| Message ID | 20210121112937.30989-1-oleksandr.mazur@plvision.eu |
|---|---|
| State | New |
| Headers | show |
| Series | [net-next] net: core: devlink: add new trap action HARD_DROP | expand |
On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: > Add new trap action HARD_DROP, which can be used by the > drivers to register traps, where it's impossible to get > packet reported to the devlink subsystem by the device > driver, because it's impossible to retrieve dropped packet > from the device itself. > In order to use this action, driver must also register > additional devlink operation - callback that is used > to retrieve number of packets that have been dropped by > the device. Are these global statistics about number of packets the hardware dropped for a specific reason or are these per-port statistics? It's a creative use of devlink-trap interface, but I think it makes sense. Better to re-use an existing interface than creating yet another one. Anyway, this patch really needs to be marked as "RFC" since we cannot add infrastructure without anyone using it. Additionally, the documentation (Documentation/networking/devlink/devlink-trap.rst) needs to be updated, netdevsim needs to be patched and the test over netdevsim (tools/testing/selftests/drivers/net/netdevsim/devlink_trap.sh) needs to be extended to cover the new functionality. More comments below. > > Signed-off-by: Oleksandr Mazur <oleksandr.mazur@plvision.eu> > --- > include/net/devlink.h | 10 ++++++++ > include/uapi/linux/devlink.h | 4 ++++ > net/core/devlink.c | 44 +++++++++++++++++++++++++++++++++++- > 3 files changed, 57 insertions(+), 1 deletion(-) > > diff --git a/include/net/devlink.h b/include/net/devlink.h > index f466819cc477..6811a614f6fd 100644 > --- a/include/net/devlink.h > +++ b/include/net/devlink.h > @@ -1294,6 +1294,16 @@ struct devlink_ops { > const struct devlink_trap_group *group, > enum devlink_trap_action action, > struct netlink_ext_ack *extack); > + /** > + * @trap_hard_drop_counter_get: Trap hard drop counter get function. > + * > + * Should be used by device drivers to report number of packets dropped > + * by the underlying device, that have been dropped because device > + * failed to pass the trapped packet. > + */ > + int (*trap_hard_drop_counter_get)(struct devlink *devlink, > + const struct devlink_trap *trap, > + u64 *p_drops); > /** > * @trap_policer_init: Trap policer initialization function. > * > diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h > index cf89c318f2ac..9247d9c7db03 100644 > --- a/include/uapi/linux/devlink.h > +++ b/include/uapi/linux/devlink.h > @@ -261,12 +261,16 @@ enum { > * enum devlink_trap_action - Packet trap action. > * @DEVLINK_TRAP_ACTION_DROP: Packet is dropped by the device and a copy is not > * sent to the CPU. > + * @DEVLINK_TRAP_ACTION_HARD_DROP: Packet was dropped by the underlying device, > + * and device cannot report packet to devlink > + * (or inject it into the kernel RX path). > * @DEVLINK_TRAP_ACTION_TRAP: The sole copy of the packet is sent to the CPU. > * @DEVLINK_TRAP_ACTION_MIRROR: Packet is forwarded by the device and a copy is > * sent to the CPU. > */ > enum devlink_trap_action { > DEVLINK_TRAP_ACTION_DROP, > + DEVLINK_TRAP_ACTION_HARD_DROP, This breaks uAPI. New values should be added at the end. > DEVLINK_TRAP_ACTION_TRAP, > DEVLINK_TRAP_ACTION_MIRROR, > }; > diff --git a/net/core/devlink.c b/net/core/devlink.c > index ee828e4b1007..5a06e00429e1 100644 > --- a/net/core/devlink.c > +++ b/net/core/devlink.c > @@ -6732,6 +6732,7 @@ devlink_trap_action_get_from_info(struct genl_info *info, > val = nla_get_u8(info->attrs[DEVLINK_ATTR_TRAP_ACTION]); > switch (val) { > case DEVLINK_TRAP_ACTION_DROP: > + case DEVLINK_TRAP_ACTION_HARD_DROP: > case DEVLINK_TRAP_ACTION_TRAP: > case DEVLINK_TRAP_ACTION_MIRROR: > *p_trap_action = val; > @@ -6820,6 +6821,37 @@ static int devlink_trap_stats_put(struct sk_buff *msg, > return -EMSGSIZE; > } > > +static int > +devlink_trap_hard_drop_stats_put(struct sk_buff *msg, > + struct devlink *devlink, > + const struct devlink_trap_item *trap_item) > +{ > + struct nlattr *attr; > + u64 drops; > + int err; > + > + err = devlink->ops->trap_hard_drop_counter_get(devlink, trap_item->trap, > + &drops); > + if (err) > + return err; > + > + attr = nla_nest_start(msg, DEVLINK_ATTR_STATS); > + if (!attr) > + return -EMSGSIZE; > + > + if (nla_put_u64_64bit(msg, DEVLINK_ATTR_STATS_RX_DROPPED, drops, > + DEVLINK_ATTR_PAD)) > + goto nla_put_failure; > + > + nla_nest_end(msg, attr); > + > + return 0; > + > +nla_put_failure: > + nla_nest_cancel(msg, attr); > + return -EMSGSIZE; > +} > + > static int devlink_nl_trap_fill(struct sk_buff *msg, struct devlink *devlink, > const struct devlink_trap_item *trap_item, > enum devlink_command cmd, u32 portid, u32 seq, > @@ -6857,7 +6889,10 @@ static int devlink_nl_trap_fill(struct sk_buff *msg, struct devlink *devlink, > if (err) > goto nla_put_failure; > > - err = devlink_trap_stats_put(msg, trap_item->stats); > + if (trap_item->action == DEVLINK_TRAP_ACTION_HARD_DROP) > + err = devlink_trap_hard_drop_stats_put(msg, devlink, trap_item); > + else > + err = devlink_trap_stats_put(msg, trap_item->stats); > if (err) > goto nla_put_failure; > > @@ -9697,6 +9732,10 @@ devlink_trap_register(struct devlink *devlink, > if (devlink_trap_item_lookup(devlink, trap->name)) > return -EEXIST; > > + if (trap->init_action == DEVLINK_TRAP_ACTION_HARD_DROP && > + !devlink->ops->trap_hard_drop_counter_get) > + return -EINVAL; > + > trap_item = kzalloc(sizeof(*trap_item), GFP_KERNEL); > if (!trap_item) > return -ENOMEM; > @@ -9876,6 +9915,9 @@ void devlink_trap_report(struct devlink *devlink, struct sk_buff *skb, > { > struct devlink_trap_item *trap_item = trap_ctx; > > + if (trap_item->action == DEVLINK_TRAP_ACTION_HARD_DROP) > + return; How can this happen? > + > devlink_trap_stats_update(trap_item->stats, skb->len); > devlink_trap_stats_update(trap_item->group_item->stats, skb->len); > > -- > 2.17.1 >
On Thu, 21 Jan 2021 14:21:52 +0200 Ido Schimmel wrote: > On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: > > Add new trap action HARD_DROP, which can be used by the > > drivers to register traps, where it's impossible to get > > packet reported to the devlink subsystem by the device > > driver, because it's impossible to retrieve dropped packet > > from the device itself. > > In order to use this action, driver must also register > > additional devlink operation - callback that is used > > to retrieve number of packets that have been dropped by > > the device. > > Are these global statistics about number of packets the hardware dropped > for a specific reason or are these per-port statistics? > > It's a creative use of devlink-trap interface, but I think it makes > sense. Better to re-use an existing interface than creating yet another > one. Not sure if I agree, if we can't trap why is it a trap? It's just a counter.
