> -----Original Message-----
> From: Jeff Daly <je...@silicom-usa.com>
> Sent: Friday, April 22, 2022 01:31
> To: Wang, Haiyue <haiyue.w...@intel.com>; dev@dpdk.org
> Cc: sta...@dpdk.org; Yang, Qiming <qiming.y...@intel.com>
> Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and linking on
> hotplug
>
>
>
> > -----Original Message-----
> > From: Wang, Haiyue <haiyue.w...@intel.com>
> > Sent: Tuesday, April 19, 2022 9:09 PM
> > To: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org
> > Cc: sta...@dpdk.org; Yang, Qiming <qiming.y...@intel.com>
> > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and linking on
> > hotplug
> >
> > Caution: This is an external email. Please take care when clicking links or
> > opening attachments.
> >
> >
> > > -----Original Message-----
> > > From: Jeff Daly <je...@silicom-usa.com>
> > > Sent: Wednesday, April 20, 2022 01:34
> > > To: Wang, Haiyue <haiyue.w...@intel.com>; dev@dpdk.org
> > > Cc: sta...@dpdk.org; Yang, Qiming <qiming.y...@intel.com>
> > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and linking
> > > on hotplug
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Wang, Haiyue <haiyue.w...@intel.com>
> > > > Sent: Monday, April 18, 2022 10:05 PM
> > > > To: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org
> > > > Cc: sta...@dpdk.org; Yang, Qiming <qiming.y...@intel.com>
> > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and linking
> > > > on hotplug
> > > >
> > > > Caution: This is an external email. Please take care when clicking
> > > > links or opening attachments.
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: Jeff Daly <je...@silicom-usa.com>
> > > > > Sent: Tuesday, April 19, 2022 05:55
> > > > > To: Wang, Haiyue <haiyue.w...@intel.com>; dev@dpdk.org
> > > > > Cc: sta...@dpdk.org
> > > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and
> > > > > linking on hotplug
> > > > >
> > > > >
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Wang, Haiyue <haiyue.w...@intel.com>
> > > > > > Sent: Thursday, April 14, 2022 8:11 AM
> > > > > > To: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org
> > > > > > Cc: sta...@dpdk.org
> > > > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and
> > > > > > linking on hotplug
> > > > > >
> > > > > > Caution: This is an external email. Please take care when
> > > > > > clicking links or opening attachments.
> > > > > >
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Jeff Daly <je...@silicom-usa.com>
> > > > > > > Sent: Thursday, April 14, 2022 18:41
> > > > > > > To: Wang, Haiyue <haiyue.w...@intel.com>; dev@dpdk.org
> > > > > > > Cc: sta...@dpdk.org
> > > > > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and
> > > > > > > linking on hotplug
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: Wang, Haiyue <haiyue.w...@intel.com>
> > > > > > > > Sent: Wednesday, April 13, 2022 11:00 PM
> > > > > > > > To: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org
> > > > > > > > Cc: sta...@dpdk.org; Stephen Douthit
> > > > > > > > <steph...@silicom-usa.com>
> > > > > > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and
> > > > > > > > linking on hotplug
> > > > > > > >
> > > > > > > > Caution: This is an external email. Please take care when
> > > > > > > > clicking links or opening attachments.
> > > > > > > >
> > > > > > > >
> > > > > > > > > -----Original Message-----
> > > > > > > > > From: Wang, Haiyue
> > > > > > > > > Sent: Thursday, April 14, 2022 10:49
> > > > > > > > > To: Jeff Daly <je...@silicom-usa.com>; dev@dpdk.org
> > > > > > > > > Cc: sta...@dpdk.org; Stephen Douthit
> > > > > > > > > <steph...@silicom-usa.com>
> > > > > > > > > Subject: RE: [PATCH v6 2/2] net/ixgbe: Fix SFP detection
> > > > > > > > > and linking on hotplug
> > > > > > > > >
> > > > > > > > > > -----Original Message-----
> > > > > > > > > > From: Jeff Daly <je...@silicom-usa.com>
> > > > > > > > > > Sent: Wednesday, April 13, 2022 01:42
> > > > > > > > > > To: dev@dpdk.org
> > > > > > > > > > Cc: sta...@dpdk.org; Stephen Douthit
> > > > > > > > > > <steph...@silicom-usa.com>; Wang, Haiyue
> > > > > > > > > > <haiyue.w...@intel.com>
> > > > > > > > > > Subject: [PATCH v6 2/2] net/ixgbe: Fix SFP detection and
> > > > > > > > > > linking on hotplug
> > > > > > > > > >
> > > > > > > > > > Currently the ixgbe driver does not ID any SFP except
> > > > > > > > > > for the first one plugged in. This can lead to no-link,
> > > > > > > > > > or incorrect speed
> > > > > > conditions.
