linux-kernel - Re: [PATCH 4/6] serial: 8250_dw: Rework IIR_NO

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]

Message-ID: <aXjEVtI7Bo5BCOiX@smile.fi.intel.com>
Date: Tue, 27 Jan 2026 15:57:42 +0200
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Jiri Slaby <jirislaby@...nel.org>,
	linux-serial <linux-serial@...r.kernel.org>,
	qianfan Zhao <qianfanguijin@....com>,
	Adriana Nicolae <adriana@...sta.com>,
	LKML <linux-kernel@...r.kernel.org>,
	"Bandal, Shankar" <shankar.bandal@...el.com>,
	"Murthy, Shanth" <shanth.murthy@...el.com>
Subject: Re: [PATCH 4/6] serial: 8250_dw: Rework IIR_NO_INT handling to stop
 interrupt storm

On Tue, Jan 27, 2026 at 03:01:46PM +0200, Ilpo Järvinen wrote:
> On Sat, 24 Jan 2026, Andy Shevchenko wrote:
> > On Fri, Jan 23, 2026 at 07:27:37PM +0200, Ilpo Järvinen wrote:

...

> > > +		d->no_int_count++;
> > > +		if (d->no_int_count > 2 && quirks & DW_UART_QUIRK_IER_KICK)
> > > +			dw8250_quirk_ier_kick(p);
> > 
> > Usual way is to use modulo. And perhaps use 4 for the sake of avoiding
> > division:
> > 
> > 		if (d->no_int_count == 3 && quirks & DW_UART_QUIRK_IER_KICK)
> > 			dw8250_quirk_ier_kick(p);
> > 
> > 		d->no_int_count = (d->no_int_count + 1) % 4;
> 
> This doesn't look equivalent code as it only fires on 4th NO_INT,

Correct, I forgot to clarify this in my original reply. Yes, bumping to
power-of-two for simplicity, but as you noticed it bumps also the loop to
"every 4th". (I was under impression that I wrote it somewhere else in
the reply, but now I see it's not the case.)

> but I guess the difference doesn't matter that much so changing to your
> suggestion so that the kick will only occurs on fourth NO_INT interrupt.

> > where 4 may be defined with meaningful name. With that u8 is more than enough.


-- 
With Best Regards,
Andy Shevchenko