From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Luck, Tony" Date: Mon, 09 Dec 2002 22:59:30 +0000 Subject: RE: [Linux-ia64] unalinged access by loadpair instruction Message-Id: List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org You do recall correctly ... checking the archives, I posted a patch on October 16, 2001. I did fix the floating point case back then too ... in fact my post claims to include the patch for both integer and FP. But I must have attached the wrong patch file. I've long since deleted all my 2.4.10 trees, so I'd have to reconstruct from scratch (can't do it from memory, those neurons have been re-assigned :-( -Tony -----Original Message----- From: David Mosberger [mailto:davidm@napali.hpl.hp.com] Sent: Monday, December 09, 2002 1:42 PM To: Hideki Yamamoto Cc: linux-ia64@linuxia64.org Subject: Re: [Linux-ia64] unalinged access by loadpair instruction It looks to me like getfpreg() and setfpreg() are pretty obviously broken: they doesn't take rotation into account at all. Seems this was missed when the integer-side was fixed (by Tony, IIRC). Anyone want to take a shot at fixing this? --david >>>>> On Mon, 09 Dec 2002 23:28:45 +0900, "Hideki Yamamoto" said: Hideki> Hi everyone, Hideki> When unaligned access is happened by loadpair instruction, Hideki> reloading by kernel is wrong. Hideki> When unalined access is happeded by loadpair instruction in Hideki> NORMAL loop, it is no problem. Please try to run the Hideki> attached programs(c.c and a.s) compiled by gcc. Hideki> However, When it is happned by loadpair instruction in SWP Hideki> loop, it fails to reload the data. Please try to run the Hideki> attached programs(c.c and aa.s) compiled by gcc. aa.s is Hideki> included loadpair in SWP loop. If there is no happening, Hideki> please change the n parameter to bigger number. Hideki> Sorry, I could not isolate whether or not it is caused by Hideki> HW bug or Kernel problem. Hideki> Let me know somebody understood the cause why it is Hideki> happened. _______________________________________________ Linux-IA64 mailing list Linux-IA64@linuxia64.org http://lists.linuxia64.org/lists/listinfo/linux-ia64