On (03/04/15 17:19), Sergey Senozhatsky wrote:
> hm... that actually doesn't look so hard.
>
> all I need to do is just change zram_add attr to RW (or create a new RO
> attribute `zram_auto_add', or whatever) and treat read access to that
> zram_add_read() or zram_auto_add_read() attr as (atomic operation):
> -- generate new device_id X
> -- create corresponding zramX device
> -- return that X as char *text to user space (or error). the
> same way we sprintf() stats.
>
> what do you think? I'll play with it.
>
that will do the trick. testing simple patch:
# cat /sys/class/zram-control/zram_add
1
# cat /sys/class/zram-control/zram_add
2
# cat /sys/class/zram-control/zram_add
3
# cat /sys/class/zram-control/zram_add
4
# cat /sys/class/zram-control/zram_add
5
dmesg:
[14339.063075] zram: Added device: zram1
[14339.972633] zram: Added device: zram2
[14340.626564] zram: Added device: zram3
[14341.242134] zram: Added device: zram4
[14341.850853] zram: Added device: zram5
try to add already existing device_id
# echo 3 > /sys/class/zram-control/zram_add
-bash: echo: write error: File exists
zram_add() error propogation to user space (-ENOMEM in this case)
# cat /sys/class/zram-control/zram_add
cat: /sys/class/zram-control/zram_add: Cannot allocate memory
will send out new patch set later today.
-ss
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
