Revision 8204f8ddaafafcae074746fcf2a05a45e6827603 authored by Darrick J. Wong on 10 August 2017, 21:20:28 UTC, committed by Darrick J. Wong on 17 August 2017, 19:40:33 UTC
Way back when we established inode block-map redo log items, it was
discovered that we needed to prevent the VFS from evicting inodes during
log recovery because any given inode might be have bmap redo items to
replay even if the inode has no link count and is ultimately deleted,
and any eviction of an unlinked inode causes the inode to be truncated
and freed too early.

To make this possible, we set MS_ACTIVE so that inodes would not be torn
down immediately upon release.  Unfortunately, this also results in the
quota inodes not being released at all if a later part of the mount
process should fail, because we never reclaim the inodes.  So, set
MS_ACTIVE right before we do the last part of log recovery and clear it
immediately after we finish the log recovery so that everything
will be torn down properly if we abort the mount.

Fixes: 17c12bcd30 ("xfs: when replaying bmap operations, don't let unlinked inodes get reaped")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
1 parent e28ae8e
Raw File
Kconfig.recursion-issue-02
# Cumulative Kconfig recursive issue
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# Test with:
#
# make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-02 allnoconfig
#
# The recursive limitations with Kconfig has some non intuitive implications on
# kconfig sematics which are documented here. One known practical implication
# of the recursive limitation is that drivers cannot negate features from other
# drivers if they share a common core requirement and use disjoint semantics to
# annotate those requirements, ie, some drivers use "depends on" while others
# use "select". For instance it means if a driver A and driver B share the same
# core requirement, and one uses "select" while the other uses "depends on" to
# annotate this, all features that driver A selects cannot now be negated by
# driver B.
#
# A perhaps not so obvious implication of this is that, if semantics on these
# core requirements are not carefully synced, as drivers evolve features
# they select or depend on end up becoming shared requirements which cannot be
# negated by other drivers.
#
# The example provided in Documentation/kbuild/Kconfig.recursion-issue-02
# describes a simple driver core layout of example features a kernel might
# have. Let's assume we have some CORE functionality, then the kernel has a
# series of bells and whistles it desires to implement, its not so advanced so
# it only supports bells at this time: CORE_BELL_A and CORE_BELL_B. If
# CORE_BELL_A has some advanced feature CORE_BELL_A_ADVANCED which selects
# CORE_BELL_A then CORE_BELL_A ends up becoming a common BELL feature which
# other bells in the system cannot negate. The reason for this issue is
# due to the disjoint use of semantics on expressing each bell's relationship
# with CORE, one uses "depends on" while the other uses "select". Another
# more important reason is that kconfig does not check for dependencies listed
# under 'select' for a symbol, when such symbols are selected kconfig them
# as mandatory required symbols. For more details on the heavy handed nature
# of select refer to Documentation/kbuild/Kconfig.select-break
#
# To fix this the "depends on CORE" must be changed to "select CORE", or the
# "select CORE" must be changed to "depends on CORE".
#
# For an example real world scenario issue refer to the attempt to remove
# "select FW_LOADER" [0], in the end the simple alternative solution to this
# problem consisted on matching semantics with newly introduced features.
#
# [0] http://lkml.kernel.org/r/1432241149-8762-1-git-send-email-mcgrof@do-not-panic.com

mainmenu "Simple example to demo cumulative kconfig recursive dependency implication"

config CORE
	tristate

config CORE_BELL_A
	tristate
	depends on CORE

config CORE_BELL_A_ADVANCED
	tristate
	select CORE_BELL_A

config CORE_BELL_B
	tristate
	depends on !CORE_BELL_A
	select CORE
back to top