Re: mmotm 2020-04-17-20-35 uploaded (squashfs)

[Matthew Wilcox <willy@infradead.org>]

On Sat, Apr 18, 2020 at 12:47:28PM -0700, Andrew Morton wrote:
> On Sat, 18 Apr 2020 08:56:31 -0700 Randy Dunlap <rdunlap@infradead.org> wrote:
> 
> > On 4/17/20 8:36 PM, akpm@linux-foundation.org wrote:
> > > The mm-of-the-moment snapshot 2020-04-17-20-35 has been uploaded to
> > > 
> > >    http://www.ozlabs.org/~akpm/mmotm/
> > > 
> > > mmotm-readme.txt says
> > > 
> > > README for mm-of-the-moment:
> > > 
> > > http://www.ozlabs.org/~akpm/mmotm/
> > > 
> > > This is a snapshot of my -mm patch queue.  Uploaded at random hopefully
> > > more than once a week.
> > > 
> > > You will need quilt to apply these patches to the latest Linus release (5.x
> > > or 5.x-rcY).  The series file is in broken-out.tar.gz and is duplicated in
> > > http://ozlabs.org/~akpm/mmotm/series
> > > 
> > > The file broken-out.tar.gz contains two datestamp files: .DATE and
> > > .DATE-yyyy-mm-dd-hh-mm-ss.  Both contain the string yyyy-mm-dd-hh-mm-ss,
> > > followed by the base kernel version against which this patch series is to
> > > be applied.
> > 
> > on x86_64:
> > 
> >   CC      fs/squashfs/decompressor_multi_percpu.o
> > ../fs/squashfs/decompressor_multi_percpu.c:75:5: error: conflicting types for ‘squashfs_decompress’
> >  int squashfs_decompress(struct squashfs_sb_info *msblk, struct buffer_head **bh,
> >      ^~~~~~~~~~~~~~~~~~~
> 
> Thanks.  Seems that file was missed.
> 
> Also, this code jumps through horrifying hoops in order to initialize
> locals at their definition site.  But the code looks so much better if
> we Just Don't Do That!

I think the code would look even better if things just Had The Right Type!

struct squashfs_sb_info {
...
-       struct squashfs_stream                  *stream;
+	union {
+		struct squashfs_stream		*stream;
+		struct squashfs_stream __percpu	*percpu_stream;
+	};

int squashfs_decompress(struct squashfs_sb_info *msblk, struct bio *bio,
		int offset, int length, struct squashfs_page_actor *output)
{
	struct squashfs_stream *stream = get_cpu_ptr(msblk->percpu_stream);
	int res = msblk->decompressor->decompress(msblk, stream->stream, bh, b,
			offset, length, output);
	...

As an aside, that calling convention could do with putting some of the
arguments into a struct so the CPU spends less time shuffling arguments
from one register to another.