the hppa64 port end up in the same situation as sparc64 where > almost nothing gets done for years because there is already a 32 bit > userland. This 'no space' problem was caused by an error in the register save/ restore on interrupts, for 64 bit (assembly.h): date: 2000/12/11 13:10:49; author: rhirst; state: Exp; lines: +10 -0 cr11 (sar) is a funny one. 5 bits on PA1.1 and 6 bit on PA2.0. For PA2.0 mtsar or mtctl always write 6 bits, but mfctl only reads 5 bits. Use mfctl,w to read all six bits. Otherwise we loose the 6th bit on a save/restore over interrupt. Caused dd of=/dev/sdaX to report no space available at random. Richard