* Re: [PATCH net] bridge: mrp: reject zero test interval to avoid OOM panic
From: Jakub Kicinski @ 2026-03-28 0:19 UTC (permalink / raw)
To: Nikolay Aleksandrov
Cc: Simon Horman, Xiang Mei, netdev, bridge, idosch, davem, edumazet,
pabeni, bestswngs
In-Reply-To: <f7fea291-bde9-4f95-8a59-1b209407ee27@blackwall.org>
On Fri, 27 Mar 2026 13:46:39 +0200 Nikolay Aleksandrov wrote:
> On 27/03/2026 13:34, Simon Horman wrote:
> > On Wed, Mar 25, 2026 at 08:24:38PM -0700, Xiang Mei wrote:
> >> br_mrp_start_test() and br_mrp_start_in_test() accept the user-supplied
> >> interval value from netlink without validation. When interval is 0,
> >> usecs_to_jiffies(0) yields 0, causing the delayed work
> >> (br_mrp_test_work_expired / br_mrp_in_test_work_expired) to reschedule
> >> itself with zero delay. This creates a tight loop on system_percpu_wq
> >> that allocates and transmits MRP test frames at maximum rate, exhausting
> >> all system memory and causing a kernel panic via OOM deadlock.
> >
> > I would suspect the primary outcome of this problem is high CPU consumption
> > rather than memory exhaustion. Is there a reason to expect that
> > the transmitted fames can't be consumed as fast as they are created?
>
> +1
> More so with CAP_NET_ADMIN you can cause all sorts of OOM and high-cpu usage
> conditions. This is a configuration error and OOM doesn't lead to panic unless
> instructed to. I don't think this is worth changing at all.
Then again if there's no practical use for 0 we should consider
the risk of getting this sort of submission over and over again?
Dunno..
^ permalink raw reply
* Re: [PATCH bpf-next 1/3] bpf: Disallow freplace on XDP with mismatched xdp_has_frags values
From: Jakub Kicinski @ 2026-03-28 0:12 UTC (permalink / raw)
To: Leon Hwang
Cc: bpf, Alexei Starovoitov, Daniel Borkmann, John Fastabend,
Andrii Nakryiko, Martin KaFai Lau, Eduard Zingerman, Song Liu,
Yonghong Song, KP Singh, Stanislav Fomichev, Hao Luo, Jiri Olsa,
Shuah Khan, David S . Miller, Jesper Dangaard Brouer,
Toke Hoiland-Jorgensen, Lorenzo Bianconi, linux-kernel,
linux-kselftest, netdev, kernel-patches-bot
In-Reply-To: <785e0b04-91e4-457f-8d6c-e0a911c0fd88@linux.dev>
On Fri, 27 Mar 2026 14:42:19 +0800 Leon Hwang wrote:
> > It may be worth adding a selftest to
> > tools/testing/selftests/drivers/net/xdp.py
> > which sets MTU to 9k, tries to attach a non-frag-capable prog
> > if that fails attaches a frag-capable prog and then checks if
> > replacing the capable prog with non-capable fails.
> > Drivers may be buggy in this regard.
>
> Do you mean adding these two tests to xdp.py?
>
> 1. Verify the failure of attaching non-frag-capable prog to mtu=9k
> driver.
> 2. Verify the failure of updating frag-capable prog with non-frag-
> capable prog via libbpf.c::bpf_link__update_program().
Not directly via libbpf just ip -f link, it uses libbpf internally
but testing with ip link is simpler. Also - there are already XDP
progs which return XDP_PASS in frag-capable and non-capable mode,
so just use those.
> As for test #2, it may require a helper to attach prog then update prog
> using libbpf's APIs.
See above.
^ permalink raw reply
* Re: [PATCH net-next 3/5] dpll: zl3073x: add ref sync and output clock type helpers
From: Jakub Kicinski @ 2026-03-27 23:59 UTC (permalink / raw)
To: Prathosh.Satish, ivecera
Cc: netdev, arkadiusz.kubalewski, jiri, mschmidt, poros, horms,
vadim.fedorenko, linux-kernel, conor+dt, krzk+dt, robh,
devicetree, pvaanane
In-Reply-To: <CY5PR11MB64621FF7F9F8E5EF85C7057EEC57A@CY5PR11MB6462.namprd11.prod.outlook.com>
On Fri, 27 Mar 2026 15:35:13 +0000 Prathosh.Satish@microchip.com wrote:
> Reviewed-by: prathosh.satish@microchip.com
That's not a correct format for a review tag, and you need to
make sure that the quoted lines are prefixed with >
Outlook style quoting does not work, it makes patchwork think
that you added all the tags that were in the quoted text again.
Ivan, could you sort this out and repost (add corrected tags from
Prathosh)? Last time I tried to fix such mess while applying it ended
poorly :(
^ permalink raw reply
* Re: [EXTERNAL] Re: [PATCH net-next v2] net: mana: Set default number of queues to 16
From: Jakub Kicinski @ 2026-03-27 23:55 UTC (permalink / raw)
To: Long Li
Cc: Konstantin Taranov, David S . Miller, Paolo Abeni, Eric Dumazet,
Andrew Lunn, Jason Gunthorpe, Leon Romanovsky, Haiyang Zhang,
KY Srinivasan, Wei Liu, Dexuan Cui, Simon Horman,
netdev@vger.kernel.org, linux-rdma@vger.kernel.org,
linux-hyperv@vger.kernel.org, linux-kernel@vger.kernel.org
In-Reply-To: <SA1PR21MB668314B1AF002E40B379F1C0CE57A@SA1PR21MB6683.namprd21.prod.outlook.com>
On Fri, 27 Mar 2026 04:00:31 +0000 Long Li wrote:
> We considered netif_get_num_default_rss_queues() but chose a fixed default based on our performance testing. On Azure VMs, typical
> workloads plateau at around 16 queues - adding more queues beyond that doesn't improve throughput but increases memory usage and
> interrupt overhead.
>
> netif_get_num_default_rss_queues() would return 32-64 on large VMs (64-128 vCPUs), which wastes resources without benefit.
>
> That said, I agree that completely ignoring the core-based heuristic isn't ideal for consistency. One option is to use
> netif_get_num_default_rss_queues() but clamp it to a maximum of MANA_DEF_NUM_QUEUES (16), so small VMs still get enough queues and
> large VMs don't over-allocate. Something like:
>
> apc->num_queues = min(netif_get_num_default_rss_queues(), MANA_DEF_NUM_QUEUES);
> apc->num_queues = min(apc->num_queues, gc->max_num_queues);
>
> For reference, it seems mlx4 does something similar - it caps at DEF_RX_RINGS (16) regardless of core count.
mlx4 is a bit ancient. And mlx5 does the wrong thing, which is why
I'm so sensitive to this issue :(
> Do you want me to send a v2?
Please send a follow up, let's leave this patch be and make an
incremental change.
Thanks!
^ permalink raw reply
* Re: [PATCH net-next v2 0/2] Add support for PIC64-HPSC/HX MDIO controller
From: Jakub Kicinski @ 2026-03-27 23:51 UTC (permalink / raw)
To: Charles Perry
Cc: netdev, Andrew Lunn, David S. Miller, Eric Dumazet, Paolo Abeni,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit,
Russell King, devicetree, linux-kernel
In-Reply-To: <acZ5WDz5qVxBDj3e@bby-cbu-swbuild03.eng.microchip.com>
On Fri, 27 Mar 2026 05:34:32 -0700 Charles Perry wrote:
> On Thu, Mar 26, 2026 at 08:33:09PM -0700, Jakub Kicinski wrote:
> > On Mon, 23 Mar 2026 15:02:52 -0700 Charles Perry wrote:
> > > .../net/microchip,pic64hpsc-mdio.yaml | 68 +++++++
> > > drivers/net/mdio/Kconfig | 7 +
> > > drivers/net/mdio/Makefile | 1 +
> > > drivers/net/mdio/mdio-pic64hpsc.c | 192 ++++++++++++++++++
> >
> > Speaking under correction from PHY maintainers but I think we need
> > a MAINTAINERS entry that will cover Microchip MDIO, or at least the
> > files you're adding. Important read:
> > https://docs.kernel.org/next/maintainer/feature-and-driver-maintainers.html
>
> Sure, I think this should go under "RISC-V MICROCHIP SUPPORT" or maybe a
> new "MICROCHIP PIC64-HPSC/HX DRIVER" entry if the former was meant only for
> Polarfire SoC.
