netdev - Re: [PATCH net-next V4 00/12] net/mlx5: ConnectX-4 100G Ethernet driver

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:	Wed, 20 May 2015 17:17:13 +0300
From:	Amir Vadai <amirv@...lanox.com>
To:	David Miller <davem@...emloft.net>
Cc:	Amir Vadai <amirv@...lanox.com>, netdev <netdev@...r.kernel.org>,
	Or Gerlitz <ogerlitz@...lanox.com>,
	Tal Alon <talal@...lanox.com>,
	Achiad Shochat <achiad@...lanox.com>,
	Saeed Mahameed <saeedm@...lanox.com>
Subject: Re: [PATCH net-next V4 00/12] net/mlx5: ConnectX-4 100G Ethernet driver

On Tue, May 19, 2015 at 7:41 PM, David Miller <davem@...emloft.net> wrote:
> From: Amir Vadai <amirv@...lanox.com>
> Date: Tue, 19 May 2015 12:25:12 +0300
>
>> On Sun, May 17, 2015 at 8:05 PM, David Miller <davem@...emloft.net> wrote:
>>> From: Amir Vadai <amirv@...lanox.com>
>>> Date: Sun, 17 May 2015 16:02:11 +0300
>>>
>>>> We didn't get a response yet regarding your comment about the irq
>>>> renaming [3].
>>>
>>> Well then, please hold off on resubmissions of this series until you
>>> do get a response and that issue is firmly resolved.
>> Hi,
>>
>> I don't mean to push you, I only want to understand what is expected
>> from me and what are the next steps:
>> How will the issue be resolved? Do you plan to answer my question [1]
>> from last week, and just too busy right now or something like that?
>
> I have not seen any response to me explaining why it's ok to change
> the IRQ name strings in the context where this will occur.
>
> Once you explain that, we can make forward progress, but only at that
> point.

Hi Dave,

Just to put us back on the same page, repeating a bit the previous chapters..

You wrote [1] that if we change these names after the request_irq() call(s), the
new name string will not propagate to /proc/interrupts output.

So, indeed, request_irq() is called when the driver is loaded (and we
don't know yet if the port types are Infiniband or Ethernet). Only later
on, we rename the name when the Ethernet interface is up and we know
its name.

Fact is that the new name does propagate to /proc/interrupts.

Also, looking in the code, I don't see a reason why shouldn't it
be properly updated. When calling request_irq(), the name argument
is not copied, but irq_desc->action->name points to it. This same pointer
is being used by show_interrupts() when /proc/interrupts is shown.

All in all, unless I somehow missed your precise question or I didn't
explain myself clearly, I don't see what is still missing in my reply,
can you please shed some light?

What I did find is that the /proc/irq/N/handler directory name
which is a copy of the original action->name doesn't change. However
AFAIK, this directory isn't being used anywhere in the kernel.

[1] - http://www.spinics.net/lists/netdev/msg328444.html

Amir.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html