Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 13 Jul 2023 12:03:55 -0500
From:   Bjorn Helgaas <helgaas@kernel.org>
To:     Radu Rendec <rrendec@redhat.com>
Cc:     Jingoo Han <jingoohan1@gmail.com>,
        Gustavo Pimentel <gustavo.pimentel@synopsys.com>,
        Lorenzo Pieralisi <lpieralisi@kernel.org>,
        Krzysztof Wilczynski <kw@linux.com>,
        Rob Herring <robh@kernel.org>,
        Bjorn Helgaas <bhelgaas@google.com>,
        Marc Zyngier <maz@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org,
        Pali =?iso-8859-1?Q?Roh=E1r?= <pali@kernel.org>
Subject: Re: [PATCH 1/1] PCI: dwc: Use regular interrupt instead of chained
Message-ID: <20230713170355.GA323073@bhelgaas>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <8715f2b298c04422db311666eb2409464c29458b.camel@redhat.com>
Precedence: bulk

On Thu, Jun 29, 2023 at 11:04:50PM -0400, Radu Rendec wrote:
> On Thu, 2023-06-29 at 17:11 -0500, Bjorn Helgaas wrote:
> ...

> > If converting from chained to normal handlers can be done safely, I
> > would definitely be in favor if doing it across all of drivers/pci/ at
> > once so they're all consistent.? Otherwise that code just gets copied
> > to new drivers, so the issue persists and spreads.
> 
> I think the conversion can be done safely, meaning that it won't break
> the drivers. And by the way, there are other IRQ drivers (outside the
> PCI space) that use chained interrupts.
> 
> Unfortunately, it seems we are going in circles. Chained interrupts are
> bad because they let IRQ storms go unnoticed and lock up the system,
> but converting them to regular interrupts is also bad because it breaks
> the userspace ABI.
> 
> I am willing to help clean up this mess, but I think first we need to
> come up with a solution that's acceptable for everybody. I was hoping
> Marc and Thomas would chime in, but unfortunately that hasn't happened
> yet - other than each of them pointing out (separately) what is wrong
> with each approach.

I don't think Marc or Thomas are going to chime in with a fully-formed
solution.  I think to make progress, you (or Pali, or somebody) will
have to try to address Marc and Thomas' comments, make a proposal, and
we can iterate on it.

Bjorn