mbox series

[0/3] btrfs: Convert kmap/memset/kunmap to memzero_user()

Message ID 20210309212137.2610186-1-ira.weiny@intel.com (mailing list archive)
Headers show
Series btrfs: Convert kmap/memset/kunmap to memzero_user() | expand

Message

Ira Weiny March 9, 2021, 9:21 p.m. UTC
From: Ira Weiny <ira.weiny@intel.com>

Previously this was submitted to convert to zero_user()[1].  zero_user() is not
the same as memzero_user() and in fact some zero_user() calls may be better off
as memzero_user().  Regardless it was incorrect to convert btrfs to
zero_user().

This series corrects this by lifting memzero_user(), converting it to
kmap_local_page(), and then using it in btrfs.

Thanks,
Ira

[1] https://lore.kernel.org/lkml/20210223192506.GY3014244@iweiny-DESK2.sc.intel.com/


Ira Weiny (3):
  iov_iter: Lift memzero_page() to highmem.h
  mm/highmem: Convert memzero_page() to kmap_local_page()
  btrfs: Use memzero_page() instead of open coded kmap pattern

 fs/btrfs/compression.c  |  5 +----
 fs/btrfs/extent_io.c    | 22 ++++------------------
 fs/btrfs/inode.c        | 33 ++++++++++-----------------------
 fs/btrfs/reflink.c      |  6 +-----
 fs/btrfs/zlib.c         |  5 +----
 fs/btrfs/zstd.c         |  5 +----
 include/linux/highmem.h |  7 +++++++
 lib/iov_iter.c          |  8 +-------
 8 files changed, 26 insertions(+), 65 deletions(-)

Comments

Andrew Morton March 10, 2021, 11:58 p.m. UTC | #1
On Tue,  9 Mar 2021 13:21:34 -0800 ira.weiny@intel.com wrote:

> Previously this was submitted to convert to zero_user()[1].  zero_user() is not
> the same as memzero_user() and in fact some zero_user() calls may be better off
> as memzero_user().  Regardless it was incorrect to convert btrfs to
> zero_user().
> 
> This series corrects this by lifting memzero_user(), converting it to
> kmap_local_page(), and then using it in btrfs.

This impacts btrfs more than MM.  I suggest the btrfs developers grab
it, with my

Acked-by: Andrew Morton <akpm@linux-foundation.org>
Ira Weiny March 11, 2021, 3:57 p.m. UTC | #2
On Wed, Mar 10, 2021 at 03:58:36PM -0800, Andrew Morton wrote:
> On Tue,  9 Mar 2021 13:21:34 -0800 ira.weiny@intel.com wrote:
> 
> > Previously this was submitted to convert to zero_user()[1].  zero_user() is not
> > the same as memzero_user() and in fact some zero_user() calls may be better off
> > as memzero_user().  Regardless it was incorrect to convert btrfs to
> > zero_user().
> > 
> > This series corrects this by lifting memzero_user(), converting it to
> > kmap_local_page(), and then using it in btrfs.
> 
> This impacts btrfs more than MM.  I suggest the btrfs developers grab
> it, with my

I thought David wanted you to take these this time?

"I can play the messenger again but now it seems a round of review is needed
and with some testing it'll be possible in some -rc. At that point you may take
the patches via the mm tree, unless Linus is ok with a late pull."

	-- https://lore.kernel.org/lkml/20210224123049.GX1993@twin.jikos.cz/

But reading that again I'm not sure what he meant.

David?

Ira

> 
> Acked-by: Andrew Morton <akpm@linux-foundation.org>
>
David Sterba March 12, 2021, 11:57 a.m. UTC | #3
On Thu, Mar 11, 2021 at 07:57:48AM -0800, Ira Weiny wrote:
> On Wed, Mar 10, 2021 at 03:58:36PM -0800, Andrew Morton wrote:
> > On Tue,  9 Mar 2021 13:21:34 -0800 ira.weiny@intel.com wrote:
> > 
> > > Previously this was submitted to convert to zero_user()[1].  zero_user() is not
> > > the same as memzero_user() and in fact some zero_user() calls may be better off
> > > as memzero_user().  Regardless it was incorrect to convert btrfs to
> > > zero_user().
> > > 
> > > This series corrects this by lifting memzero_user(), converting it to
> > > kmap_local_page(), and then using it in btrfs.
> > 
> > This impacts btrfs more than MM.  I suggest the btrfs developers grab
> > it, with my
> 
> I thought David wanted you to take these this time?
> 
> "I can play the messenger again but now it seems a round of review is needed
> and with some testing it'll be possible in some -rc. At that point you may take
> the patches via the mm tree, unless Linus is ok with a late pull."
> 
> 	-- https://lore.kernel.org/lkml/20210224123049.GX1993@twin.jikos.cz/
> 
> But reading that again I'm not sure what he meant.

As Linus had some objections I was not sure it was still feasible for
the merge window, but this is now sorted. This new patchset does further
changes in MM and the btrfs part is a straightforward cleanup. I've
noticed Andrew added the patches to his queue which I'd prefer so I've
added my reviewed-by to the third patch. Thanks.