[linux-lvm] Inconsistent LVM code in 2.4.23-rc1

[linux-lvm] Inconsistent LVM code in 2.4.23-rc1

[James G. Sack (jim)]

[Hope this reaches the right place -- I'm not on any list  -- please
reply/cc direct]

Linux kernel 2.4.23-rc1 has LVM patched code that differs from LVM 1.0.8
(as well as cvs HEAD) from sistina.com

Q: is the 2.4.23 code OK?
  if yes, should it be in cvs
  if no, should somebody tell Marcelo?



In essance, there are extra calls to vfree(snap_lv_ptr) at 4 places in
lvm_do_vg_create() by virtue of a "goto copy_fault" --AND-- what looks
like a correction to vg_ptr->pe_allocated 

A simple diff looks like
------------------------
1587,1588c1587
< 	if (minor >= ABS_MAX_VG) {
< 		kfree(vg_ptr);
---
> 	if (minor >= ABS_MAX_VG)
1590d1588
< 	}
1658c1656,1657
< 				goto copy_fault;
---
> 				lvm_do_vg_remove(minor);
> 				return -EFAULT;
1669c1668,1669
< 				goto copy_fault;
---
> 				lvm_do_vg_remove(minor);
> 				return -EFAULT;
1679c1679,1680
< 			goto copy_fault;
---
> 			lvm_do_vg_remove(minor);
> 			return -EFAULT;
1682c1683,1684
< 			goto copy_fault;
---
> 			lvm_do_vg_remove(minor);
> 			return -EFAULT;
1697,1700d1698
< copy_fault:
< 	lvm_do_vg_remove(minor);
< 	vfree(snap_lv_ptr);
< 	return -EFAULT;
2597c2595
< 		vg_ptr->pe_allocated -= old_lv->lv_allocated_le;
---
> 		vg_ptr->pe_allocated -= old_lv->lv_allocated_snapshot_le;
2598a2597
> 		old_lv->lv_allocated_snapshot_le = new_lv->lv_allocated_le;
---------------------------------------------------------------------------

Regards,
..jim

[linux-lvm] Re: Inconsistent LVM code in 2.4.23-rc1

[Heinz J . Mauelshagen]

On Tue, Nov 18, 2003 at 10:33:56AM -0800, James G. Sack (jim) wrote:
> [Hope this reaches the right place -- I'm not on any list  -- please
> reply/cc direct]
> 
> Linux kernel 2.4.23-rc1 has LVM patched code that differs from LVM 1.0.8
> (as well as cvs HEAD) from sistina.com
> 
> Q: is the 2.4.23 code OK?

Yes, use the code in mainline.

>   if yes, should it be in cvs

Yes.

Will check code differences and get back to Marcelo if necessary.

Regards,
Heinz    -- The LVM Guy --


>   if no, should somebody tell Marcelo?
> 
> 
> 
> In essance, there are extra calls to vfree(snap_lv_ptr) at 4 places in
> lvm_do_vg_create() by virtue of a "goto copy_fault" --AND-- what looks
> like a correction to vg_ptr->pe_allocated 
> 
> A simple diff looks like
> ------------------------
> 1587,1588c1587
> < 	if (minor >= ABS_MAX_VG) {
> < 		kfree(vg_ptr);
> ---
> > 	if (minor >= ABS_MAX_VG)
> 1590d1588
> < 	}
> 1658c1656,1657
> < 				goto copy_fault;
> ---
> > 				lvm_do_vg_remove(minor);
> > 				return -EFAULT;
> 1669c1668,1669
> < 				goto copy_fault;
> ---
> > 				lvm_do_vg_remove(minor);
> > 				return -EFAULT;
> 1679c1679,1680
> < 			goto copy_fault;
> ---
> > 			lvm_do_vg_remove(minor);
> > 			return -EFAULT;
> 1682c1683,1684
> < 			goto copy_fault;
> ---
> > 			lvm_do_vg_remove(minor);
> > 			return -EFAULT;
> 1697,1700d1698
> < copy_fault:
> < 	lvm_do_vg_remove(minor);
> < 	vfree(snap_lv_ptr);
> < 	return -EFAULT;
> 2597c2595
> < 		vg_ptr->pe_allocated -= old_lv->lv_allocated_le;
> ---
> > 		vg_ptr->pe_allocated -= old_lv->lv_allocated_snapshot_le;
> 2598a2597
> > 		old_lv->lv_allocated_snapshot_le = new_lv->lv_allocated_le;
> ---------------------------------------------------------------------------
> 
> Regards,
> ..jim

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-