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: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Mon, 12 Apr 2021 14:43:45 +0200
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Cédric Le Goater <clg@...d.org>,
        Nicholas Piggin <npiggin@...il.com>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH] genirq: reduce irqdebug bouncing cachelines

Cédric,

On Mon, Apr 12 2021 at 11:06, Cédric Le Goater wrote:
> On 4/10/21 1:58 PM, Thomas Gleixner wrote:
>> --- a/kernel/irq/spurious.c
>> +++ b/kernel/irq/spurious.c
>> @@ -274,7 +274,7 @@ void note_interrupt(struct irq_desc *des
>>  	unsigned int irq;
>>  
>>  	if (desc->istate & IRQS_POLL_INPROGRESS ||
>> -	    irq_settings_is_polled(desc))
>> +	    irq_settings_is_polled(desc) | irq_settings_no_debug(desc))
>
> Shouldn't it be '||' instead  ?

It could. But that's intentionally '|'. Why?

Because that lets the compiler merge the bit checks into one and
therefore spares one conditional branch.

>>  		return;
>>  
>>  	if (bad_action_ret(action_ret)) {
>> 
>
> We could test irq_settings_no_debug() directly under handle_nested_irq() 
> and handle_irq_event_percpu() to avoid calling note_interrupt(), just 
> like we do for noirqdebug.

We can do that, but then we should not just make it:

   if (!irqnodebug && !irq_settings_no_debug(desc))
   	note_interrupt(...);

Instead have only one condition:

   if (!irq_settings_no_debug(desc))
   	note_interrupt(...);

See the uncompiled delta patch below.

Thanks,

        tglx
---
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -1690,6 +1690,9 @@ static int
 				irq_settings_set_no_debug(desc);
 		}
 
+		if (noirqdebug)
+			irq_settings_set_no_debug(desc);
+
 		if (new->flags & IRQF_ONESHOT)
 			desc->istate |= IRQS_ONESHOT;
 
--- a/kernel/irq/spurious.c
+++ b/kernel/irq/spurious.c
@@ -274,7 +274,7 @@ void note_interrupt(struct irq_desc *des
 	unsigned int irq;
 
 	if (desc->istate & IRQS_POLL_INPROGRESS ||
-	    irq_settings_is_polled(desc) | irq_settings_no_debug(desc))
+	    irq_settings_is_polled(desc))
 		return;
 
 	if (bad_action_ret(action_ret)) {
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -481,7 +481,7 @@ void handle_nested_irq(unsigned int irq)
 	for_each_action_of_desc(desc, action)
 		action_ret |= action->thread_fn(action->irq, action->dev_id);
 
-	if (!noirqdebug)
+	if (!irq_settings_no_debug(desc))
 		note_interrupt(desc, action_ret);
 
 	raw_spin_lock_irq(&desc->lock);
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -197,7 +197,7 @@ irqreturn_t handle_irq_event_percpu(stru
 
 	add_interrupt_randomness(desc->irq_data.irq, flags);
 
-	if (!noirqdebug)
+	if (!irq_settings_no_debug(desc))
 		note_interrupt(desc, retval);
 	return retval;
 }

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