Bug 475283

Summary: kpartx tries to create devices beyond the end of the disk.
Product: [Fedora] Fedora Reporter: Nate Clark <redhat>
Component: device-mapper-multipathAssignee: Milan Broz <mbroz>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: low    
Version: rawhideCC: agk, bmarzins, dwysocha, heinzm, kwizart, lex.lists, lvm-team, mbroz, prockai, pvrabec
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-03-06 13:16:06 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Subtracts the start of the extended partition from the overall volume offset.
none
Proposed patch none

Description Nate Clark 2008-12-08 19:34:48 UTC 
Description of problem:
kpartx when trying to create the corresponding linear devices on a dos partition with an extended partition fails. This is occurring because kpartx creates partitions that are inside a logical partition on top of the device that represents the extended partition.  The problem comes from the fact that when this is done instead of changing the offset of the partition to the start of the extended partition it continues to use the offset from the beginning of the entire volume.

Example DOS MBR:
# partition table of /dev/sda
unit: sectors

/dev/sda1 : start=       32, size=   206816, Id=83
/dev/sda2 : start=   206848, size=   206848, Id=83
/dev/sda3 : start=   413696, size=  6293504, Id=82
/dev/sda4 : start=  6707200, size= 56207360, Id= 5
/dev/sda5 : start=  6707232, size=  5244896, Id=83
/dev/sda6 : start= 11952160, size=  5244896, Id=83
/dev/sda7 : start= 17197088, size= 22022112, Id=83
/dev/sda8 : start= 39219232, size= 22022112, Id=83

bad dmtable made by kpartx:
mpath0p7: 0 22022112 linear 253:4 22442080
mpath0p6: 0 5244896 linear 253:4 11952192
mpath0p5: 0 5244896 linear 253:4 6707232
mpath0p4: 0 56207360 linear 253:0 6707200
mpath0p3: 0 6293504 linear 253:0 413696
mpath0p2: 0 206848 linear 253:0 206848
mpath0p1: 0 206816 linear 253:0 32

Corrected dmtable:
archive0p8: 0 22022112 linear 253:4 32512032
archive0p7: 0 22022112 linear 253:4 10489888
archive0p6: 0 5244896 linear 253:4 5244960
archive0p5: 0 5244896 linear 253:4 32
archive0p4: 0 56207360 linear 253:0 6707200
archive0p3: 0 6293504 linear 253:0 413696
archive0p2: 0 206848 linear 253:0 206848
archive0p1: 0 206816 linear 253:0 32

Attached is a patch which I used to work around this problem. So far it has worked for me.

Version-Release number of selected component (if applicable):
device-mapper-multipath-0.4.8-7.fc10.i386

How reproducible: 100%


Steps to Reproduce:
1. Create a partition table with extended partitions on a multipath device
2. Execute "kpartx -a <mpathdev>"
  
Actual results:
Fails to create some of the partitions.

Expected results:
Partitions created as linear device mapper devs.

Additional info:
Comment 1 Nate Clark 2008-12-08 19:38:45 UTC 
Created attachment 326169 [details]
Subtracts the start of the extended partition from the overall volume offset.

Patch mentioned in description.
Comment 2 Heinz Mauelshagen 2009-01-09 15:21:12 UTC 
*** Bug 479116 has been marked as a duplicate of this bug. ***
Comment 3 Nicolas Chauvet (kwizart) 2009-01-26 16:50:54 UTC 
Is there any improvements about how to fix this bug ?
Because this is really weird that it was working previously without the patch !!!
Or the problem might be elsewhere ?
Comment 4 Nate Clark 2009-01-26 17:08:48 UTC 
I think this use to work in previous versions of kpartx because all dm-linear devices use to sit on top of the same device. In this version only primary partitions sit on top of the base device and extended partitions sit on top of the appropriate partition.
Comment 5 Nicolas Chauvet (kwizart) 2009-03-05 11:52:36 UTC 
This patch is still needed in current rawhide !
Any chances of it to be fixed by the F-11 beta ?
Comment 6 Milan Broz 2009-03-06 12:45:46 UTC 
Ok, that's apparently serious problem, I'll fix that in rawhide.
Comment 7 Milan Broz 2009-03-06 12:48:04 UTC 
Created attachment 334287 [details]
Proposed patch

(patch from Hannes who originally added that code to kpartx - not sure why it is missing there)
Comment 8 Milan Broz 2009-03-06 13:16:06 UTC 
Fixed in device-mapper-multipath-0.4.8-9.fc11 build.