netdev - RE: [PATCH v4 2/2] PTP: add support for one-shot output

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: <B79D786B7111A34A8CF09F833429C493BCA528D5@ORSMSX109.amr.corp.intel.com>
Date:   Tue, 24 Sep 2019 20:23:30 +0000
From:   "Hall, Christopher S" <christopher.s.hall@...el.com>
To:     "Keller, Jacob E" <jacob.e.keller@...el.com>,
        Felipe Balbi <felipe.balbi@...ux.intel.com>,
        Richard Cochran <richardcochran@...il.com>
CC:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH v4 2/2] PTP: add support for one-shot output

> -----Original Message-----
> From: Keller, Jacob E
> Sent: Tuesday, September 24, 2019 12:23 PM
> To: Felipe Balbi <felipe.balbi@...ux.intel.com>; Richard Cochran
> <richardcochran@...il.com>
> Cc: Hall, Christopher S <christopher.s.hall@...el.com>;
> netdev@...r.kernel.org; linux-kernel@...r.kernel.org
> Subject: RE: [PATCH v4 2/2] PTP: add support for one-shot output
> 
> 
> 
> > -----Original Message-----
> > From: netdev-owner@...r.kernel.org [mailto:netdev-owner@...r.kernel.org]
> On
> > Behalf Of Felipe Balbi
> > Sent: Tuesday, September 10, 2019 11:16 PM
> > To: Richard Cochran <richardcochran@...il.com>
> > Cc: Hall, Christopher S <christopher.s.hall@...el.com>;
> netdev@...r.kernel.org;
> > linux-kernel@...r.kernel.org; Felipe Balbi
> <felipe.balbi@...ux.intel.com>
> > Subject: [PATCH v4 2/2] PTP: add support for one-shot output
> >
> > Some controllers allow for a one-shot output pulse, in contrast to
> > periodic output. Now that we have extensible versions of our IOCTLs, we
> > can finally make use of the 'flags' field to pass a bit telling driver
> > that if we want one-shot pulse output.
> >
> > Signed-off-by: Felipe Balbi <felipe.balbi@...ux.intel.com>
> > ---
> >
> > Changes since v3:
> > 	- Remove bogus bitwise negation
> >
> > Changes since v2:
> > 	- Add _PEROUT_ to bit macro
> >
> > Changes since v1:
> > 	- remove comment from .flags field
> >
> >  include/uapi/linux/ptp_clock.h | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/include/uapi/linux/ptp_clock.h
> b/include/uapi/linux/ptp_clock.h
> > index 9a0af3511b68..f16301015949 100644
> > --- a/include/uapi/linux/ptp_clock.h
> > +++ b/include/uapi/linux/ptp_clock.h
> > @@ -38,8 +38,8 @@
> >  /*
> >   * Bits of the ptp_perout_request.flags field:
> >   */
> > -#define PTP_PEROUT_VALID_FLAGS (0)
> > -
> > +#define PTP_PEROUT_ONE_SHOT (1<<0)
> > +#define PTP_PEROUT_VALID_FLAGS	(PTP_PEROUT_ONE_SHOT)
> >  /*
> >   * struct ptp_clock_time - represents a time value
> >   *
> > @@ -77,7 +77,7 @@ struct ptp_perout_request {
> >  	struct ptp_clock_time start;  /* Absolute start time. */
> >  	struct ptp_clock_time period; /* Desired period, zero means disable.
> */
> >  	unsigned int index;           /* Which channel to configure. */
> > -	unsigned int flags;           /* Reserved for future use. */
> > +	unsigned int flags;
> >  	unsigned int rsv[4];          /* Reserved for future use. */
> >  };
> >
> > --
> > 2.23.0
> 
> Hi Felipe,
> 
> Do you have any examples for how you envision using this? I don't see any
> drivers or other code on the list for doing so.
> 
> Additionally, it seems weird because we do not have support for specifying
> the pulse width. I guess you leave that up to driver choice?
> 
> Thanks,
> Jake

Jake,

Good catch on the terminology. This is an API that produces edges not pulses.
This flag causes the PEROUT ioctl to ignore the period argument and produce a
single edge. Currently, the igb driver implements the same function, but uses
a "magic" invalid period specification to signal that the period argument
should be ignored (use_freq == 0):

		if (on && ((ns <= 70000000LL) || (ns == 125000000LL) ||
			   (ns == 250000000LL) || (ns == 500000000LL))) {
			if (ns < 8LL)
				return -EINVAL;
			use_freq = 1;
		}

The proposal is to support this function without magic period specifications
using an explicit flag instead. An example use case is pulse-per-second
output. While PPS is periodic, time-aware GPIO is driven by (an
unadjustable) Always Running Timer (ART). It's necessary to schedule each
edge in software to produce PPS synced with system time.

Chris