RE: [PATCH v5 02/15] ring-buffer: Page size per ring buffer

[David Laight <David.Laight@ACULAB.COM>]

From: Steven Rostedt
> Sent: 20 December 2023 13:01
> 
> On Wed, 20 Dec 2023 08:48:02 +0000
> David Laight <David.Laight@ACULAB.COM> wrote:
> 
> > From: Steven Rostedt
> > > Sent: 19 December 2023 18:54
> > > From: "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com>
> > >
> > > Currently the size of one sub buffer page is global for all buffers and
> > > it is hard coded to one system page. In order to introduce configurable
> > > ring buffer sub page size, the internal logic should be refactored to
> > > work with sub page size per ring buffer.
> > >
> > ...
> > > -	nr_pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE);
> > > +	/* Default buffer page size - one system page */
> > > +	buffer->subbuf_size = PAGE_SIZE - BUF_PAGE_HDR_SIZE;
> > > +
> > > +	/* Max payload is buffer page size - header (8bytes) */
> > > +	buffer->max_data_size = buffer->subbuf_size - (sizeof(u32) * 2);
> > > +
> > > +	nr_pages = DIV_ROUND_UP(size, buffer->subbuf_size);
> >
> > While not new, does this really make any sense for systems with 64k pages?
> > Wouldn't it be better to have units of 4k?
> 
> Unfortunately, it has to be PAGE_SIZE (and for now it's a power of 2 to
> make masking easy). It's used for splice and will also be used for memory
> mapping with user space.

Perhaps then the sysctl to set the size should be powers of 4k
with a minimum size of PAGE_SIZE.
Then you don't have to know the page size when setting things up.

I'm also guessing that no Linux kernels have a PAGE_SIZE of 2k?
IIRC some old mmu (maybe 68020 era) used 2k pages.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)