Re: [PATCH v1 2/2] zram: remove init_lock in zram_make_request

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

On (01/28/15 17:15), Minchan Kim wrote:
> Admin could reset zram during I/O operation going on so we have
> used zram->init_lock as read-side lock in I/O path to prevent
> sudden zram meta freeing.
> 
> However, the init_lock is really troublesome.
> We can't do call zram_meta_alloc under init_lock due to lockdep splat
> because zram_rw_page is one of the function under reclaim path and
> hold it as read_lock while other places in process context hold it
> as write_lock. So, we have used allocation out of the lock to avoid
> lockdep warn but it's not good for readability and fainally, I met
> another lockdep splat between init_lock and cpu_hotpulug from
> kmem_cache_destroy during wokring zsmalloc compaction. :(
> 
> Yes, the ideal is to remove horrible init_lock of zram in rw path.
> This patch removes it in rw path and instead, put init_done bool
> variable to check initialization done with smp_[wmb|rmb] and
> srcu_[un]read_lock to prevent sudden zram meta freeing
> during I/O operation.
> 
> Signed-off-by: Minchan Kim <minchan@kernel.org>
> ---
>  drivers/block/zram/zram_drv.c | 85 +++++++++++++++++++++++++++++++------------
>  drivers/block/zram/zram_drv.h |  5 +++
>  2 files changed, 66 insertions(+), 24 deletions(-)
> 
> diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> index a598ada817f0..b33add453027 100644
> --- a/drivers/block/zram/zram_drv.c
> +++ b/drivers/block/zram/zram_drv.c
> @@ -32,6 +32,7 @@
>  #include <linux/string.h>
>  #include <linux/vmalloc.h>
>  #include <linux/err.h>
> +#include <linux/srcu.h>
>  
>  #include "zram_drv.h"
>  
> @@ -53,9 +54,31 @@ static ssize_t name##_show(struct device *d,		\
>  }									\
>  static DEVICE_ATTR_RO(name);
>  
> -static inline int init_done(struct zram *zram)
> +static inline bool init_done(struct zram *zram)
>  {
> -	return zram->meta != NULL;
> +	/*
> +	 * init_done can be used without holding zram->init_lock in
> +	 * read/write handler(ie, zram_make_request) but we should make sure
> +	 * that zram->init_done should set up after meta initialization is
> +	 * done. Look at setup_init_done.
> +	 */
> +	bool ret = zram->init_done;

I don't like re-introduced ->init_done.
another idea... how about using `zram->disksize == 0' instead of
`->init_done' (previously `->meta != NULL')? should do the trick.


and I'm not sure I get this rmb...

	-ss