linux-kernel - Re: [PATCH v7 3/3] PCI: Add tango MSI controller support

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: <a714b871-dd6c-4db0-3539-aabf7274938a@arm.com>
Date:   Wed, 14 Jun 2017 10:32:24 +0100
From:   Marc Zyngier <marc.zyngier@....com>
To:     Marc Gonzalez <marc_gonzalez@...madesigns.com>,
        Thomas Gleixner <tglx@...utronix.de>
Cc:     Bjorn Helgaas <helgaas@...nel.org>,
        Robin Murphy <robin.murphy@....com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Liviu Dudau <liviu.dudau@....com>,
        David Laight <david.laight@...lab.com>,
        linux-pci <linux-pci@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        LKML <linux-kernel@...r.kernel.org>, Mason <slash.tmp@...e.fr>
Subject: Re: [PATCH v7 3/3] PCI: Add tango MSI controller support

On 14/06/17 10:19, Marc Gonzalez wrote:
> On 14/06/2017 11:00, Marc Gonzalez wrote:
> 
>> The MSI controller in Tango supports 256 message-signaled interrupts,
>> and a single doorbell address.
>>
>> Signed-off-by: Marc Gonzalez <marc_gonzalez@...madesigns.com>
>> ---
>> Changes from v6 to v7
>> o Call spin_lock() not spin_lock_irqsave() in the ISR
>> ---
>>  drivers/pci/host/pcie-tango.c | 225 ++++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 225 insertions(+)
> 
> Someone on IRC suggested testing the driver with LOCKDEP.
> 
> If I understand the warning below correctly, I am not supposed
> to call irq_domain_set_info() while holding used_msi_lock?

Indeed. This creates an AB/BA situation, which will eventually deadlock.
Once again, lockdep saves the day.

> NB: in probe, my driver calls
> 
> 	add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
> 
> This should not break LOCKDEP analysis, right?

It doesn't. The code is provably wrong, and lockdep proved that it is wrong.

	M.
-- 
Jazz is not dead. It just smells funny...