Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 20 Jan 2023 16:29:19 +0800
From:   Herbert Xu <herbert@gondor.apana.org.au>
To:     Giovanni Cabiddu <giovanni.cabiddu@intel.com>
Cc:     linux-crypto@vger.kernel.org, giovanni.cabiddu@intel.com
Subject: Re: [RFC] crypto: qat - enable polling for compression
Message-ID: <Y8pQ38EkAK/DVTJ2@gondor.apana.org.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230118171411.8088-1-giovanni.cabiddu@intel.com>
Precedence: bulk

Giovanni Cabiddu <giovanni.cabiddu@intel.com> wrote:
> When a request is synchronous, it is more efficient to submit it and
> poll for a response without going through the interrupt path.

Can you quantify how polling is more efficient? Does it actually make
an observable difference?

> This patch adds logic in the transport layer to poll the response ring
> and enables polling for compression in the QAT driver.
> 
> This is an initial and not complete implementation. The reason why it
> has been sent as RFC is to discuss about ways to mark a request as
> synchronous from the acomp APIs.

What existing (or future) users would benefit from this?

You could poll based on a request flag, e.g., the existing MAY_SLEEP
flag.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt