2023-06-19 08:56:46

by Petr Oros

[permalink] [raw]
Subject: [PATCH net] ice: ice_vsi_release cleanup

Since commit 6624e780a577fc ("ice: split ice_vsi_setup into smaller
functions") ice_vsi_release does things twice. There is unregister
netdev which is unregistered in ice_deinit_eth also.

It also unregisters the devlink_port twice which is also unregistered
in ice_deinit_eth(). This double deregistration is hidden because
devl_port_unregister ignores the return value of xa_erase.

[ 68.642167] Call Trace:
[ 68.650385] ice_devlink_destroy_pf_port+0xe/0x20 [ice]
[ 68.655656] ice_vsi_release+0x445/0x690 [ice]
[ 68.660147] ice_deinit+0x99/0x280 [ice]
[ 68.664117] ice_remove+0x1b6/0x5c0 [ice]

[ 171.103841] Call Trace:
[ 171.109607] ice_devlink_destroy_pf_port+0xf/0x20 [ice]
[ 171.114841] ice_remove+0x158/0x270 [ice]
[ 171.118854] pci_device_remove+0x3b/0xc0
[ 171.122779] device_release_driver_internal+0xc7/0x170
[ 171.127912] driver_detach+0x54/0x8c
[ 171.131491] bus_remove_driver+0x77/0xd1
[ 171.135406] pci_unregister_driver+0x2d/0xb0
[ 171.139670] ice_module_exit+0xc/0x55f [ice]

Fixes: 6624e780a577 ("ice: split ice_vsi_setup into smaller functions")
Signed-off-by: Petr Oros <[email protected]>
---
drivers/net/ethernet/intel/ice/ice_lib.c | 27 ------------------------
1 file changed, 27 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 11ae0e41f518a1..284a1f0bfdb545 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -3272,39 +3272,12 @@ int ice_vsi_release(struct ice_vsi *vsi)
return -ENODEV;
pf = vsi->back;

- /* do not unregister while driver is in the reset recovery pending
- * state. Since reset/rebuild happens through PF service task workqueue,
- * it's not a good idea to unregister netdev that is associated to the
- * PF that is running the work queue items currently. This is done to
- * avoid check_flush_dependency() warning on this wq
- */
- if (vsi->netdev && !ice_is_reset_in_progress(pf->state) &&
- (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state))) {
- unregister_netdev(vsi->netdev);
- clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
- }
-
- if (vsi->type == ICE_VSI_PF)
- ice_devlink_destroy_pf_port(pf);
-
if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
ice_rss_clean(vsi);

ice_vsi_close(vsi);
ice_vsi_decfg(vsi);

- if (vsi->netdev) {
- if (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state)) {
- unregister_netdev(vsi->netdev);
- clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
- }
- if (test_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state)) {
- free_netdev(vsi->netdev);
- vsi->netdev = NULL;
- clear_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state);
- }
- }
-
/* retain SW VSI data structure since it is needed to unregister and
* free VSI netdev when PF is not in reset recovery pending state,\
* for ex: during rmmod.
--
2.41.0



2023-06-19 10:41:46

by Paul Menzel

[permalink] [raw]
Subject: Re: [Intel-wired-lan] [PATCH net] ice: ice_vsi_release cleanup

Dear Petr,


Thank you for your patch.

Am 19.06.23 um 10:49 schrieb Petr Oros:
> Since commit 6624e780a577fc ("ice: split ice_vsi_setup into smaller
> functions") ice_vsi_release does things twice. There is unregister
> netdev which is unregistered in ice_deinit_eth also.
>
> It also unregisters the devlink_port twice which is also unregistered
> in ice_deinit_eth(). This double deregistration is hidden because
> devl_port_unregister ignores the return value of xa_erase.
>
> [ 68.642167] Call Trace:
> [ 68.650385] ice_devlink_destroy_pf_port+0xe/0x20 [ice]
> [ 68.655656] ice_vsi_release+0x445/0x690 [ice]
> [ 68.660147] ice_deinit+0x99/0x280 [ice]
> [ 68.664117] ice_remove+0x1b6/0x5c0 [ice]
>
> [ 171.103841] Call Trace:
> [ 171.109607] ice_devlink_destroy_pf_port+0xf/0x20 [ice]
> [ 171.114841] ice_remove+0x158/0x270 [ice]
> [ 171.118854] pci_device_remove+0x3b/0xc0
> [ 171.122779] device_release_driver_internal+0xc7/0x170
> [ 171.127912] driver_detach+0x54/0x8c
> [ 171.131491] bus_remove_driver+0x77/0xd1
> [ 171.135406] pci_unregister_driver+0x2d/0xb0
> [ 171.139670] ice_module_exit+0xc/0x55f [ice]

