lkml.org 
[lkml]   [2020]   [Oct]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRE: [PATCH net] ravb: Fix bit fields checking in ravb_hwtstamp_get()
Date
Hello Sergei,

> -----Original Message-----
> From: Sergei Shtylyov [mailto:sergei.shtylyov@gmail.com]
> Sent: Saturday, October 17, 2020 10:49 PM
> To: Gabbasov, Andrew <Andrew_Gabbasov@mentor.com>
> Cc: linux-renesas-soc@vger.kernel.org; netdev@vger.kernel.org; linux-kernel@vger.kernel.org; David S. Miller
> <davem@davemloft.net>; geert+renesas@glider.be; Julia Lawall <julia.lawall@inria.fr>; Behme, Dirk - Bosch
> <dirk.behme@de.bosch.com>; Eugeniu Rosca <erosca@de.adit-jv.com>
> Subject: Re: [PATCH net] ravb: Fix bit fields checking in ravb_hwtstamp_get()
>
> Hello!
>
> On 10/1/20 10:13 AM, Andrew Gabbasov wrote:
>
> The patch was set to the "Changes Requested" state -- most probably because of this
> mail. Though unintentionally, it served to throttle actions on this patch. I did only
> remember about this patch yesterday... :-)
>
> [...]
> >> In the function ravb_hwtstamp_get() in ravb_main.c with the existing
> > values
> >> for RAVB_RXTSTAMP_TYPE_V2_L2_EVENT (0x2) and RAVB_RXTSTAMP_TYPE_ALL
> >> (0x6)
> >>
> >> if (priv->tstamp_rx_ctrl & RAVB_RXTSTAMP_TYPE_V2_L2_EVENT)
> >> config.rx_filter = HWTSTAMP_FILTER_PTP_V2_L2_EVENT;
> >> else if (priv->tstamp_rx_ctrl & RAVB_RXTSTAMP_TYPE_ALL)
> >> config.rx_filter = HWTSTAMP_FILTER_ALL;
> >>
> >> if the test on RAVB_RXTSTAMP_TYPE_ALL should be true, it will never be
> >> reached.
> >>
> >> This issue can be verified with 'hwtstamp_config' testing program
> >> (tools/testing/selftests/net/hwtstamp_config.c). Setting filter type to
> > ALL
> >> and subsequent retrieving it gives incorrect value:
> >>
> >> $ hwtstamp_config eth0 OFF ALL
> >> flags = 0
> >> tx_type = OFF
> >> rx_filter = ALL
> >> $ hwtstamp_config eth0
> >> flags = 0
> >> tx_type = OFF
> >> rx_filter = PTP_V2_L2_EVENT
> >>
> >> Correct this by converting if-else's to switch.
> >
> > Earlier you proposed to fix this issue by changing the value
> > of RAVB_RXTSTAMP_TYPE_ALL constant to 0x4.
> > Unfortunately, simple changing of the constant value will not
> > be enough, since the code in ravb_rx() (actually determining
> > if timestamp is needed)
> >
> > u32 get_ts = priv->tstamp_rx_ctrl & RAVB_RXTSTAMP_TYPE;
> > [...]
> > get_ts &= (q == RAVB_NC) ?
> > RAVB_RXTSTAMP_TYPE_V2_L2_EVENT :
> > ~RAVB_RXTSTAMP_TYPE_V2_L2_EVENT;
> >
> > will work incorrectly and will need to be fixed too, making this
> > piece of code more complicated.
> >
> > So, it's probably easier and safer to keep the constant value and
> > the code in ravb_rx() intact, and just fix the get ioctl code,
> > where the issue is actually located.
>
> We have one more issue with the current driver: bit 2 of priv->tstamp_rx_ctrl
> can only be set as a part of the ALL mask, not individually. I'm now thinking we
> should set RAVB_RXTSTAMP_TYPE[_ALL] to 2 (and probably just drop the ALL mask)...

[skipped]

> Yeah, that's better. But do we really need am anonymous bit 2 that can't be
> toggled other than via passing the ALL mask?

The driver supports setting timestamps either for all packets or for some
particular kind of packets (events). Bit 1 in internal mask corresponds
to this selected kind. Bit 2 corresponds to all other packets, and ALL mask
combines both variants. Although bit 2 can't be controlled individually
(since there is no much sense to Request stamping of only packets, other than
events, moreover, there is no user-visible filter constant to represent it),
and that's why is anonymous, it provides a convenient way to handle stamping
logic in ravb_rx(), so I don't see an immediate need to get rid of it.

Thanks.

Best regards,
Andrew

\
 
 \ /
  Last update: 2020-10-19 09:41    [W:0.079 / U:0.492 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site