From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <SRS0=HZd4=LD=kernel.crashing.org=benh@kernel.org>
Return-Path: <SRS0=HZd4=LD=kernel.crashing.org=benh@kernel.org>
Message-ID: <648c06a779a535266afa5ee7bd96ffba8f14d0ca.camel@kernel.crashing.org>
Subject: Re: [PATCH 1/1] PCI/AER: prevent pcie_do_fatal_recovery from using
 device after it is removed
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Sinan Kaya <okaya@kernel.org>, Keith Busch <keith.busch@intel.com>
Cc: poza@codeaurora.org, Bjorn Helgaas <helgaas@kernel.org>,
        Thomas Tai
	 <thomas.tai@oracle.com>, bhelgaas@google.com,
        linux-pci@vger.kernel.org, linux-pci-owner@vger.kernel.org,
        Sam Bobroff <sam.bobroff@au1.ibm.com>
Date: Tue, 21 Aug 2018 08:19:02 +1000
In-Reply-To: <184814fb-0caf-b260-4415-8f96fbdf4831@kernel.org>
References: <20180819021922.GE128050@bhelgaas-glaptop.roam.corp.google.com>
	 <fe891b63-12c7-61ab-a59a-ae24ad766754@kernel.org>
	 <ab5cb5b277f1c1c1362ef597ca7b8c6a2cd32cdf.camel@kernel.crashing.org>
	 <908ff33ded8f31830f95a8889d8540f1@codeaurora.org>
	 <5027d857bb59edfd33442003aa618ece1bc9cd52.camel@kernel.crashing.org>
	 <b2e2ac5c92d8a91d00606754d1046578@codeaurora.org>
	 <2ecd1fd6d763810d45697f846fa876b58a193b1b.camel@kernel.crashing.org>
	 <20180820155325.GA16148@localhost.localdomain>
	 <cf6de256280be7a2948db6dd2b5bfae92e70e12f.camel@kernel.crashing.org>
	 <6aa71d74-e4dc-c627-1496-981278388bce@kernel.org>
	 <20180820213544.GA16805@localhost.localdomain>
	 <0193e0c5a52d8f2e93958f0683f3acc81cef2bc9.camel@kernel.crashing.org>
	 <fd538f20-0754-f127-ceb9-693fec90b4af@kernel.org>
	 <44195c0d7159ce4cd01b660002b2f5af5d53678b.camel@kernel.crashing.org>
	 <184814fb-0caf-b260-4415-8f96fbdf4831@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
List-ID: <linux-pci.vger.kernel.org>

On Mon, 2018-08-20 at 18:13 -0400, Sinan Kaya wrote:
> 
> I proposed this as one of the possible solutions but presence detect is
> optional 

Is it optional on hotplug capable bridges ? I thought it wasn't ...

> and also presence detect interrupt can be delivered after

We don't care about the interrupt, we just need to look at the value of
the change latch.

> link layer interrupt as well. No guarantees with respect to the order of
> link layer interrupt and presence detect interrupts delivery.
> 
> I instead look at fatal error pending bit in device status register to
> decide if the link down was initiated due to a pending fatal error or
> somebody actually removed the card.
> 
> If fatal error is pending, wait until it is cleared. If link is healthy,
> return gracefully.
> 
> Otherwise, proceed with the hotplug removal.

That could work too but it would be safer to also use the change in
presence detect if available.

Cheers,
Ben.