[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAC5umyhEpUAunH=eZVfihPTd_fLYFVrG4hsRRQryCUNgncDoUA@mail.gmail.com>
Date: Sun, 8 Feb 2015 13:03:34 +0900
From: Akinobu Mita <akinobu.mita@...il.com>
To: LKML <linux-kernel@...r.kernel.org>
Cc: Akinobu Mita <akinobu.mita@...il.com>,
Jens Axboe <axboe@...nel.dk>,
Andrew Morton <akpm@...ux-foundation.org>,
linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH] brd: detect zero writes for saving ram
2015-02-08 0:18 GMT+09:00 Akinobu Mita <akinobu.mita@...il.com>:
> 2015-02-08 0:03 GMT+09:00 Akinobu Mita <akinobu.mita@...il.com>:
>> This introduces a module parameter to detect zero writes and not to
>> allocate memory. Read requests for unallocated (unwritten) region
>> end up by reading zero. So this can save zeroed memory consumption
>> with extra overhead for the detection.
>>
>> This feature is useful for testing filesystems and user programs to
>> huge files without huge real storage. So this change also extends
>> the upper limit on the size of the RAM disk.
>
> I have just noticed about zram and I'll check I can use it instead.
zram currently needs to allocate memory proportional to the size of
the disk ((disksize >> PAGE_SHIFT) * sizeof(struct zram_table_entry)).
If disksize is TB or PB class, it requires huge amount of memory.
But brd with this patch can create 1PB filesystem. So there is a
reason for this new brd feature to exist.
# modprobe brd zero_detect=1 rd_size=$((1024*1024*1024*1024))
# mkfs.xfs /dev/ram0
# mount /dev/ram0 /mnt
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists