[linux-lvm] Problems with 0.8pre tools (fwd)

[linux-lvm] Problems with 0.8pre tools (fwd)

[Michael Marxmeier]

Forwarded message ...

-------- Original Message --------
Date: Wed, 2 Feb 2000 23:39:30 -0500 (EST)
From: Jim Cain <me@jimcain.net>
Subject: Problems with 0.8pre tools

All,

I had some problems getting the tools to build. First, in
tools/lib/liblvm.h, the included file <linux/lvm.h> did not compile
because the reference to spinlock_t on line 573 was unknown. I added
an
#include <linux/spinlock.h> to liblvm.h, and everything in the lib
directory built fine.

Next, in tools/lvmchange.c on line 160, the identifier LVM_RESET was
unknown. This is defined in <linux/lvm.h>, but only if LVM_TOTAL_RESET
is
defined. I removed the #ifdef around it so that LVM_RESET is always
defined. After that change, the tools built okay.

Well, not completely okay. Because when I do the following:

pvcreate /dev/hd[bc]1
vgcreate vg01 /dev/hd[bc]1

the vgcreate command dumps core on a segmentation fault. I'm wondering
if
any of the fixes I made were incomplete or the wrong way to fix it.

This is being built on kernel 2.3.40 with the 0.8i patch.

Thanks for the help,
Jim

[linux-lvm] Problems with 0.8pre tools (fwd)

[Michael Marxmeier]

Forwarded message ...

-------- Original Message --------
Date: Thu, 3 Feb 2000 00:32:37 -0500 (EST)
Resent-Date: Thu, 03 Feb 2000 12:12:14 MEZ
From: Jim Cain <me@jimcain.net>
X-Sender: jec@polonius.jec.net
Subject: Re: Problems with 0.8pre tools

On Wed, 2 Feb 2000, Jim Cain wrote:

> On Wed, 2 Feb 2000, Jim Cain wrote:
> 
> > All,
> > 
> > I had some problems getting the tools to build. First, in
> > tools/lib/liblvm.h, the included file <linux/lvm.h> did not compile
> > because the reference to spinlock_t on line 573 was unknown. I added an
> > #include <linux/spinlock.h> to liblvm.h, and everything in the lib
> > directory built fine.
> > 
> > Next, in tools/lvmchange.c on line 160, the identifier LVM_RESET was
> > unknown. This is defined in <linux/lvm.h>, but only if LVM_TOTAL_RESET is
> > defined. I removed the #ifdef around it so that LVM_RESET is always
> > defined. After that change, the tools built okay.
> > 
> > Well, not completely okay. Because when I do the following:
> > 
> > pvcreate /dev/hd[bc]1
> > vgcreate vg01 /dev/hd[bc]1
> > 
> > the vgcreate command dumps core on a segmentation fault. I'm wondering if
> > any of the fixes I made were incomplete or the wrong way to fix it.
> > 
> > This is being built on kernel 2.3.40 with the 0.8i patch.
> 
> I should have also mentioned that I made the changes to
> tools/lib/lvm_tab_vg_check_exist_all_vg.c that Heinz describes in his
> message on Jan. 1. I get no output from vgcreate at all, just the seg
> fault.

Duh. I suppose I should do a little more digging on my own before I
email
to the list. I had a typo in the changes I made to that file, so I
fixed
that and just to be safe I did a make clean before the make.

I'm now creating PVs, VGs and LVs to my heart's content.

To Heinz and contributors: I must give you a sincere THANK YOU for
creating this fantastic software. LVM on Linux is something I've
wanted
for a long time.

Cheers,
Jim