>
> I'll add something in v3.
Thanks! FWIW I'd prefer the latter - smaller entries make the
responsibility clear. Under a big arch entry the maintainers are
usually seeing too many random patches to act. You can add
_both_ dedicated a entry and add the files to RISC-V MICROCHIP,
if you prefer, too.
^ permalink raw reply
* Re: [PATCH net-next 1/2] devlink: unify devlink_shd_get_priv() into devlink_priv()
From: Jakub Kicinski @ 2026-03-27 23:47 UTC (permalink / raw)
To: Przemek Kitszel
Cc: Jiri Pirko, netdev, Tony Nguyen, intel-wired-lan,
Aleksandr Loktionov, edumazet, horms, pabeni, davem,
Michal Schmidt
In-Reply-To: <f930e404-412c-46c0-9e3c-feee4d3c70fb@intel.com>
On Fri, 27 Mar 2026 08:42:47 +0100 Przemek Kitszel wrote:
> > If we are worried about misuse we should instead add an accessor
> > for "individual" (better name welcome) instance and WARN_ON()
> > when devlink_priv() is used in the shared setup.
>
> that would require the same amount of code as this patch (curr ver)
> has, only with WARN_ON() instead proper value (IOW: we detect what
> developer wanted, and give them big warning instead)
I guess I don't understand why you're so certain that you know what
the developer wanted. I'm of the opinion that the "individual" devlink
instances should not exist at all. Y'all want them, and claim to have
use cases for them. And yet, it is somehow not valid to get their priv.
^ permalink raw reply
* Re: [PATCH v1 net-next] tun: Ignore tun in netdev_lock_pos().
From: Willem de Bruijn @ 2026-03-27 23:16 UTC (permalink / raw)
To: Kuniyuki Iwashima, Eric Dumazet
Cc: David S. Miller, Jakub Kicinski, Paolo Abeni, Willem de Bruijn,
Jason Wang, Andrew Lunn, Simon Horman, Kuniyuki Iwashima, netdev
In-Reply-To: <CAAVpQUB=LtWXq_bzET2DOkyw_y5qRsuUf_fNCRE+TtiVHFktZw@mail.gmail.com>
Kuniyuki Iwashima wrote:
> On Fri, Mar 27, 2026 at 2:48 PM Eric Dumazet <edumazet@google.com> wrote:
> >
> > On Fri, Mar 27, 2026 at 2:34 PM Kuniyuki Iwashima <kuniyu@google.com> wrote:
> > >
> > > dev->type could be any value on tun due to TUNSETLINK.
> > >
> > > Let's not warn about it in netdev_lock_pos().
> > >
> > > Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
> >
> > Hmmm... should we instead have a list of supported types ?
> >
> > And reject user space requests for unsupported ones ?
>
> I chose a safer way but maybe we could reuse tun_get_addr_len()
> and define the default case there as unsupported ?
>
> I guess it's okay even if it breaks userspace because we can
> just add a single line to restore the support.
We shouldn't risk breaking userspace, over such a sanity check.
Presumably any truly newly added type will have a non-zero addr_len.
While tun sets this to zero for weird inputs. Maybe just filter on
that in the WARN_ONCE?
