linux-kernel - Re: [PATCH 4/5] pci: handled return value of platform_get

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:   Thu, 12 Mar 2020 09:11:02 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Marc Gonzalez <marc.w.gonzalez@...e.fr>
Cc:     Aman Sharma <amanharitsh123@...il.com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Andrew Murray <amurray@...goodpenguin.co.uk>,
        Linus Walleij <linus.walleij@...aro.org>,
        Ryder Lee <ryder.lee@...iatek.com>,
        Karthikeyan Mitran <m.karthikeyan@...iveil.co.in>,
        Hou Zhiqiang <Zhiqiang.Hou@....com>,
        Mans Rullgard <mans@...sr.com>,
        Matthias Brugger <matthias.bgg@...il.com>,
        linux-pci@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, linux-mediatek@...ts.infradead.org,
        Marc Zyngier <maz@...nel.org>
Subject: Re: [PATCH 4/5] pci: handled return value of platform_get_irq
 correctly

[+cc another Marc]

On Thu, Mar 12, 2020 at 10:53:06AM +0100, Marc Gonzalez wrote:
> On 11/03/2020 20:19, Aman Sharma wrote:
> 
> > diff --git a/drivers/pci/controller/pcie-tango.c b/drivers/pci/controller/pcie-tango.c
> > index 21a208da3f59..18c2c4313eb5 100644
> > --- a/drivers/pci/controller/pcie-tango.c
> > +++ b/drivers/pci/controller/pcie-tango.c
> > @@ -273,9 +273,9 @@ static int tango_pcie_probe(struct platform_device *pdev)
> >  		writel_relaxed(0, pcie->base + SMP8759_ENABLE + offset);
> >  
> >  	virq = platform_get_irq(pdev, 1);
> > -	if (virq <= 0) {
> > +	if (virq < 0) {
> >  		dev_err(dev, "Failed to map IRQ\n");
> > -		return -ENXIO;
> > +		return virq;
> >  	}
> >  
> >  	irq_dom = irq_domain_create_linear(fwnode, MSI_MAX, &dom_ops, pcie);
> 
> Weee, here we go again :-)
> 
> https://patchwork.kernel.org/patch/11066455/
> https://patchwork.kernel.org/patch/10006651/
> 
> Last time around, my understanding was that, going forward,
> the best solution was:
> 
> 	virq = platform_get_irq(...)
> 	if (virq <= 0)
> 		return virq ? : -ENODEV;
> 
> i.e. map 0 to -ENODEV, pass other errors as-is, remove the dev_err
> 
> @Bjorn/Lorenzo did you have a change of heart?

Yes.  In 10006651 (Oct 20, 2017), I thought:

  irq = platform_get_irq(pdev, 0);
  if (irq <= 0)
    return -ENODEV;

was fine.  In 11066455 (Aug 7, 2019), I said I thought I was wrong and
that:

  platform_get_irq() is a generic interface and we have to be able to
  interpret return values consistently.  The overwhelming consensus
  among platform_get_irq() callers is to treat "irq < 0" as an error,
  and I think we should follow suit.
  ...
  I think the best pattern is:

    irq = platform_get_irq(pdev, i);
    if (irq < 0)
      return irq;

I still think what I said in 2019 is the right approach.  I do see
your comment in 10006651 about this pattern:

  if (virq <= 0)
    return virq ? : -ENODEV;

but IMHO it's too complicated for general use.  Admittedly, it's not
*very* complicated, but it's a relatively unusual C idiom and I
stumble over it every time I see it.  If 0 is a special case I think
it should be mapped to a negative error in arch-specific code, which I
think is what Linus T suggested in [1].

I think there's still a large consensus that "irq < 0" is the error
case.  In the tree today we have about 1400 callers of
platform_get_irq() and platform_get_irq_byname() [2].  Of those,
almost 900 check for "irq < 0" [3], while only about 150 check for
"irq <= 0" [4] and about 15 use some variant of a "irq ? : -ENODEV"
pattern.

The bottom line is that in drivers/pci, I'd like to see either a
single style or a compelling argument for why some checks should be
"irq < 0" and others should be "irq <= 0".

[1] https://yarchive.net/comp/linux/zero.html
[2] $ git grep "=.*platform_get_irq" | wc -l
    1422
[3] $ git grep -A4 "=.*platform_get_irq" | grep "<\s*0" | wc -l
    894
[4] $ git grep -A4 "=.*platform_get_irq" | grep "<=\s*0" | wc -l
    151
[5] $ git grep -A4 "=.*platform_get_irq" | grep "return.*?.*:.*;" | wc -l
    15