Prev: sched/cpu-timers: task->signal is stable now
Next: perf/tracing: Fix regression of perf losing kprobe events
From: Kyungmin Park on 11 Jun 2010 04:00
On Thu, Jun 10, 2010 at 5:19 PM, Christoph Hellwig <hch(a)infradead.org> wrote:
> On Fri, Jun 04, 2010 at 12:13:16AM +0300, Adrian Hunter wrote:
>> SD/MMC cards provide an erase operation that is too inefficient to
>> allow file systems to use it, however it is still useful for
>> userspace tools because it is still 10 - 100 times faster than
>> writing zeroes. �Allow the MMC block driver to provide its own
>> BLKDISCARD implementation for this purpose.

I heard if we give trim command without proper align, it writes some
data to align its data.
then user data can be corrupt.
The main problem is we don't know exact used block at filesystem. so
user can corrupt their data easily.

And as you know each manufacturers use their own superpage and align.

>
> That's no good reason. �Almost all current SSD TRIM implementations
> have the same issue, that's why filesustems require the -o discard
> option to enabled it. �In addition I've not actually seen any numbers
> from you on a current kernel.

I hope to see the performance results. In my test, no improvement.

Thank you,
Kyungmin Park

>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo(a)vger.kernel.org
> More majordomo info at �http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: sched/cpu-timers: task->signal is stable now
Next: perf/tracing: Fix regression of perf losing kprobe events