linux-kernel - Re: [PATCH v2 4/5] mfd: tqmx86: make IRQ setup errors non-fatal

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] [day] [month] [year] [list]

Date: Mon, 24 Jun 2024 11:47:58 +0100
From: Lee Jones <lee@...nel.org>
To: Matthias Schiffer <matthias.schiffer@...tq-group.com>
Cc: linux-kernel@...r.kernel.org,
	Gregor Herburger <gregor.herburger@...group.com>,
	linux@...tq-group.com
Subject: Re: [PATCH v2 4/5] mfd: tqmx86: make IRQ setup errors non-fatal

On Mon, 24 Jun 2024, Matthias Schiffer wrote:

> On Thu, 2024-06-20 at 17:35 +0100, Lee Jones wrote:
> > 
> > On Mon, 17 Jun 2024, Matthias Schiffer wrote:
> > 
> > > GPIO IRQ setup can fail either because an invalid IRQ was passed as a
> > > parameter, or because the GPIO controller does not support interrupts.
> > > Neither is severe enough to stop the whole probe; simply disable IRQ
> > > support in the GPIO resource when setup fails.
> > > 
> > > Signed-off-by: Matthias Schiffer <matthias.schiffer@...tq-group.com>
> > > ---
> > > 
> > > v2: no changes (was patch 3/4)
> > > 
> > >  drivers/mfd/tqmx86.c | 7 ++++---
> > >  1 file changed, 4 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/mfd/tqmx86.c b/drivers/mfd/tqmx86.c
> > > index 5aa51ead00a28..7f9ccd83278dd 100644
> > > --- a/drivers/mfd/tqmx86.c
> > > +++ b/drivers/mfd/tqmx86.c
> > > @@ -259,13 +259,14 @@ static int tqmx86_probe(struct platform_device *pdev)
> > >  		err = tqmx86_setup_irq(dev, "GPIO", gpio_irq, io_base,
> > >  				       TQMX86_REG_IO_EXT_INT_GPIO_SHIFT);
> > >  		if (err)
> > > -			return err;
> > > +			gpio_irq = 0;
> > > +	}
> > >  
> > > +	if (gpio_irq)
> > 
> > Stacking identical if()s one after another doesn't sound very efficient.
> > 
> > Why not put the contents of this one inside the one above?
> 
> The intention was to have the "else" branch be executed both when gpio_irq was reset to 0 in the
> above error path, and when gpio_irq was 0 in the first place (so the above section running
> tqmx86_setup_irq() hasn't even been executed).
> 
> I got a better idea now however - by initializing flags to 0 and only setting it together with the
> IRQ in the success path, no (!gpio_irq) branch is needed here at all. Will change in v3.

That sounds like a better approach, thanks.

-- 
Lee Jones [李琼斯]