> > > > > > > > > >
> > > > > > > > > > For example:
> > > > > > > > > >
> > > > > > > > > > * If link is initially established with a 1G SFP, and
> > > > > > > > > > later a 1G/10G multispeed part is later installed, then
> > > > > > > > > > the MAC link setup functions are never called to change
> > > > > > > > > > from 1000BASE-X to 10GBASE-R mode, and the link stays
> > > > > > > > > > running at the
> > > > slower rate.
> > > > > > > > > >
> > > > > > > > > > * If link is initially established with a 1G SFP, and
> > > > > > > > > > later a 10G only module is later installed, no link is
> > > > > > > > > > established, since we are still trasnsmitting in
> > > > > > > > > > 1000BASE-X mode to a 10GBASE-R
> > > > > > only partner.
> > > > > > > > > >
> > > > > > > > > > Refactor the SFP ID/setup, and link setup code, to more
> > > > > > > > > > closely match the flow of the mainline kernel driver
> > > > > > > > > > which does not have these issues. In that driver a
> > > > > > > > > > service task runs periodically to handle these
> > > > > > > > > > operations based on bit flags that have been set
> > > > > > > > > > (usually via interrupt or userspace request), and then
> > > > > > > > > > get cleared once the requested subtask has
> > > > been completed.
> > > > > > > > > >
> > > > > > > > > > Fixes: af75078fece ("first public release")
> > > > > > > > > > Cc: sta...@dpdk.org
> > > > > > > > > >
> > > > > > > > > > Signed-off-by: Stephen Douthit
> > > > > > > > > > <steph...@silicom-usa.com>
> > > > > > > > > > Signed-off-by: Jeff Daly <je...@silicom-usa.com>
> > > > > > > > > > ---
> > > > > > > > > > drivers/net/ixgbe/ixgbe_ethdev.c | 533
> > > > > > > > > > +++++++++++++++++++++++--------
> > > > > > > > > > +++++++++++++++++++++++drivers/net/ixgbe/ixgbe_ethdev.h
> > > > > > > > > > +++++++++++++++++++++++|
> > > > > > > > > > 14 +-
> > > > > > > > > > 2 files changed, 410 insertions(+), 137 deletions(-)
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > struct ixgbe_stat_mapping_registers { @@ -510,7 +509,7
> > > > > > > > > > @@ struct ixgbe_adapter {
> > > > > > > > > > uint8_t pflink_fullchk;
> > > > > > > > > > uint8_t mac_ctrl_frame_fwd;
> > > > > > > > > > rte_atomic32_t link_thread_running;
> > > > > > > > > > - pthread_t link_thread_tid;
> > > > > > > > > > + pthread_t service_thread_tid;
> > > > > > > > >
> > > > > > > > > No need to rename this variable,
> > > > > > > >
> > > > > > > > Let's do link related service now, so we can keep it, I
> > > > > > > > missed to add my comment. ;-)
> > > > > > > >
> > > > > > >
> > > > > > > I don't understand this reply, are you still asking to rework
> > > > > > > the patch or
> > > > > > not?
> > > > > > >
> > > > > >
> > > > > > Different thing.
> > > > > >
> > > > > > 1. This var can be kept to trace the created thread. (change
> > > > > > less code to
> > > > keep
> > > > > > the patch clean.)
> > > > > > 2. Yes, two patches.
> > > > > >
> > > > >
> > > > > ok, I guess I'm just being thick-headed here, but I still don't
> > > > > understand why you are saying it should be split into
> > > > > 2 patches. if I understand *what* you are asking, you're saying
> > > > > make the original thread periodic to continuously
> > > >
> > > > Well, ...