Thu, Jan 21, 2021 at 06:36:05PM CET, kuba@kernel.org wrote: >On Thu, 21 Jan 2021 14:21:52 +0200 Ido Schimmel wrote: >> On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: >> > Add new trap action HARD_DROP, which can be used by the >> > drivers to register traps, where it's impossible to get >> > packet reported to the devlink subsystem by the device >> > driver, because it's impossible to retrieve dropped packet >> > from the device itself. >> > In order to use this action, driver must also register >> > additional devlink operation - callback that is used >> > to retrieve number of packets that have been dropped by >> > the device. >> >> Are these global statistics about number of packets the hardware dropped >> for a specific reason or are these per-port statistics? >> >> It's a creative use of devlink-trap interface, but I think it makes >> sense. Better to re-use an existing interface than creating yet another >> one. > >Not sure if I agree, if we can't trap why is it a trap? >It's just a counter. +1
Thu, Jan 21, 2021 at 06:36:05PM CET, kuba@kernel.org wrote: >On Thu, 21 Jan 2021 14:21:52 +0200 Ido Schimmel wrote: >> On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: >> > Add new trap action HARD_DROP, which can be used by the >> > drivers to register traps, where it's impossible to get >> > packet reported to the devlink subsystem by the device >> > driver, because it's impossible to retrieve dropped packet >> > from the device itself. >> > In order to use this action, driver must also register >> > additional devlink operation - callback that is used >> > to retrieve number of packets that have been dropped by >> > the device. >> >> Are these global statistics about number of packets the hardware dropped >> for a specific reason or are these per-port statistics? >> >> It's a creative use of devlink-trap interface, but I think it makes >> sense. Better to re-use an existing interface than creating yet another >> one. > >Not sure if I agree, if we can't trap why is it a trap? >It's just a counter. >+1 Device might be unable to trap only the 'DROP' packets, and this information should be transparent for the user. I agree on the statement, that new action might be an overhead. I could continue on with the solution Ido Schimmel proposed: since no new action would be needed and no UAPI changes are required, i could simply do the dropped statistics (additional field) output added upon trap stats queiring. (In case if driver registerd callback, of course; and do so only for DROP actions)
Mon, Jan 25, 2021 at 01:24:27PM CET, oleksandr.mazur@plvision.eu wrote: >Thu, Jan 21, 2021 at 06:36:05PM CET, kuba@kernel.org wrote: >>On Thu, 21 Jan 2021 14:21:52 +0200 Ido Schimmel wrote: >>> On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: >>> > Add new trap action HARD_DROP, which can be used by the >>> > drivers to register traps, where it's impossible to get >>> > packet reported to the devlink subsystem by the device >>> > driver, because it's impossible to retrieve dropped packet >>> > from the device itself. >>> > In order to use this action, driver must also register >>> > additional devlink operation - callback that is used >>> > to retrieve number of packets that have been dropped by >>> > the device. >>> >>> Are these global statistics about number of packets the hardware dropped >>> for a specific reason or are these per-port statistics? >>> >>> It's a creative use of devlink-trap interface, but I think it makes >>> sense. Better to re-use an existing interface than creating yet another >>> one. >> >>Not sure if I agree, if we can't trap why is it a trap? >>It's just a counter. > >>+1 >Device might be unable to trap only the 'DROP' packets, and this information should be transparent for the user. > >I agree on the statement, that new action might be an overhead. >I could continue on with the solution Ido Schimmel proposed: since no new action would be needed and no UAPI changes are required, i could simply do the dropped statistics (additional field) output added upon trap stats queiring. >(In case if driver registerd callback, of course; and do so only for DROP actions) It is not "a trap". You just need to count dropped packet. You don't trap anything. That is why I don't think this has anything to do with "trap" infra.