It’d be more specific in the commit message summary/title, and use
imperative mood. Maybe:

ice: Unregister netdev and devlink_port only once


Kind regards,

Paul


> Fixes: 6624e780a577 ("ice: split ice_vsi_setup into smaller functions")
> Signed-off-by: Petr Oros <[email protected]>
> ---
> drivers/net/ethernet/intel/ice/ice_lib.c | 27 ------------------------
> 1 file changed, 27 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
> index 11ae0e41f518a1..284a1f0bfdb545 100644
> --- a/drivers/net/ethernet/intel/ice/ice_lib.c
> +++ b/drivers/net/ethernet/intel/ice/ice_lib.c
> @@ -3272,39 +3272,12 @@ int ice_vsi_release(struct ice_vsi *vsi)
> return -ENODEV;
> pf = vsi->back;
>
> - /* do not unregister while driver is in the reset recovery pending
> - * state. Since reset/rebuild happens through PF service task workqueue,
> - * it's not a good idea to unregister netdev that is associated to the
> - * PF that is running the work queue items currently. This is done to
> - * avoid check_flush_dependency() warning on this wq
> - */
> - if (vsi->netdev && !ice_is_reset_in_progress(pf->state) &&
> - (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state))) {
> - unregister_netdev(vsi->netdev);
> - clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
> - }
> -
> - if (vsi->type == ICE_VSI_PF)
> - ice_devlink_destroy_pf_port(pf);
> -
> if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
> ice_rss_clean(vsi);
>
> ice_vsi_close(vsi);
> ice_vsi_decfg(vsi);
>
> - if (vsi->netdev) {
> - if (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state)) {
> - unregister_netdev(vsi->netdev);
> - clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
> - }
> - if (test_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state)) {
> - free_netdev(vsi->netdev);
> - vsi->netdev = NULL;
> - clear_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state);
> - }
> - }
> -
> /* retain SW VSI data structure since it is needed to unregister and
> * free VSI netdev when PF is not in reset recovery pending state,\
> * for ex: during rmmod.

2023-06-19 12:52:06

by Maciej Fijalkowski

[permalink] [raw]
Subject: Re: [Intel-wired-lan] [PATCH net] ice: ice_vsi_release cleanup

On Mon, Jun 19, 2023 at 10:49:48AM +0200, Petr Oros wrote:
> Since commit 6624e780a577fc ("ice: split ice_vsi_setup into smaller
> functions") ice_vsi_release does things twice. There is unregister
> netdev which is unregistered in ice_deinit_eth also.
>
> It also unregisters the devlink_port twice which is also unregistered
> in ice_deinit_eth(). This double deregistration is hidden because
> devl_port_unregister ignores the return value of xa_erase.
>
> [ 68.642167] Call Trace:
> [ 68.650385] ice_devlink_destroy_pf_port+0xe/0x20 [ice]
> [ 68.655656] ice_vsi_release+0x445/0x690 [ice]
> [ 68.660147] ice_deinit+0x99/0x280 [ice]
> [ 68.664117] ice_remove+0x1b6/0x5c0 [ice]
>
> [ 171.103841] Call Trace:
> [ 171.109607] ice_devlink_destroy_pf_port+0xf/0x20 [ice]
> [ 171.114841] ice_remove+0x158/0x270 [ice]
> [ 171.118854] pci_device_remove+0x3b/0xc0
> [ 171.122779] device_release_driver_internal+0xc7/0x170
> [ 171.127912] driver_detach+0x54/0x8c
> [ 171.131491] bus_remove_driver+0x77/0xd1
> [ 171.135406] pci_unregister_driver+0x2d/0xb0
> [ 171.139670] ice_module_exit+0xc/0x55f [ice]

Hi Petr,
can you tell us when in particular this call trace was observed?

CC: Michal Swiatkowski <[email protected]>

>
> Fixes: 6624e780a577 ("ice: split ice_vsi_setup into smaller functions")
> Signed-off-by: Petr Oros <[email protected]>
> ---
> drivers/net/ethernet/intel/ice/ice_lib.c | 27 ------------------------
> 1 file changed, 27 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
> index 11ae0e41f518a1..284a1f0bfdb545 100644
> --- a/drivers/net/ethernet/intel/ice/ice_lib.c
> +++ b/drivers/net/ethernet/intel/ice/ice_lib.c
> @@ -3272,39 +3272,12 @@ int ice_vsi_release(struct ice_vsi *vsi)
> return -ENODEV;
> pf = vsi->back;
>
> - /* do not unregister while driver is in the reset recovery pending
> - * state. Since reset/rebuild happens through PF service task workqueue,
> - * it's not a good idea to unregister netdev that is associated to the
> - * PF that is running the work queue items currently. This is done to
> - * avoid check_flush_dependency() warning on this wq
> - */
> - if (vsi->netdev && !ice_is_reset_in_progress(pf->state) &&
> - (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state))) {
> - unregister_netdev(vsi->netdev);
> - clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
> - }
> -
> - if (vsi->type == ICE_VSI_PF)
> - ice_devlink_destroy_pf_port(pf);
> -
> if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
> ice_rss_clean(vsi);
>
> ice_vsi_close(vsi);
> ice_vsi_decfg(vsi);
>
> - if (vsi->netdev) {
> - if (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state)) {
> - unregister_netdev(vsi->netdev);
> - clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
> - }
> - if (test_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state)) {
> - free_netdev(vsi->netdev);
> - vsi->netdev = NULL;
> - clear_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state);
> - }
> - }
> -
> /* retain SW VSI data structure since it is needed to unregister and
> * free VSI netdev when PF is not in reset recovery pending state,\
> * for ex: during rmmod.
> --
> 2.41.0
>
> _______________________________________________
> Intel-wired-lan mailing list
> [email protected]
> https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

