linux-kernel - Re: [PATCH 3/3] x86,pci,acpi: Handle invalid

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, 21 Apr 2010 16:43:42 -0700
From:	Yinghai <yinghai.lu@...cle.com>
To:	"H. Peter Anvin" <hpa@...or.com>
CC:	Bjorn Helgaas <bjorn.helgaas@...com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	Jesse Barnes <jbarnes@...tuousgeek.org>,
	Andy Isaacson <adi@...apodia.org>, guenter.roeck@...csson.com,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
	"x86@...nel.org" <x86@...nel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Thomas Renninger <trenn@...e.de>
Subject: Re: [PATCH 3/3] x86,pci,acpi: Handle invalid _CRS

On 04/21/2010 04:10 PM, H. Peter Anvin wrote:
> On 04/21/2010 04:04 PM, Bjorn Helgaas wrote:
>> On Wednesday 21 April 2010 04:33:28 pm H. Peter Anvin wrote:
>>> Do you have opinions on patches 1-2 of the series?
>>>
>>> I'm getting concerned about how the size of the patchset has grown, and
>>> we're past -rc5 already... but it is a regression so we can't just defer
>>> it to .35.
>>
>> Part 1: the essential part of this seems to be the trim_bios_range()
>> change, and that part is not too big.  In v4, Yinghai also removed
>> probe_roms_32.c.  That sounds like the right thing to do, but I'd
>> rather have that in separate patch so it doesn't obfuscate the other
>> change, and I don't know whether it *has* to be done for .34; maybe
>> it could be deferred.
> 
> I would agree with that.

then use -v3 please

-v4: also don't trim [0xa0000, 0x100000] for mrst.

> 
>> Part 2: IMHO, we're putting way too much crap in kernel/resource.c.
>> A name like "reserve_region_with_split_check_child()" is a pretty
>> good clue that we've lost our way somewhere.  But that's mostly a
>> cosmetic thing, and the end result does seem to be something that
>> fixes the current regression.
> 
> It's not just a good clue we have lost our way, it's also completely
> impossible for anyone but Yinghai to divine what the intended semantics
> are supposed to be.  This *greatly* concerns me, especially given
> previous track record.

I don't know.

insert_resource_expand_to_fit() is added by Linus.

And at least he knew old reserve_region_with_split()

> 
> Even the checkin comment is almost unparsable, which makes it very
> likely that someone is going to trip up on some of this in the future.
> I really would like to get a better description.
> 
> The use of a string match in:
> 
> +       if (check_child && !conflict->child && strstr(conflict->name,
> "PCI Bus"))
> ^^^^^^^^^
> 
> ... screams "wrong! ugly! bad!" in my opinion.  I utterly fail to see
> how that could be acceptable under any circumstances.  I thought that
> had been flagged earlier in the conversation, but it is apparently still
> there.


the string checking is to make sure pci device that is hooked into bus0 directly, but pci bar is falling into
0xa0000 - 0x100000.
so can not put "reserved" holder under them.

YH
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/