On Thu, 14 Apr 2022 15:11:46 +0000
Jeff Daly <je...@silicom-usa.com> wrote:
> > -----Original Message-----
> > From: Thomas Monjalon <tho...@monjalon.net>
> > Sent: Thursday, April 14, 2022 8:19 AM
> > To: Wang, Haiyue <haiyue.w...@intel.com>
> > Cc: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org; Stephen Douthit
> > <steph...@silicom-usa.com>; qi.z.zh...@intel.com;
> > john.mcnam...@intel.com
> > Subject: Re: [PATCH] net/ixgbe: Treat 1G Cu SFPs as 1G SX on the X550
> > devices
> >
> > Caution: This is an external email. Please take care when clicking links or
> > opening attachments.
> >
> >
> > 14/04/2022 14:13, Wang, Haiyue:
> > > From: Thomas Monjalon <tho...@monjalon.net>
> > > > 14/04/2022 03:31, Wang, Haiyue:
> > > > > From: je...@silicom-usa.com <je...@silicom-usa.com>
> > > > > > From: Stephen Douthit <steph...@silicom-usa.com>
> > > > > >
> > > > > > 1G Cu SFPs are not officially supported on the X552/X553 family
> > > > > > of devices but treat them as 1G SX modules since they usually
> > > > > > work. Print a warning though since support isn't validated,
> > > > > > similar to what already happens for other unofficially supported
> > > > > > SFPs enabled via the allow_unsupported_sfps parameter inherited
> > from the mainline Linux driver.
> > > > > >
> > > > > > Signed-off-by: Stephen Douthit <steph...@silicom-usa.com>
> > > > > > Signed-off-by: Jeff Daly <je...@silicom-usa.com>
> > > > > > ---
> > > > > > drivers/net/ixgbe/base/ixgbe_x550.c | 14 +++++++++++++-
> > > > > > 1 file changed, 13 insertions(+), 1 deletion(-)
> > > > > >
> > > > > > diff --git a/drivers/net/ixgbe/base/ixgbe_x550.c
> > > > > > b/drivers/net/ixgbe/base/ixgbe_x550.c
> > > > > > index 8810d1658e..8d1bc6c80d 100644
> > > > > > --- a/drivers/net/ixgbe/base/ixgbe_x550.c
> > > > > > +++ b/drivers/net/ixgbe/base/ixgbe_x550.c
> > > > > > @@ -1538,9 +1538,21 @@ STATIC s32
> > > > > > ixgbe_supported_sfp_modules_X550em(struct ixgbe_hw *hw, bool
> > > > *linear)
> > > > >
> > > > > NACK.
> > > > >
> > > > > As for 1G Cu SFP treating it as 1G SX, some 1G-Base-T SFP modules
> > > > > require the use of RX_ILOS and some Intel Ethernet products don't
> > support that.
> > > >
> > > > So what is the solution?
> > > >
> > > > > And the DPDK keeps the same design with kernel.
> > > >
> > > > It should not be a justification for limiting DPDK features.
> > >
> > > Um, this is upstream version driver to keep the same behavior.
> > >
> > > There are also some kind of custom release ...
> >
> > I don't understand.
> > Upstream DPDK (and Linux) must support a maximum of hardware and
> > setup.
> > Why rejecting adding such compatibility?
> >
>
> so, I will ask a question directly in case people just aren't inclined to
> make a suggestion
> (and perhaps this should be also directed to the Linux kernel driver mailing
> list), but
> if there's a driver option: module_param(allow_unsupported_sfp, uint, 0) to
> allow
> enabling non-official support of some SFPs, then I can't image that it
> wouldn't also be
> acceptable to add: module_param(cu_sfp_as sx, uint, 0) to be able to select
> whether
> to enable this specific handling as well?
>
> if a patch of this nature is acceptable to Linux driver maintainers, then it
> would also be
> here as well according to your explanation of the NACK, correct?
Makes sense for DPDK to have a similar option to enable (at your own risk)
SFP's.
But:
- there is no equivalent of module_params in DPDK; you will have to think of
something
- should print message that when enabled the driver is no longer supported.
use at your own risk.
