[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=wj73=Os1p=W7D2va=Rd81ZKvjb35yWgXQNgn1hXNRpAbw@mail.gmail.com>
Date: Mon, 6 Mar 2023 10:48:04 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Vernon Yang <vernon2gm@...il.com>
Cc: tytso@....edu, Jason@...c4.com, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
jejb@...ux.ibm.com, martin.petersen@...cle.com,
yury.norov@...il.com, andriy.shevchenko@...ux.intel.com,
linux@...musvillemoes.dk, james.smart@...adcom.com,
dick.kennedy@...adcom.com, linux-kernel@...r.kernel.org,
wireguard@...ts.zx2c4.com, netdev@...r.kernel.org,
linux-scsi@...r.kernel.org
Subject: Re: [PATCH 3/5] scsi: lpfc: fix lpfc_cpu_affinity_check() if no
further cpus set
On Mon, Mar 6, 2023 at 8:07 AM Vernon Yang <vernon2gm@...il.com> wrote:
>
> - if (new_cpu == nr_cpumask_bits)
> + if (new_cpu >= nr_cpumask_bits)
This all should use "nr_cpu_ids", not "nr_cpumask_bits".
But I really suspect that it should all be rewritten to not do that
thing over and over, but just use a helper function for it.
int lpfc_next_present_cpu(int n, int alternate)
{
n = cpumask_next(n, cpu_present_mask);
if (n >= nr_cpu_ids)
n = alternate;
return n;
}
and then you could just use
start_cpu = lpfc_next_present_cpu(new_cpu, first_cpu);
or similar.
Linus
PS. We "kind of" already have a helper function for this:
cpumask_next_wrap(). But it's really meant for a different pattern
entirely, so let's not confuse things.
Powered by blists - more mailing lists