linux-kernel - Re: [PATCH v2 2/3] genhd, efi: add efi partition metadata to hd

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, 4 Aug 2010 09:44:55 -0500
From:	Will Drewry <wad@...omium.org>
To:	Kay Sievers <kay.sievers@...y.org>, Tejun Heo <tj@...nel.org>
Cc:	Karel Zak <kzak@...hat.com>, linux-kernel@...r.kernel.org,
	Jens Axboe <axboe@...nel.dk>,
	"David S. Miller" <davem@...emloft.net>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Joe Perches <joe@...ches.com>, Jan Blunck <jblunck@...e.de>,
	Greg Kroah-Hartman <gregkh@...e.de>
Subject: Re: [PATCH v2 2/3] genhd, efi: add efi partition metadata to 
	hd_structs

On Wed, Aug 4, 2010 at 5:14 AM, Kay Sievers <kay.sievers@...y.org> wrote:
> On Wed, Aug 4, 2010 at 11:00, Karel Zak <kzak@...hat.com> wrote:
>> On Tue, Aug 03, 2010 at 09:04:42PM -0500, Will Drewry wrote:
>>> This change extends the partition_meta_info structure to
>>> support EFI GPT-specific metadata and ensures that data
>>> is copied in on partition scanning.
>>
>> Why do want to store GPT-specific data (efi_guid_t) to
>> partition_meta_info? I think it would be better to use label and uuid
>> in a generic format (e.g. string or u8 uuid[16]) -- then you don't
>> have to use things like union, disklabel specific code to compare
>> uuids, etc.  IMHO your current code is too complicated.
>
> I don't mind having the raw data and the type accessible. It might be
> useful for something we don't know about and it basically comes for
> free.

I'll bump out the uuid at least, but it might be worth keeping an
extended meta info option.  But it's a lot less work to ditch it, so
I'm happy enough either way.


> But the only thing we are really interested in is the UUID, which,
> like Tejun already suggested, we should probably store
> format-independent, and have it always accessible. That way, we would
> not need any type-specific parser, we just handle the normal DCE
> format.

I'll bump it out and make it the efi code generic-ify its uuid.  Out
of curiousity, were you and Tejun thinking of keeping it as a 36 byte
string or as the 16 byte packed value.  While less space efficient,
I'd prefer the u8[36] as it avoids dealing with versioning when
parsing the user-supplied string.  Instead, each partition type can
properly unparse its uuids according to what they expect.

Seem reasonable?

> I don't think we should support any of the labels anyway in root= and
> similar, because they never really worked reliably with duplicates,
> and just ask for trouble.

Agreed. I don't think labels make sense, but we may later want to
support partition types (as I mention in my other mail).
--
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/