^ permalink raw reply
* Re: [PATCH net v2] rtnetlink: add missing netlink_ns_capable() check for peer netns
From: Kuniyuki Iwashima @ 2026-03-27 22:38 UTC (permalink / raw)
To: Nikolaos Gkarlis; +Cc: netdev, kuba
In-Reply-To: <20260327221203.430454-1-nickgarlis@gmail.com>
On Fri, Mar 27, 2026 at 3:12 PM Nikolaos Gkarlis <nickgarlis@gmail.com> wrote:
>
> rtnl_newlink() lacks a CAP_NET_ADMIN capability check on the peer
> network namespace when creating paired devices (veth, vxcan,
> netkit). This allows an unprivileged user with a user namespace
> to create interfaces in arbitrary network namespaces, including
> init_net.
>
> Add a netlink_ns_capable() check for CAP_NET_ADMIN in the peer
> namespace before allowing device creation to proceed.
>
> Fixes: 81adee47dfb6 ("net: Support specifying the network namespace upon device creation.")
> Signed-off-by: Nikolaos Gkarlis <nickgarlis@gmail.com>
> ---
> v2:
> - Removed "Reported-by" tag
> - Fixed "Fixes" tag with the help of Kuniyuki Iwashima (thanks !)
>
> net/core/rtnetlink.c | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
> index fae8034efbf..ee7d9d65b2c 100644
> --- a/net/core/rtnetlink.c
> +++ b/net/core/rtnetlink.c
> @@ -4059,8 +4059,15 @@ static int rtnl_newlink(struct sk_buff *skb, struct nlmsghdr *nlh,
> ret = PTR_ERR(peer_net);
> goto put_ops;
> }
> - if (peer_net)
> + if (peer_net) {
> + if (!netlink_ns_capable(skb, peer_net->user_ns,
> + CAP_NET_ADMIN)) {
I think this can be moved after the last rtnl_link_get_net_ifla()
in rtnl_get_peer_net().
If netns is not fetched from the peer tb, rtnl_link_get_net_capable()
just below this chunk will perform the same check redundantly.
> + put_net(peer_net);
> + ret = -EPERM;
> + goto put_ops;
> + }
> rtnl_nets_add(&rtnl_nets, peer_net);
> + }
> }
> }
>
> --
> 2.34.1
>
^ permalink raw reply
* Re: [PATCH v1 net-next] tun: Ignore tun in netdev_lock_pos().
From: Kuniyuki Iwashima @ 2026-03-27 22:22 UTC (permalink / raw)
To: Eric Dumazet
Cc: David S. Miller, Jakub Kicinski, Paolo Abeni, Willem de Bruijn,
Jason Wang, Andrew Lunn, Simon Horman, Kuniyuki Iwashima, netdev
In-Reply-To: <CANn89iLshTzN95t+GV0Tk97n_BCDzjAGDNTb=vSyF1rQNqwe_A@mail.gmail.com>
On Fri, Mar 27, 2026 at 2:48 PM Eric Dumazet <edumazet@google.com> wrote:
>
> On Fri, Mar 27, 2026 at 2:34 PM Kuniyuki Iwashima <kuniyu@google.com> wrote:
> >
> > dev->type could be any value on tun due to TUNSETLINK.
> >
> > Let's not warn about it in netdev_lock_pos().
> >
> > Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
>
> Hmmm... should we instead have a list of supported types ?
>
> And reject user space requests for unsupported ones ?
I chose a safer way but maybe we could reuse tun_get_addr_len()
and define the default case there as unsupported ?
I guess it's okay even if it breaks userspace because we can
just add a single line to restore the support.
