From: "Alexander E. Patrakov" <patrakov@ums.usu.ru>
To: Carl-Daniel Hailfinger <c-d.hailfinger.kernel.2004@gmx.net>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH] Partitioned loop devices, support for 127 Partitions on SATA, IDE and SCSI
Date: Thu, 11 Nov 2004 18:55:06 +0500 [thread overview]
Message-ID: <200411111855.06266.patrakov@ums.usu.ru> (raw)
In-Reply-To: <4192F069.90209@gmx.net>
On Thursday 11 November 2004 09:54, Carl-Daniel Hailfinger wrote:
> Alexander E. Patrakov schrieb:
> > Carl-Daniel Hailfinger wrote:
> >>Hi,
> >>
> >>having seen the problems people have when switching from traditional IDE
> >>drivers to libata if they have more than 15 partitions, I decided to do
> >>something against it. With this patch (and recreating /dev/loop* nodes)
> >>it is possible to support up to 127 partitions per loop device
> >>regardless what the underlying device supports. It works for me
> >>and has the added bonus that it will be in compatibility mode as long
> >>as you don't specify the max_part parameter.
> >
> > Why not just use EVMS? Partition code is supposed to be moved to
> > userspace anyway.
>
> Because my solution works fine with userspace partitioning code (I tested
> with partx from util-linux) and has the big advantage that partitions
> actually appear at the right place in /sys/block/loopN/loopNpM. Most
> other solutions for many partitions per device failed to make the
> relationship between parent device and partition visible in sysfs.
> I haven't checked yet how EVMS handles this. Could you post
> find /sys/block/$SOME_EVMS_DISK/ -type d
> for a normal disk which is completely managed by EVMS so I can verify
> whether that would be satisfactory. Thanks.
That is not satisfactory since this relation is not expressed in sysfs at all
(device-mapper issue, discuss it on dm-devel if you want). EVMS is using
normal dm devices (like LVM2 does), so, although there are subdirectories
of /sys/block/hdc, they correspond only to normal ("obsolete", kernel-space,
unusable together with EVMS) partitions.
patrakov@lfs:~$ find /sys/block/hdc/ -type d
/sys/block/hdc/
/sys/block/hdc/queue
/sys/block/hdc/queue/iosched
/sys/block/hdc/hdc1
/sys/block/hdc/hdc2
and so on
(The last two entries correspond to unusable devices)
Real block devices with filesystems on them devices are there:
patrakov@lfs:~$ find /sys/block/dm* -type d
/sys/block/dm-0
/sys/block/dm-1
/sys/block/dm-2
/sys/block/dm-3
/sys/block/dm-4
/sys/block/dm-5
/sys/block/dm-6
/sys/block/dm-7
/sys/block/dm-8
EVMS installation is non-intrusive if you want to use it just for loop devices
(just enable device mapper in vanilla linux-2.6.9 and install userspace
tools), so you can try it without any risk.
--
Alexander E. Patrakov
prev parent reply other threads:[~2004-11-11 13:54 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-11-10 4:31 Carl-Daniel Hailfinger
2004-11-10 6:05 ` Alexander E. Patrakov
2004-11-11 4:54 ` Carl-Daniel Hailfinger
2004-11-11 13:55 ` Alexander E. Patrakov [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200411111855.06266.patrakov@ums.usu.ru \
--to=patrakov@ums.usu.ru \
--cc=c-d.hailfinger.kernel.2004@gmx.net \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®