zoned-vm-stats-add-nr_anon.patch

zoned-vm-stats-add-nr_anon.patch

[Martin Bligh]

Could we rename nr_mapped to something else if we're going to change
it's meaning? Perhaps split nr_mapped into nr_mapped_file and
nr_mapped_anon or something?

Otherwise any poor schmuck (eg me) who has to work across different
kernel versions is going to get rather confused in the future.

In my mind, "nr_mapped" is a good name for the number of pages which
are mapped, so excluding the anon pages from that seems to make
the naming non-obvious. similarly, I presume we can have anon pages
on transition to or from swap that are not mapped, and yet will
not be reflected here, so nr_anon doesn't seem like a wholly
accurate name to me either?

M.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: zoned-vm-stats-add-nr_anon.patch

[Christoph Lameter]

On Tue, 20 Jun 2006, Martin Bligh wrote:

> Could we rename nr_mapped to something else if we're going to change
> it's meaning? Perhaps split nr_mapped into nr_mapped_file and
> nr_mapped_anon or something?

Yes we did that. nr_mapped was split into NR_MAPPED and NR_ANON. Please 
read the description for V4 if this patchset that was posted last week.

> In my mind, "nr_mapped" is a good name for the number of pages which
> are mapped, so excluding the anon pages from that seems to make
> the naming non-obvious. similarly, I presume we can have anon pages
> on transition to or from swap that are not mapped, and yet will
> not be reflected here, so nr_anon doesn't seem like a wholly

The same confusion exist for nr_dirty. Should we also rename nr_dirty to 
nr_dirty_file?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: zoned-vm-stats-add-nr_anon.patch

[Martin Bligh]

Christoph Lameter wrote:
> On Tue, 20 Jun 2006, Martin Bligh wrote:
> 
> 
>>Could we rename nr_mapped to something else if we're going to change
>>it's meaning? Perhaps split nr_mapped into nr_mapped_file and
>>nr_mapped_anon or something?
> 
> 
> Yes we did that. nr_mapped was split into NR_MAPPED and NR_ANON. Please 
> read the description for V4 if this patchset that was posted last week.

Yeah, but ... that's what I'm concerned about, the naming of it.
Splitting it makes sense, just needs to be renamed something else, I think.

>>In my mind, "nr_mapped" is a good name for the number of pages which
>>are mapped, so excluding the anon pages from that seems to make
>>the naming non-obvious. similarly, I presume we can have anon pages
>>on transition to or from swap that are not mapped, and yet will
>>not be reflected here, so nr_anon doesn't seem like a wholly
> 
> The same confusion exist for nr_dirty. Should we also rename nr_dirty to 
> nr_dirty_file?

Sure. Naming is important, IMHO. People reading code make involuntary
assumptions as they read code, it's inevitable.

M.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: zoned-vm-stats-add-nr_anon.patch

[Christoph Lameter]

On Tue, 20 Jun 2006, Martin Bligh wrote:

> Sure. Naming is important, IMHO. People reading code make involuntary
> assumptions as they read code, it's inevitable.

Ok. I will make a pass over all the counter before sending the next 
patchset to Andrew and make sure that they get a proper suffix if their 
role is restricted.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: zoned-vm-stats-add-nr_anon.patch

[Martin Bligh]

Christoph Lameter wrote:
> On Tue, 20 Jun 2006, Martin Bligh wrote:
> 
> 
>>Sure. Naming is important, IMHO. People reading code make involuntary
>>assumptions as they read code, it's inevitable.
> 
> 
> Ok. I will make a pass over all the counter before sending the next 
> patchset to Andrew and make sure that they get a proper suffix if their 
> role is restricted.

Awesome - thanks very much.

M.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>