^ permalink raw reply
* RE: [PATCH iwl-net] ice: fix PTP timestamping broken by SyncE code on E825C
From: Loktionov, Aleksandr @ 2026-03-27 22:21 UTC (permalink / raw)
To: Oros, Petr, netdev@vger.kernel.org
Cc: stable@vger.kernel.org, Oros, Petr, Nguyen, Anthony L,
Kitszel, Przemyslaw, Andrew Lunn, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Richard Cochran,
Kubalewski, Arkadiusz, Nitka, Grzegorz, Vecera, Ivan,
intel-wired-lan@lists.osuosl.org, linux-kernel@vger.kernel.org
In-Reply-To: <20260327074658.2963328-1-poros@redhat.com>
> -----Original Message-----
> From: Petr Oros <poros@redhat.com>
> Sent: Friday, March 27, 2026 8:47 AM
> To: netdev@vger.kernel.org
> Cc: stable@vger.kernel.org; Oros, Petr <poros@redhat.com>; Nguyen,
> Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> <przemyslaw.kitszel@intel.com>; Andrew Lunn <andrew+netdev@lunn.ch>;
> David S. Miller <davem@davemloft.net>; Eric Dumazet
> <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>; Paolo Abeni
> <pabeni@redhat.com>; Richard Cochran <richardcochran@gmail.com>;
> Kubalewski, Arkadiusz <arkadiusz.kubalewski@intel.com>; Nitka,
> Grzegorz <grzegorz.nitka@intel.com>; Loktionov, Aleksandr
> <aleksandr.loktionov@intel.com>; Vecera, Ivan <ivecera@redhat.com>;
> intel-wired-lan@lists.osuosl.org; linux-kernel@vger.kernel.org
> Subject: [PATCH iwl-net] ice: fix PTP timestamping broken by SyncE
> code on E825C
>
> The E825C SyncE support added in commit ad1df4f2d591 ("ice: dpll:
> Support E825-C SyncE and dynamic pin discovery") introduced a SyncE
> reconfiguration block in ice_ptp_link_change() that prevents
> ice_ptp_port_phy_restart() from being called in several error paths.
> Without the PHY restart, PTP timestamps stop working after any link
> change event.
>
> There are three ways the PHY restart gets blocked:
>
> 1. When DPLL initialization fails (e.g. missing ACPI firmware node
> properties), ICE_FLAG_DPLL is not set and the function returns
> early
> before reaching the PHY restart.
>
> 2. When ice_tspll_bypass_mux_active_e825c() fails to read the CGU
> register, WARN_ON_ONCE fires and the function returns early.
>
> 3. When ice_tspll_cfg_synce_ethdiv_e825c() fails to configure the
> clock divider for an active pin, same early return.
>
> SyncE and PTP are independent features. SyncE reconfiguration failures
> must not prevent the PTP PHY restart that is essential for timestamp
> recovery after link changes.
>
> Fix by making the entire SyncE block conditional on ICE_FLAG_DPLL
> without an early return, and replacing the WARN_ON_ONCE + return error
> handling inside the loop with dev_err_once + break. The function
> always proceeds to ice_ptp_port_phy_restart() regardless of SyncE
> errors.
