Re: [RFC][PATCH] zram: drop partial_io support

[Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>]

Hello Minchan,

On (12/16/15 10:01), Minchan Kim wrote:
> Hello Sergey,
> 
> Sorry for the late response. I am in long vavacation now but today,
> I get small time to sit down on computer. :)

Have a good one!

> On Mon, Dec 14, 2015 at 09:38:55PM +0900, Sergey Senozhatsky wrote:
> > Hello,
> > 
> > I've been thinking about this for some time, but didn't have a chance
> > to properly investigate so far. My question is: why do we even bother
> > with partial IO in zram?
> 
> It was done before I involved zram actively so I should spend a time
> to search the reason.

Thanks a lot! I appreciate this. I was about to do the same but still
a bit too busy.

> Firstly, author was Jerome.
> http://lists.openwall.net/linux-kernel/2011/06/10/318
> 
> And Nitin wanted to increase logical block size 64K instead of
> making complex part by partial I/O.
> http://lists.openwall.net/linux-kernel/2011/06/10/402
> 
> And Jeff and Martin said there is no problem to increase
> logical_block_size from unsigned short if people are aware of
> the implications bigger blocks have on the filesystems they put on top.
> 
> http://lists.openwall.net/linux-kernel/2011/06/14/289
> http://lists.openwall.net/linux-kernel/2011/06/14/324
> 
> Jerome finally found severe problem which FAT fs are unable to
> cope with 64K logical blocks at least.
> http://lists.openwall.net/linux-kernel/2011/07/01/196
>
> That's why Nitin decided to suppport partial IO in zram.
> And I think it does make sense.

uhhh... ok, I see. Thanks.

	-ss