linux-kernel - Re: [PATCH] PCI: Fix issue with "pci=disable_acs

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] [day] [month] [year] [list]

Message-ID: <f7f06f17-b644-dfea-e355-f44c8e920a33@deltatee.com>
Date:   Fri, 5 Apr 2019 16:11:46 -0600
From:   Logan Gunthorpe <logang@...tatee.com>
To:     Bjorn Helgaas <helgaas@...nel.org>
Cc:     linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
        Stephen Bates <sbates@...thlin.com>,
        Andrew Maier <andrew.maier@...eticom.com>
Subject: Re: [PATCH] PCI: Fix issue with "pci=disable_acs_redir" parameter
 being ignored

On 2019-04-05 3:39 p.m., Bjorn Helgaas wrote:
>> -static const char *disable_acs_redir_param;
>> +static char disable_acs_redir_param[COMMAND_LINE_SIZE];
> 
> I certainly acknowledge the problem, but I'm a little hesitant to add
> a static buffer of 256-4096 bytes (2048 on x86, arm64, powerpc, sparc)
> for a relatively low-usage situation.  The memory usage doesn't seem
> in proportion to the value-add.
> 
> Ugh, and we allocate another similar buffer for
> resource_alignment_param[], which I would guess is also rarely used.

Yeah, I was looking for a better solution and eventually decided to just
copy what was done in the past.

> Since disable_acs_redir_param[] and resource_alignment_param[] are
> both read-only and big enough to hold the entire command line, we
> should be able to share a single buffer between them if we made the
> parsers a little smarter.  

That sounds like a good idea that I didn't think of. If I can't find
something better, I'll do that.

> In fact, I bet there's already a static
> copy lying around somewhere for /proc/cmdline.

There are lots of copies of cmdline around, but none that are available
for early param. I did a sampling of some of the other early params but
none seem to be anywhere near as complex as the PCI one and I haven't
yet found one that needs to keep a copy of any or all of the command line.

In any case, I'll keep digging and try to get a better patch for you
shortly.

Logan