>
> Fixes: ad1df4f2d591 ("ice: dpll: Support E825-C SyncE and dynamic pin
> discovery")
> Signed-off-by: Petr Oros <poros@redhat.com>
> ---
> drivers/net/ethernet/intel/ice/ice_ptp.c | 22 ++++++++++++----------
> 1 file changed, 12 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c
> b/drivers/net/ethernet/intel/ice/ice_ptp.c
> index 094e96219f4565..60bc47099432a2 100644
> --- a/drivers/net/ethernet/intel/ice/ice_ptp.c
> +++ b/drivers/net/ethernet/intel/ice/ice_ptp.c
> @@ -1296,12 +1296,10 @@ void ice_ptp_link_change(struct ice_pf *pf,
> bool linkup)
> if (pf->hw.reset_ongoing)
> return;
>
> - if (hw->mac_type == ICE_MAC_GENERIC_3K_E825) {
> + if (hw->mac_type == ICE_MAC_GENERIC_3K_E825 &&
> + test_bit(ICE_FLAG_DPLL, pf->flags)) {
> int pin, err;
>
> - if (!test_bit(ICE_FLAG_DPLL, pf->flags))
> - return;
> -
> mutex_lock(&pf->dplls.lock);
> for (pin = 0; pin < ICE_SYNCE_CLK_NUM; pin++) {
> enum ice_synce_clk clk_pin;
> @@ -1314,15 +1312,19 @@ void ice_ptp_link_change(struct ice_pf *pf,
> bool linkup)
> port_num,
> &active,
> clk_pin);
> - if (WARN_ON_ONCE(err)) {
> - mutex_unlock(&pf->dplls.lock);
> - return;
> + if (err) {
> + dev_err_once(ice_pf_to_dev(pf),
> + "Failed to read SyncE bypass
> mux for pin %d, err %d\n",
> + pin, err);
> + break;
> }
>
> err = ice_tspll_cfg_synce_ethdiv_e825c(hw,
> clk_pin);
> - if (active && WARN_ON_ONCE(err)) {
> - mutex_unlock(&pf->dplls.lock);
> - return;
> + if (active && err) {
> + dev_err_once(ice_pf_to_dev(pf),
> + "Failed to configure SyncE ETH
> divider for pin %d, err %d\n",
> + pin, err);
> + break;
> }
> }
> mutex_unlock(&pf->dplls.lock);
> --
> 2.52.0
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
^ permalink raw reply
* [PATCH net v2] rtnetlink: add missing netlink_ns_capable() check for peer netns
From: Nikolaos Gkarlis @ 2026-03-27 22:12 UTC (permalink / raw)
To: netdev; +Cc: kuba, nickgarlis, kuniyu
rtnl_newlink() lacks a CAP_NET_ADMIN capability check on the peer
network namespace when creating paired devices (veth, vxcan,
netkit). This allows an unprivileged user with a user namespace
to create interfaces in arbitrary network namespaces, including
init_net.
Add a netlink_ns_capable() check for CAP_NET_ADMIN in the peer
namespace before allowing device creation to proceed.
Fixes: 81adee47dfb6 ("net: Support specifying the network namespace upon device creation.")
Signed-off-by: Nikolaos Gkarlis <nickgarlis@gmail.com>
---
v2:
- Removed "Reported-by" tag
- Fixed "Fixes" tag with the help of Kuniyuki Iwashima (thanks !)
net/core/rtnetlink.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index fae8034efbf..ee7d9d65b2c 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -4059,8 +4059,15 @@ static int rtnl_newlink(struct sk_buff *skb, struct nlmsghdr *nlh,
ret = PTR_ERR(peer_net);
goto put_ops;
}
- if (peer_net)
+ if (peer_net) {
+ if (!netlink_ns_capable(skb, peer_net->user_ns,
+ CAP_NET_ADMIN)) {
+ put_net(peer_net);
+ ret = -EPERM;
+ goto put_ops;
+ }
rtnl_nets_add(&rtnl_nets, peer_net);
+ }
}
}
--
2.34.1
^ permalink raw reply related
* Re: [PATCH v1 net-next] tun: Ignore tun in netdev_lock_pos().
From: Eric Dumazet @ 2026-03-27 21:48 UTC (permalink / raw)
To: Kuniyuki Iwashima
Cc: David S. Miller, Jakub Kicinski, Paolo Abeni, Willem de Bruijn,
Jason Wang, Andrew Lunn, Simon Horman, Kuniyuki Iwashima, netdev
In-Reply-To: <20260327213441.3781433-1-kuniyu@google.com>
On Fri, Mar 27, 2026 at 2:34 PM Kuniyuki Iwashima <kuniyu@google.com> wrote:
>
> dev->type could be any value on tun due to TUNSETLINK.
>
> Let's not warn about it in netdev_lock_pos().
>
> Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Hmmm... should we instead have a list of supported types ?
And reject user space requests for unsupported ones ?
^ permalink raw reply
* [PATCH v1 net-next] tun: Ignore tun in netdev_lock_pos().
