[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <38da93959bc948c992fc7fc1efd4d25c@AcuMS.aculab.com>
Date: Thu, 25 Mar 2021 15:59:56 +0000
From: David Laight <David.Laight@...LAB.COM>
To: "'Matthew Wilcox (Oracle)'" <willy@...radead.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Andrew Morton" <akpm@...ux-foundation.org>,
Christoph Lameter <cl@...ux.com>,
"Pekka Enberg" <penberg@...nel.org>,
David Rientjes <rientjes@...gle.com>,
"Joonsoo Kim" <iamjoonsoo.kim@....com>,
Vlastimil Babka <vbabka@...e.cz>,
"Uladzislau Rezki" <urezki@...il.com>,
Nicholas Piggin <npiggin@...il.com>
Subject: RE: [PATCH v2 0/4] vmalloc: Improve vmalloc(4MB) performance
From: Matthew Wilcox
> Sent: 24 March 2021 15:05
>
> I may have gone a little too far with the first patch. It turns out we
> have a lot of code using the vmalloc caller functionality, and I decided
> to convert it all to unsigned long instead of leaving some paths as void *
> and casting later.
What is the 'address of the caller' needed for?
If it gets printed anywhere it ought to be a pointer type
throughout so that it will get hashed and not leak the
real address.
Even then it is probably better to use a typed pointer
(eg to an undefined structure) than 'void *'.
That might also pick up more incorrect uses.
Of course things like _RET_IP_ would need changing to return
the same pointer type.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
Powered by blists - more mailing lists