Re: [PATCH 1/2] fs: Do not dispatch FITRIM through separate super_operation

[Greg Freemyer <greg.freemyer@gmail.com>]

On Thu, Nov 25, 2010 at 9:53 AM, Mark Lord <kernel@teksavvy.com> wrote:
> On 10-11-24 11:41 PM, Greg Freemyer wrote:
>>
>> If you see a significant speed difference between mke2fs and running
>> wiper.sh on that same filesystem immediately after formatting, then
>> their is likely a bug worth chasing.
>
>
> You're right.  They're both equally slow on a blank Sandforce drive,
> taking about 83-90 seconds to TRIM 100GB.
>
> The difference being that mke2fs just sits there with zero indication
> or information as to what the huge delay is.
>
> It's a pity the TRIM commands are so sluggish on the Sandforce.
> It's a generation newer than the original Vertex, and TRIM is way faster
> on the older Vertex.  Not that either of them do it quickly enough that
> I'd ever want the kernel doing it automatically "on my behalf".
>
> The strange thing, is that when we TRIM the entire raw device,
> using ATA SECURITY ERASE, it takes less than 2 seconds.
>
> Cheers

Mark,

I'm away from my systems today, but is there an easy way to tweak
wiper.sh or hdparm to cause only a single range per trim command.

I'm curious if for the Sandforce that would still be in the 90 second
range, or closer to 64x that.

Or if there is a readily available userspace tool to invoke FITRIM and
you have a 2.6.37-rc? kernel, you could just time that on a newly
formatted drive.

Greg