> > > >
> > > > Your patch merges the original 'ixgbe_setup_link' task into one,
> > > > this will make us hard to review the whole design. So what I said
> > > > is: firstly, let's change the thread to a service thread to handle
> > > > the 'ixgbe_setup_link' subtask firstly. Which is 'ixgbe_link_service'
> > > > in your whole patch.
> > > >
> > >
> > > still not 100%, are you suggesting that the original
> > > ixgbe_dev_setup_link_thread_handler()
> > > which currently is not periodic and only really calls
> > > ixgbe_setup_link() be changed to be a periodic task that essentially
> > > does what the patch's ixgbe_link_service() function does which would
> > > only be checking whether link config is needed and if so calls
> > ixgbe_setup_link() as before?
> > >
> > > if I'm following the code correctly, it ends up going down to
> > > ixgbe_check_mac_link_generic() which looks at SDP0 (in the case of
> > > needing xtalk fix) which incorrectly will set sfp_cage_full when in
> > > fact the PRESENT# signal (or MOD_ABS#) is active low. the code is
> > > extremely convoluted in it's effort to be smart so I may be missing
> > > something, but I *believe* that what we end up with is completely
> > unnecessary probing of i2c busses looking for SFPs that don't exist. even
> > when making it periodic, I don't think it's going to end up with working
> > code.
> >
> >
> > I'm lost ...
> >
> > The 'periodic' is the service thread running mode, but the subtask is ONLY
> > called when be scheduled, like:
> >
> > if (!(adapter->flags & IXGBE_FLAG_NEED_LINK_CONFIG))
> > return;
> >
>
> this makes no sense. the *only* time right now when
> IXGBE_FLAG_NEED_LINK_CONFIG is set is
> immediately before the thread is created! there's no 'hotplug' part of the
> patch vs fixing the
> 'original issue' (your words, below). the 'original issue' is that the code
> doesn't work for hotplug
> (SFP cages), so there's no breaking up the patch to fix any original issue
> first. the only issue is
> that for SFP cages, the code doesn't work.
No!
Your code is right, but looks like you still be lost in your hotplug issue.
+static void *
+ixgbe_dev_service_thread_handler(void *param)
+{
+ struct rte_eth_dev *dev = (struct rte_eth_dev *)param;
+ struct ixgbe_hw *hw =
+ IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ uint64_t start, ticks, service_ms;
+ uint32_t speed;
+ s32 err;
+ bool link_up;
+
+ while (1) {
+ ixgbe_sfp_service(dev);
+ ixgbe_link_service(dev);
^
|
+- 1st patch, just handle the
'ixgbe_link_service_thread_handler' issue.
I've no words to comment again, hope this makes thing clear.
>
> > >
> > > > Small patch is good for us to review, and try to do one thing.
> > > >
> > > > Hope this time, I can make myself clear. ;-)
> > > >
> > > > > do ixgbe_link_setup() ? I believe the problem with the setup is
> > > > > that the sfp_type is only detected once at initialization time and
> > > > > if nothing is in the cage then the code just returns
> > > > > IXGBE_SUCCESS, in which case making this task periodic is useless.
> > > > > the whole issue of hotplug is only addressed by the entire patch
> > > > > which
> > > > > 1) makes the
> > > > > task periodic, 2) changes the actions of the task to look for
> > > > > whether the cage has something in it and whether its been changed
> > > > > and needs to be configured again.
> > > > >
> > > > >
> > > > > > > > > we can separate this patch as least into two patches:
> > > > > > > >
> > > > > > > >
> > > > > > > > >
> > > > > > > > > 1st, change the thread handle
> > > > 'ixgbe_dev_setup_link_thread_handler'
> > > > > > > > > from
> > > > > > > > >
> > > > > > > > > run-once to as periodical, to handle the original issue.
> > > > > > > > >
> > > > > > > > > The name 'ixgbe_dev_setup_link_thread_handler' may be not
> > > > > > > > > suitable now, as it is a service thread.
> > > > > > > > >
> > > > > > > > > We can change it to "'ixgbe_link_service_thread_handler'"
> > > > > > > > > to reflect the change purpose.
> > > > > > > > >
> > > > > > > > > 2nd, add the SFP hotplug in this patch.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > > };
> > > > > > > > > >
> > > > > > > > > > --
> > > > > > > > > > 2.25.1
