* issuing flush to all namespaces
@ 2022-07-14 8:08 Klaus Jensen
2022-07-14 8:11 ` Christoph Hellwig
0 siblings, 1 reply; 3+ messages in thread
From: Klaus Jensen @ 2022-07-14 8:08 UTC (permalink / raw)
To: linux-nvme; +Cc: Keith Busch, Christoph Hellwig, Jens Axboe, Sagi Grimberg
[-- Attachment #1: Type: text/plain, Size: 457 bytes --]
Hi,
Since the (very reasonable) changes that cause the driver to check that
the nsid of the command corresponds to the nsid of the namespace being
targeted it is no longer possible to issue a flush to all namespaces:
# nvme flush /dev/nvme0n1 -n 0xffffffff
flush: Invalid argument
Now, I think we can all agree that Flush is the odd one out here, but
can we do anything about it? The char devices also check the nsid, so
they can't be used either.
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: issuing flush to all namespaces
2022-07-14 8:08 issuing flush to all namespaces Klaus Jensen
@ 2022-07-14 8:11 ` Christoph Hellwig
2022-07-14 8:16 ` Klaus Jensen
0 siblings, 1 reply; 3+ messages in thread
From: Christoph Hellwig @ 2022-07-14 8:11 UTC (permalink / raw)
To: Klaus Jensen
Cc: linux-nvme, Keith Busch, Christoph Hellwig, Jens Axboe,
Sagi Grimberg
On Thu, Jul 14, 2022 at 10:08:41AM +0200, Klaus Jensen wrote:
> Hi,
>
> Since the (very reasonable) changes that cause the driver to check that
> the nsid of the command corresponds to the nsid of the namespace being
> targeted it is no longer possible to issue a flush to all namespaces:
>
> # nvme flush /dev/nvme0n1 -n 0xffffffff
> flush: Invalid argument
>
> Now, I think we can all agree that Flush is the odd one out here, but
> can we do anything about it? The char devices also check the nsid, so
> they can't be used either.
Don't do it. Sending flush to all namespaces wasn't even clearly
defined until I think NVMe 1.4 with the supported bits. It doesn't
really make much sense and there is no point in supporting it in Linux.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: issuing flush to all namespaces
2022-07-14 8:11 ` Christoph Hellwig
@ 2022-07-14 8:16 ` Klaus Jensen
0 siblings, 0 replies; 3+ messages in thread
From: Klaus Jensen @ 2022-07-14 8:16 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: linux-nvme, Keith Busch, Jens Axboe, Sagi Grimberg
[-- Attachment #1: Type: text/plain, Size: 878 bytes --]
On Jul 14 10:11, Christoph Hellwig wrote:
> On Thu, Jul 14, 2022 at 10:08:41AM +0200, Klaus Jensen wrote:
> > Hi,
> >
> > Since the (very reasonable) changes that cause the driver to check that
> > the nsid of the command corresponds to the nsid of the namespace being
> > targeted it is no longer possible to issue a flush to all namespaces:
> >
> > # nvme flush /dev/nvme0n1 -n 0xffffffff
> > flush: Invalid argument
> >
> > Now, I think we can all agree that Flush is the odd one out here, but
> > can we do anything about it? The char devices also check the nsid, so
> > they can't be used either.
>
> Don't do it. Sending flush to all namespaces wasn't even clearly
> defined until I think NVMe 1.4 with the supported bits. It doesn't
> really make much sense and there is no point in supporting it in Linux.
Alright, I hear you :) Thanks!
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-07-14 8:16 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-07-14 8:08 issuing flush to all namespaces Klaus Jensen
2022-07-14 8:11 ` Christoph Hellwig
2022-07-14 8:16 ` Klaus Jensen
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.