On Mon, Jan 25, 2021 at 03:56:14PM +0100, Jiri Pirko wrote: > Mon, Jan 25, 2021 at 01:24:27PM CET, oleksandr.mazur@plvision.eu wrote: > >Thu, Jan 21, 2021 at 06:36:05PM CET, kuba@kernel.org wrote: > >>On Thu, 21 Jan 2021 14:21:52 +0200 Ido Schimmel wrote: > >>> On Thu, Jan 21, 2021 at 01:29:37PM +0200, Oleksandr Mazur wrote: > >>> > Add new trap action HARD_DROP, which can be used by the > >>> > drivers to register traps, where it's impossible to get > >>> > packet reported to the devlink subsystem by the device > >>> > driver, because it's impossible to retrieve dropped packet > >>> > from the device itself. > >>> > In order to use this action, driver must also register > >>> > additional devlink operation - callback that is used > >>> > to retrieve number of packets that have been dropped by > >>> > the device. > >>> > >>> Are these global statistics about number of packets the hardware dropped > >>> for a specific reason or are these per-port statistics? > >>> > >>> It's a creative use of devlink-trap interface, but I think it makes > >>> sense. Better to re-use an existing interface than creating yet another > >>> one. > >> > >>Not sure if I agree, if we can't trap why is it a trap? > >>It's just a counter. > > > >>+1 > >Device might be unable to trap only the 'DROP' packets, and this information should be transparent for the user. > > > >I agree on the statement, that new action might be an overhead. > >I could continue on with the solution Ido Schimmel proposed: since no new action would be needed and no UAPI changes are required, i could simply do the dropped statistics (additional field) output added upon trap stats queiring. > >(In case if driver registerd callback, of course; and do so only for DROP actions) > > It is not "a trap". You just need to count dropped packet. You don't > trap anything. That is why I don't think this has anything to do with > "trap" infra. From [1] I understand that it is a trap and the action can be switched, but when it is 'drop', the hardware can provide statistics about number of packets that were discarded in hardware. If this is correct, then the suggestion in [2] looks valid to me. [1] https://lore.kernel.org/netdev/AM0P190MB073828252FFDA3215387765CE4A00@AM0P190MB0738.EURP190.PROD.OUTLOOK.COM/ [2] https://lore.kernel.org/netdev/20210123160348.GB2799851@shredder.lan/
diff --git a/include/net/devlink.h b/include/net/devlink.h index f466819cc477..6811a614f6fd 100644 --- a/include/net/devlink.h +++ b/include/net/devlink.h @@ -1294,6 +1294,16 @@ struct devlink_ops { const struct devlink_trap_group *group, enum devlink_trap_action action, struct netlink_ext_ack *extack); + /** + * @trap_hard_drop_counter_get: Trap hard drop counter get function. + * + * Should be used by device drivers to report number of packets dropped + * by the underlying device, that have been dropped because device + * failed to pass the trapped packet. + */ + int (*trap_hard_drop_counter_get)(struct devlink *devlink, + const struct devlink_trap *trap, + u64 *p_drops); /** * @trap_policer_init: Trap policer initialization function. * diff --git a/include/uapi/linux/devlink.h b/include/uapi/linux/devlink.h index cf89c318f2ac..9247d9c7db03 100644 --- a/include/uapi/linux/devlink.h +++ b/include/uapi/linux/devlink.h @@ -261,12 +261,16 @@ enum { * enum devlink_trap_action - Packet trap action. * @DEVLINK_TRAP_ACTION_DROP: Packet is dropped by the device and a copy is not * sent to the CPU. + * @DEVLINK_TRAP_ACTION_HARD_DROP: Packet was dropped by the