2023-06-19 13:06:04

by Petr Oros

[permalink] [raw]
Subject: Re: [Intel-wired-lan] [PATCH net] ice: ice_vsi_release cleanup

Maciej Fijalkowski píše v Po 19. 06. 2023 v 14:38 +0200:
> On Mon, Jun 19, 2023 at 10:49:48AM +0200, Petr Oros wrote:
> > Since commit 6624e780a577fc ("ice: split ice_vsi_setup into smaller
> > functions") ice_vsi_release does things twice. There is unregister
> > netdev which is unregistered in ice_deinit_eth also.
> >
> > It also unregisters the devlink_port twice which is also
> > unregistered
> > in ice_deinit_eth(). This double deregistration is hidden because
> > devl_port_unregister ignores the return value of xa_erase.
> >
> > [   68.642167] Call Trace:
> > [   68.650385]  ice_devlink_destroy_pf_port+0xe/0x20 [ice]
> > [   68.655656]  ice_vsi_release+0x445/0x690 [ice]
> > [   68.660147]  ice_deinit+0x99/0x280 [ice]
> > [   68.664117]  ice_remove+0x1b6/0x5c0 [ice]
> >
> > [  171.103841] Call Trace:
> > [  171.109607]  ice_devlink_destroy_pf_port+0xf/0x20 [ice]
> > [  171.114841]  ice_remove+0x158/0x270 [ice]
> > [  171.118854]  pci_device_remove+0x3b/0xc0
> > [  171.122779]  device_release_driver_internal+0xc7/0x170
> > [  171.127912]  driver_detach+0x54/0x8c
> > [  171.131491]  bus_remove_driver+0x77/0xd1
> > [  171.135406]  pci_unregister_driver+0x2d/0xb0
> > [  171.139670]  ice_module_exit+0xc/0x55f [ice]
>
> Hi Petr,
> can you tell us when in particular this call trace was observed?

I found out that the devlink port is deregistered twice. This calltrace
was generated by dump_stack() placed in devlink_port_unregister -> it
is testing trace, not from production.

Regards,
Petr


>
> CC: Michal Swiatkowski <[email protected]>
>
> >
> > Fixes: 6624e780a577 ("ice: split ice_vsi_setup into smaller
> > functions")
> > Signed-off-by: Petr Oros <[email protected]>
> > ---
> >  drivers/net/ethernet/intel/ice/ice_lib.c | 27 --------------------
> > ----
> >  1 file changed, 27 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c
> > b/drivers/net/ethernet/intel/ice/ice_lib.c
> > index 11ae0e41f518a1..284a1f0bfdb545 100644
> > --- a/drivers/net/ethernet/intel/ice/ice_lib.c
> > +++ b/drivers/net/ethernet/intel/ice/ice_lib.c
> > @@ -3272,39 +3272,12 @@ int ice_vsi_release(struct ice_vsi *vsi)
> >                 return -ENODEV;
> >         pf = vsi->back;
> >  
> > -       /* do not unregister while driver is in the reset recovery
> > pending
> > -        * state. Since reset/rebuild happens through PF service
> > task workqueue,
> > -        * it's not a good idea to unregister netdev that is
> > associated to the
> > -        * PF that is running the work queue items currently. This
> > is done to
> > -        * avoid check_flush_dependency() warning on this wq
> > -        */
> > -       if (vsi->netdev && !ice_is_reset_in_progress(pf->state) &&
> > -           (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state))) {
> > -               unregister_netdev(vsi->netdev);
> > -               clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi->state);
> > -       }
> > -
> > -       if (vsi->type == ICE_VSI_PF)
> > -               ice_devlink_destroy_pf_port(pf);
> > -
> >         if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
> >                 ice_rss_clean(vsi);
> >  
> >         ice_vsi_close(vsi);
> >         ice_vsi_decfg(vsi);
> >  
> > -       if (vsi->netdev) {
> > -               if (test_bit(ICE_VSI_NETDEV_REGISTERED, vsi-
> > >state)) {
> > -                       unregister_netdev(vsi->netdev);
> > -                       clear_bit(ICE_VSI_NETDEV_REGISTERED, vsi-
> > >state);
> > -               }
> > -               if (test_bit(ICE_VSI_NETDEV_ALLOCD, vsi->state)) {
> > -                       free_netdev(vsi->netdev);
> > -                       vsi->netdev = NULL;
> > -                       clear_bit(ICE_VSI_NETDEV_ALLOCD, vsi-
> > >state);
> > -               }
> > -       }
> > -
> >         /* retain SW VSI data structure since it is needed to
> > unregister and
> >          * free VSI netdev when PF is not in reset recovery pending
> > state,\
> >          * for ex: during rmmod.
> > --
> > 2.41.0
> >
> > _______________________________________________
> > Intel-wired-lan mailing list
> > [email protected]
> > https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
>