[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ef2785806bb34b4688888063edab0588@AcuMS.aculab.com>
Date: Thu, 15 Feb 2024 19:41:33 +0000
From: David Laight <David.Laight@...LAB.COM>
To: 'Guenter Roeck' <linux@...ck-us.net>, John David Anglin
<dave.anglin@...l.net>, Charlie Jenkins <charlie@...osinc.com>, "Palmer
Dabbelt" <palmer@...belt.com>, Andrew Morton <akpm@...ux-foundation.org>,
Helge Deller <deller@....de>, "James E.J. Bottomley"
<James.Bottomley@...senpartnership.com>, Parisc List
<linux-parisc@...r.kernel.org>, Al Viro <viro@...iv.linux.org.uk>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH v8 2/2] lib: checksum: Use aligned accesses for
ip_fast_csum and csum_ipv6_magic tests
...
> > Descriptors need 8-byte alignment for efficiency on 64-bit parisc. The pc and gp are accessed
> > using ldd instructions.
> >
>
> How about the patch below ?
I think you still need something to avoid the misalignment trap
in the 'false' case.
If they only need to be aligned 'for efficiency' then I'd assume
the cpu (or whatever normally processes them) is ok with 4-byte
alignment even though 'ldd' (8 byte load?) faults.
Which would mean you need to read them with two 4-byte loads.
Especially if 'fn' isn't just one of a couple of specific values
that can be forced to be aligned.
But the code might just be completely broken.
(as suggested elsewhere).
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
Powered by blists - more mailing lists