underlying device, + * and device cannot report packet to devlink + * (or inject it into the kernel RX path). * @DEVLINK_TRAP_ACTION_TRAP: The sole copy of the packet is sent to the CPU. * @DEVLINK_TRAP_ACTION_MIRROR: Packet is forwarded by the device and a copy is * sent to the CPU. */ enum devlink_trap_action { DEVLINK_TRAP_ACTION_DROP, + DEVLINK_TRAP_ACTION_HARD_DROP, DEVLINK_TRAP_ACTION_TRAP, DEVLINK_TRAP_ACTION_MIRROR, }; diff --git a/net/core/devlink.c b/net/core/devlink.c index ee828e4b1007..5a06e00429e1 100644 --- a/net/core/devlink.c +++ b/net/core/devlink.c @@ -6732,6 +6732,7 @@ devlink_trap_action_get_from_info(struct genl_info *info, val = nla_get_u8(info->attrs[DEVLINK_ATTR_TRAP_ACTION]); switch (val) { case DEVLINK_TRAP_ACTION_DROP: + case DEVLINK_TRAP_ACTION_HARD_DROP: case DEVLINK_TRAP_ACTION_TRAP: case DEVLINK_TRAP_ACTION_MIRROR: *p_trap_action = val; @@ -6820,6 +6821,37 @@ static int devlink_trap_stats_put(struct sk_buff *msg, return -EMSGSIZE; } +static int +devlink_trap_hard_drop_stats_put(struct sk_buff *msg, + struct devlink *devlink, + const struct devlink_trap_item *trap_item) +{ + struct nlattr *attr; + u64 drops; + int err; + + err = devlink->ops->trap_hard_drop_counter_get(devlink, trap_item->trap, + &drops); + if (err) + return err; + + attr = nla_nest_start(msg, DEVLINK_ATTR_STATS); + if (!attr) + return -EMSGSIZE; + + if (nla_put_u64_64bit(msg, DEVLINK_ATTR_STATS_RX_DROPPED, drops, + DEVLINK_ATTR_PAD)) + goto nla_put_failure; + + nla_nest_end(msg, attr); + + return 0; + +nla_put_failure: + nla_nest_cancel(msg, attr); + return -EMSGSIZE; +} + static int devlink_nl_trap_fill(struct sk_buff *msg, struct devlink *devlink, const struct devlink_trap_item *trap_item, enum devlink_command cmd, u32 portid, u32 seq, @@ -6857,7 +6889,10 @@ static int devlink_nl_trap_fill(struct sk_buff *msg, struct devlink *devlink, if (err) goto nla_put_failure; - err = devlink_trap_stats_put(msg, trap_item->stats); + if (trap_item->action == DEVLINK_TRAP_ACTION_HARD_DROP) + err = devlink_trap_hard_drop_stats_put(msg, devlink, trap_item); + else + err = devlink_trap_stats_put(msg, trap_item->stats); if (err) goto nla_put_failure; @@ -9697,6 +9732,10 @@ devlink_trap_register(struct devlink *devlink, if (devlink_trap_item_lookup(devlink, trap->name)) return -EEXIST; + if (trap->init_action == DEVLINK_TRAP_ACTION_HARD_DROP && + !devlink->ops->trap_hard_drop_counter_get) + return -EINVAL; + trap_item = kzalloc(sizeof(*trap_item), GFP_KERNEL); if (!trap_item) return -ENOMEM; @@ -9876,6 +9915,9 @@ void devlink_trap_report(struct devlink *devlink, struct sk_buff *skb, { struct devlink_trap_item *trap_item = trap_ctx; + if (trap_item->action == DEVLINK_TRAP_ACTION_HARD_DROP) + return; + devlink_trap_stats_update(trap_item->stats, skb->len); devlink_trap_stats_update(trap_item->group_item->stats, skb->len);
Add new trap action HARD_DROP, which can be used by the drivers to register traps, where it's impossible to get packet reported to the devlink subsystem by the device driver, because it's impossible to retrieve dropped packet from the device itself. In order to use this action, driver must also register additional devlink operation - callback that is used to retrieve number of packets that have been dropped by the device. Signed-off-by: Oleksandr Mazur <oleksandr.mazur@plvision.eu> --- include/net/devlink.h | 10 ++++++++ include/uapi/linux/devlink.h | 4 ++++ net/core/devlink.c | 44 +++++++++++++++++++++++++++++++++++- 3 files changed, 57 insertions(+), 1 deletion(-)