Re: [PATCH v2] mm/zsmalloc: add statistics support

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Ganesh,

On Tue, Dec 23, 2014 at 10:26:12AM +0800, Ganesh Mahendran wrote:
> Hello Minchan
> 
> 2014-12-20 10:25 GMT+08:00 Minchan Kim <minchan@xxxxxxxxxx>:
> > Hey Ganesh,
> >
> > On Sat, Dec 20, 2014 at 09:43:34AM +0800, Ganesh Mahendran wrote:
> >> 2014-12-20 8:23 GMT+08:00 Minchan Kim <minchan@xxxxxxxxxx>:
> >> > On Fri, Dec 19, 2014 at 04:17:56PM -0800, Andrew Morton wrote:
> >> >> On Sat, 20 Dec 2014 09:10:43 +0900 Minchan Kim <minchan@xxxxxxxxxx> wrote:
> >> >>
> >> >> > > It involves rehashing a lengthy argument with Greg.
> >> >> >
> >> >> > Okay. Then, Ganesh,
> >> >> > please add warn message about duplicaed name possibility althoug
> >> >> > it's unlikely as it is.
> >> >>
> >> >> Oh, getting EEXIST is easy with this patch.  Just create and destroy a
> >> >> pool 2^32 times and the counter wraps ;) It's hardly a serious issue
> >> >> for a debugging patch.
> >> >
> >> > I meant that I wanted to change from index to name passed from caller like this
> >> >
> >> > zram:
> >> >         zs_create_pool(GFP_NOIO | __GFP_HIGHMEM, zram->disk->first_minor);
> >> >
> >> > So, duplication should be rare. :)
> >>
> >> We still can not know whether the name is duplicated if we do not
> >> change the debugfs API.
> >> The API does not return the errno to us.
> >>
> >> How about just zsmalloc decides the name of the pool-id, like pool-x.
> >> When the pool-id reaches
> >> 0xffff.ffff, we print warn message about duplicated name, and stop
> >> creating the debugfs entry
> >> for the user.
> >
> > The idea is from the developer point of view to implement thing easy
> > but my point is we should take care of user(ie, admin) rather than
> > developer(ie, we).
> 
> Yes. I got it.
> 
> >
> > For user, /sys/kernel/debug/zsmalloc/zram0 would be more
> > straightforward and even it doesn't need zram to export
> > /sys/block/zram0/pool-id.
> 
> BTW, If we add a new argument in zs_create_pool(). It seems we also need to
> add argument in zs_zpool_create(). So, zpool/zswap/zbud will be
> modified to support
> the new API.
> Is that acceptable?

I think it's doable.
The zpool_create_pool has already zswap_zpool_type.
Ccing maintainers for double check.

Many thanks.


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

-- 
Kind regards,
Minchan Kim

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



[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]