From: Kuniyuki Iwashima @ 2026-03-27 21:34 UTC (permalink / raw)
To: David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni,
Willem de Bruijn, Jason Wang, Andrew Lunn
Cc: Simon Horman, Kuniyuki Iwashima, Kuniyuki Iwashima, netdev
dev->type could be any value on tun due to TUNSETLINK.
Let's not warn about it in netdev_lock_pos().
Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
---
drivers/net/tun.c | 2 +-
include/linux/if_tun.h | 12 ++++++++++++
net/core/dev.c | 22 +++++++++++++---------
3 files changed, 26 insertions(+), 10 deletions(-)
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index c492fda6fc15..48194e567656 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -2338,7 +2338,7 @@ static int tun_fill_info(struct sk_buff *skb, const struct net_device *dev)
return -EMSGSIZE;
}
-static struct rtnl_link_ops tun_link_ops __read_mostly = {
+struct rtnl_link_ops tun_link_ops __read_mostly = {
.kind = DRV_NAME,
.priv_size = sizeof(struct tun_struct),
.setup = tun_setup,
diff --git a/include/linux/if_tun.h b/include/linux/if_tun.h
index 80166eb62f41..89d9b7b0a044 100644
--- a/include/linux/if_tun.h
+++ b/include/linux/if_tun.h
@@ -20,6 +20,13 @@ struct tun_msg_ctl {
};
#if defined(CONFIG_TUN) || defined(CONFIG_TUN_MODULE)
+extern struct rtnl_link_ops tun_link_ops;
+
+static inline bool netdev_is_tun(const struct net_device *dev)
+{
+ return dev->rtnl_link_ops == &tun_link_ops;
+}
+
struct socket *tun_get_socket(struct file *);
struct ptr_ring *tun_get_tx_ring(struct file *file);
@@ -45,6 +52,11 @@ void tun_ptr_free(void *ptr);
struct file;
struct socket;
+static inline bool netdev_is_tun(const struct net_device *dev)
+{
+ return false;
+}
+
static inline struct socket *tun_get_socket(struct file *f)
{
return ERR_PTR(-EINVAL);
diff --git a/net/core/dev.c b/net/core/dev.c
index 200d44883fc1..bd91e0a3dbdf 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -89,6 +89,7 @@
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
+#include <linux/if_tun.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
@@ -526,24 +527,27 @@ static const char *const netdev_lock_name[] = {
static struct lock_class_key netdev_xmit_lock_key[ARRAY_SIZE(netdev_lock_type)];
static struct lock_class_key netdev_addr_lock_key[ARRAY_SIZE(netdev_lock_type)];
-static inline unsigned short netdev_lock_pos(unsigned short dev_type)
+static inline unsigned short netdev_lock_pos(const struct net_device *dev)
{
+ unsigned short dev_type = dev->type;
int i;
for (i = 0; i < ARRAY_SIZE(netdev_lock_type); i++)
if (netdev_lock_type[i] == dev_type)
return i;
+
/* the last key is used by default */
- WARN_ONCE(1, "netdev_lock_pos() could not find dev_type=%u\n", dev_type);
+ WARN_ONCE(!netdev_is_tun(dev),
+ "netdev_lock_pos() could not find dev_type=%u\n", dev_type);
return ARRAY_SIZE(netdev_lock_type) - 1;
}
-static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
- unsigned short dev_type)
+static inline void netdev_set_xmit_lockdep_class(const struct net_device *dev,
+ spinlock_t *lock)
{
int i;
- i = netdev_lock_pos(dev_type);
+ i = netdev_lock_pos(dev);
lockdep_set_class_and_name(lock, &netdev_xmit_lock_key[i],
netdev_lock_name[i]);
}
@@ -552,14 +556,14 @@ static inline void netdev_set_addr_lockdep_class(struct net_device *dev)
