* A NVME command always generates some additional commands?
@ 2014-06-02 19:32 Angelo Brito
2014-06-02 20:08 ` Keith Busch
0 siblings, 1 reply; 3+ messages in thread
From: Angelo Brito @ 2014-06-02 19:32 UTC (permalink / raw)
Hello guys,
We are developing a NVMe device and your device is under tests right
now. So we are monitoring what is requested, what is executed and what
are the answers. Then every time we request a NVMe I/O Command some
additional 23 I/O Read Commands are generated to read some blocks next
to the end of our memory or next to the start. Is it an expected
behavior? I didn't find a reason at the driver for these commands to
be generated.
I'm using Red Hat Enterprise Linux with kernel 3.10.0-rc5linux and the
default driver version.
What do you think? Any suggestions of how to shut these commands down?
Regards,
Angelo Silva Brito.
http://about.me/angelobrito
_________________________________________________
^ permalink raw reply [flat|nested] 3+ messages in thread
* A NVME command always generates some additional commands?
2014-06-02 19:32 A NVME command always generates some additional commands? Angelo Brito
@ 2014-06-02 20:08 ` Keith Busch
2014-06-02 20:51 ` Angelo Brito
0 siblings, 1 reply; 3+ messages in thread
From: Keith Busch @ 2014-06-02 20:08 UTC (permalink / raw)
On Mon, 2 Jun 2014, Angelo Brito wrote:
> We are developing a NVMe device and your device is under tests right
> now. So we are monitoring what is requested, what is executed and what
> are the answers. Then every time we request a NVMe I/O Command some
> additional 23 I/O Read Commands are generated to read some blocks next
> to the end of our memory or next to the start. Is it an expected
> behavior? I didn't find a reason at the driver for these commands to
> be generated.
Reads are generated everytime you close a block device opened for writing.
^ permalink raw reply [flat|nested] 3+ messages in thread
* A NVME command always generates some additional commands?
2014-06-02 20:08 ` Keith Busch
@ 2014-06-02 20:51 ` Angelo Brito
0 siblings, 0 replies; 3+ messages in thread
From: Angelo Brito @ 2014-06-02 20:51 UTC (permalink / raw)
Thanks Keith!
I just needed a confirmation that it was an expected behavior!
Regards,
Angelo Silva Brito.
Graduate in Engenharia da Computa??o - UFPE
http://about.me/angelobrito
_________________________________________________
On Mon, Jun 2, 2014@5:08 PM, Keith Busch <keith.busch@intel.com> wrote:
> On Mon, 2 Jun 2014, Angelo Brito wrote:
>>
>> We are developing a NVMe device and your device is under tests right
>> now. So we are monitoring what is requested, what is executed and what
>> are the answers. Then every time we request a NVMe I/O Command some
>> additional 23 I/O Read Commands are generated to read some blocks next
>> to the end of our memory or next to the start. Is it an expected
>> behavior? I didn't find a reason at the driver for these commands to
>> be generated.
>
>
> Reads are generated everytime you close a block device opened for writing.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2014-06-02 20:51 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-06-02 19:32 A NVME command always generates some additional commands? Angelo Brito
2014-06-02 20:08 ` Keith Busch
2014-06-02 20:51 ` Angelo Brito
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox