linux-kernel - Re: [PATCH 07/16] spi: bcm63xx-hsspi: Add polling mode support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <dee51504-ccd2-3c27-ffc8-2e392f2bd781@broadcom.com>
Date:   Wed, 11 Jan 2023 12:13:43 -0800
From:   William Zhang <william.zhang@...adcom.com>
To:     Mark Brown <broonie@...nel.org>
Cc:     Linux SPI List <linux-spi@...r.kernel.org>,
        Broadcom Kernel List <bcm-kernel-feedback-list@...adcom.com>,
        anand.gore@...adcom.com, tomer.yacoby@...adcom.com,
        dan.beygelman@...adcom.com, joel.peshkin@...adcom.com,
        f.fainelli@...il.com, jonas.gorski@...il.com,
        kursad.oney@...adcom.com, dregan@...l.com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 07/16] spi: bcm63xx-hsspi: Add polling mode support



On 01/10/2023 02:49 PM, Mark Brown wrote:
> On Mon, Jan 09, 2023 at 12:10:30PM -0800, William Zhang wrote:
>> On 01/09/2023 11:06 AM, Mark Brown wrote:
> 
>>> You can put whatever logic is needed in the code - for something like
>>> this an architecture based define isn't ideal but is probably good
>>> enough if need be (though I'd not be surprised if it turned out that
>>> there was also some performance benefit for the MIPS systems too, at
>>> least for smaller transfers).
> 
>> I just don't know what other logic I can put in the driver to select
>> interrupt or polling mode.  Only the end user know if performance or cpu
>> usage is more important to their application.
> 
> Usually you can take a reasonable guess as to what would be a good point
> to start switching, typically for short enough transfers the overhead of
> setting up DMA, waiting for interrupts and tearing things down is very
> much larger than the cost of just doing PIO - a bunch of other drivers
> have pick a number logic of some kind, often things like FIFO sizes are
> a good key for where to look.  A lot of the time this is good enough,
> and it means that users have much better facilities for making tradeoffs
> if they have a range of transfer sizes available - it's not an either/or
> thing but based on some features of the individual message/transfer.
> 
> It is true that for people with heavy SPI traffic or otherwise very
> demanding requirements for a specific system and software stack
> additional tuning might produce better results, exposing some sysfs
> knobs to allow tuning of parameters at runtime would be helpful for them
> and I'd certainly be happy to see that added.
> 
Thanks for the explanation. I saw the spi-uniphier.c and spi-bcm2835.c 
doing the trick you mentioned(thanks Kursad for pointing out).  In our 
case, even the maximum fifo size usage(512bytes), the polling still have 
better performance than interrupt. The MTD test result included in this 
patch is based on maximum fifo usage. So there is no benefit to switch 
to interrupt based on transfer size.

Does the spi framework has any requirement on how much time that the 
driver's transfer_one function can spend?  I can see the polling 
function might take quite some time in busy loop if the clock is low, 
for example, at 100Hz(slowest clock this controller can go), it takes 
512x8/100Hz ~= 41ms to complete.  If this is something in concern,  I 
can do the interrupt switch based on a time limit value if interrupt is 
available.

Download attachment "smime.p7s" of type "application/pkcs7-signature" (4212 bytes)