Hi Mark, thanks for your reply and your comments. I will make sure to follow your suggestions next time I try to submit a patch to the mailing list. It doesn't make sense to try again with this patch, because the project in which context it was created is already closed.
Regards, Daniel Daniel Sobe Engineer, BL Car NXP Semiconductors Germany GmbH Am Waldschlösschen 1, Dresden, Germany Tel: +49 351 32023 504 E-Mail: [email protected], www.nxp.com Geschäftsführung: Ruediger Stroh (Vors.), Frans Scheper, Kurt Sievers, Erik Just-Wartiainen / Aufsichtsratsvorsitzender: Gernot Fiedler / Sitz: Hamburg / Registergericht: Hamburg HRB 84 865 The information contained in this message is confidential and may be legally privileged. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, dissemination, or reproduction is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. Unless otherwise recorded in a written agreement, all sales transactions by NXP Semiconductors are subject to our general terms and conditions of commercial sale. These are published at: www.nxp.com/profile/terms/index.html ________________________________________ From: Mark Brown <[email protected]> Sent: Saturday, August 16, 2014 5:13 PM To: Daniel Sobe Cc: [email protected] Subject: Re: [PATCH] (re-send) spi-imx: improve timing of bursts On Fri, Mar 28, 2014 at 11:05:49AM +0000, Daniel Sobe wrote: > > Hi, > > The patch below addresses 2 improvements: Please follow the patch submission process covered in SubmittingPatches - in particular you need to send patches to the relevant developers as well as the mailing lists, you need to format your changelogs in the standard Linux style (including word wrapping within paragraphs) and you need to follow the kernel coding style. If you only send your patch to the list there's a good chance it'll be missed. > 1) Burstsize of a transfer adjusted up to the maximum burstsize of the > hardware --> no unwanted CS toggling between data words > 2) Variable trigger level allows to increase priority for data transfer, > enables continuous transfers up to the burstsize. --> remove gaps every 64 > words > The original behaviour of the driver is untouched. The additional features > can be enabled via sysfs. Why would you ever want /CS toggling between words - such behaviour is just plain buggy? > +// (sysfs-)tunable attributes > +static int allow_big_bursts = 0; // if set to 1 this enables 32-bit big > bursts > +static int tx_fifo_thres = -1; // if positive, use TX data request > interrupt when applicable These would need to be per device rather than global. It's also not clear to me that we don't want to enable some level of FIFO threashold usage by default, perhaps aiming for something just slightly below the full FIFO so that we minimise the number of extra interrupts but increase throughput on lightly loaded systems - I'm not even sure that making this behaviour tunable is needed. -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
