Bug 1595923

Summary: Unsafe permission of allocations around resizes.
Product: Red Hat Enterprise Linux 7 Reporter: Sweet Tea Dorminy <sweettea>
Component: kmod-kvdoAssignee: Sweet Tea Dorminy <sweettea>
Status: CLOSED ERRATA QA Contact: Jakub Krysl <jkrysl>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 7.6CC: awalsh, jkrysl, limershe
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: 6.1.1.115 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-10-30 09:39:49 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Sweet Tea Dorminy 2018-06-27 19:23:32 UTC 
Description of problem:

This is a multi-part minor bug.

prepareToResizePhysical/Logical do their allocations wholly on a dm thread, not on a VDO thread. As such, they do not need to mark the layer as allocationsAllowed; their thread is registered with a NULL pointer to decide whether allocations are allowed, so they are always disallowed. Furthermore, a comment should be added that they may not do any work on threads, lest a allocation block on IO through VDO that requires the use of the blocked thread.

resizePhysical/Logical do no allocations, but have allocationsAllowed. The setting of allocationsAllowed is both no-effect currently, and dangerous: were there to be an allocation from something during the period currently allowed, it would potentially wait for a page to write out through the (suspended) VDO, deadlocking the thread allocating until the VDO were unsuspended.


Version-Release number of selected component (if applicable):
All

How reproducible:
Exceeding rare

Steps to Reproduce:
1. Wholly theoretical at this point

Actual results:
Possible deadlock

Expected results:
No deadlock
Comment 3 Jakub Krysl 2018-09-18 13:55:49 UTC 
regression testing passed with vdo vdo-6.1.1.120-3.el7.
Comment 5 errata-xmlrpc 2018-10-30 09:39:49 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2018:3094