linux-kernel - Re: preempt.h: some SOFTIRQ_OFFSET should be SOFTIRQ

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAK1hOcNETR+J6nYqidQKyybHmbHgiPvrm604vK=rLqRVOd6XTg@mail.gmail.com>
Date:   Wed, 13 Feb 2019 11:37:45 +0100
From:   Denys Vlasenko <vda.linux@...glemail.com>
To:     Frederic Weisbecker <frederic@...nel.org>
Cc:     Frederic Weisbecker <fweisbec@...il.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: preempt.h: some SOFTIRQ_OFFSET should be SOFTIRQ_MASK?

On Wed, Feb 13, 2019 at 5:05 AM Frederic Weisbecker <frederic@...nel.org> wrote:
> On Tue, Feb 05, 2019 at 07:34:31PM +0100, Denys Vlasenko wrote:
> > SOFTIRQ is a counter.
> > Why here:
> >
> > #define in_serving_softirq()    (softirq_count() & SOFTIRQ_OFFSET)
> > #define in_task()               (!(preempt_count() & \
> >                                    (NMI_MASK | HARDIRQ_MASK | SOFTIRQ_OFFSET)))
> >
> > we check only lowest bit?
>
> So we have SOFTIRQ_OFFSET that is used when serving softirqs.
> And we have SOFTIRQ_DISABLE_OFFSET that is used when we disable
> softirqs.
>
> I think the choice is right on both tests above, or am I missing something?

>From my reading of the following:

 * We put the hardirq and softirq counter into the preemption
 * counter. The bitmask has the following meaning:
 * - bits 0-7 are the preemption count (max preemption depth: 256)
 * - bits 8-15 are the softirq count (max # of softirqs: 256)
 * The hardirq count could in theory be the same as the number of
 * interrupts in the system, but we run all interrupt handlers with
 * interrupts disabled, so we cannot have nesting interrupts. Though
 * there are a few palaeontologic drivers which reenable interrupts in
 * the handler, so we need more than one bit here.
 *         PREEMPT_MASK:        0x000000ff
 *         SOFTIRQ_MASK:        0x0000ff00
 *         HARDIRQ_MASK:        0x000f0000
 *             NMI_MASK:        0x00100000
 * PREEMPT_NEED_RESCHED:        0x80000000
 */
#define PREEMPT_BITS    8
#define SOFTIRQ_BITS    8
#define HARDIRQ_BITS    4
#define NMI_BITS        1

it seems that 8-bit SOFTIRQ_MASK contains depth count of nested softirqs.
Therefore, to test whether we are in softirq, you need:

define in_serving_softirq()    (softirq_count() & SOFTIRQ_MASK)

But existing code uses SOFTIRQ_OFFSET, not SOFTIRQ_MASK, for the mask.
This means that if nest count is e.g. 2, in_serving_softirq() will return
"false".