On 10/12/2026 7:58 PM, Keith Busch wrote:
On Mon, Oct 12, 2026 at 12:17:38PM +0530, Naman Jain wrote:
The disable_irq_nosync still does non-posted transactions for msi and
msi-x. NVMe has a shortcut for msi, but there's currently nothing for
msi-x. For that case, I have a prep patch I consulted with Thomas, and
the result is below. I'm delayed sending it out while catching up from
travel, but I think this will be even more efficient.
Thanks for sharing. I'll give it a try. IIUC, after your patch goes in, I
should be using disable_irq_nowait() in above and see better performance.
Yes, exactly. It should save hundreds of CPU cycles on every switch from
irq to polling.
Understood.
Please correct me if I am wrong. I'll monitor your patch in LKML for future
versions.
I need a user for the new API. NVMe is the only one I have in mind
right now, so this new irq disable would have to be tied to the series
that does the adaptive polling.
Acked. For this series, there are couple of solutions floating around in
LKML, and we are trying to converge on the final approach. Once
finalized, if we still need to call disable_irq_sync(), we can make use
of this new API. In any case, if I happen to send the next version of
this, I can use the new APIs, creating dependency on your changes.
Regards,
Naman