{
int i;
- i = netdev_lock_pos(dev->type);
+ i = netdev_lock_pos(dev);
lockdep_set_class_and_name(&dev->addr_list_lock,
&netdev_addr_lock_key[i],
netdev_lock_name[i]);
}
#else
-static inline void netdev_set_xmit_lockdep_class(spinlock_t *lock,
- unsigned short dev_type)
+static inline void netdev_set_xmit_lockdep_class(const struct net_device *dev,
+ spinlock_t *lock)
{
}
@@ -11210,7 +11214,7 @@ static void netdev_init_one_queue(struct net_device *dev,
{
/* Initialize queue lock */
spin_lock_init(&queue->_xmit_lock);
- netdev_set_xmit_lockdep_class(&queue->_xmit_lock, dev->type);
+ netdev_set_xmit_lockdep_class(dev, &queue->_xmit_lock);
queue->xmit_lock_owner = -1;
netdev_queue_numa_node_write(queue, NUMA_NO_NODE);
queue->dev = dev;
--
2.53.0.1018.g2bb0e51243-goog
^ permalink raw reply related
* Re: [PATCH] net: mana: fix use-after-free in add_adev() error path
From: Hardik Garg @ 2026-03-27 21:27 UTC (permalink / raw)
To: Guangshuo Li, K. Y. Srinivasan, Haiyang Zhang, Wei Liu,
Dexuan Cui, Long Li, Andrew Lunn, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Saurabh Sengar,
Erni Sri Satya Vennela, Shradha Gupta, Dipayaan Roy, Aditya Garg,
Shiraz Saleem, Leon Romanovsky, linux-hyperv, netdev,
linux-kernel
Cc: stable
In-Reply-To: <20260318154041.638747-1-lgs201920130244@gmail.com>
On 3/18/2026 8:40 AM, Guangshuo Li wrote:
> If auxiliary_device_add() fails, add_adev() calls
> auxiliary_device_uninit(adev), whose release callback adev_release()
> frees the containing struct mana_adev.
>
> The current error path then falls through to init_fail and accesses
> adev->id. Since adev is embedded in struct mana_adev, this may lead
> to a use-after-free.
>
> Fix it by storing the allocated auxiliary device id in a local
> variable and using that saved id in the cleanup path after
> auxiliary_device_uninit().
>
> Fixes: a69839d4327d ("net: mana: Add support for auxiliary device")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> ---
> drivers/net/ethernet/microsoft/mana/mana_en.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c
> index 1ad154f9db1a..70d71594c599 100644
> --- a/drivers/net/ethernet/microsoft/mana/mana_en.c
> +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c
> @@ -3362,6 +3362,7 @@ static int add_adev(struct gdma_dev *gd, const char *name)
> {
> struct auxiliary_device *adev;
> struct mana_adev *madev;
> + int id;
> int ret;
>
> madev = kzalloc(sizeof(*madev), GFP_KERNEL);
> @@ -3372,7 +3373,8 @@ static int add_adev(struct gdma_dev *gd, const char *name)
> ret = mana_adev_idx_alloc();
> if (ret < 0)
> goto idx_fail;
> - adev->id = ret;
> + id = ret;
> + adev->id = id;
>
> adev->name = name;
> adev->dev.parent = gd->gdma_context->dev;
> @@ -3398,7 +3400,7 @@ static int add_adev(struct gdma_dev *gd, const char *name)
> auxiliary_device_uninit(adev);
>
> init_fail:
> - mana_adev_idx_free(adev->id);
> + mana_adev_idx_free(id);
>
> idx_fail:
> kfree(madev);
Reviewed-by: Hardik Garg <hargar@linux.microsoft.com>
Thanks,
Hardik
^ permalink raw reply
* RE: [PATCH iwl-net] ice: fix PTP timestamping broken by SyncE code on E825C
From: Nitka, Grzegorz @ 2026-03-27 21:17 UTC (permalink / raw)
To: Oros, Petr, netdev@vger.kernel.org
Cc: stable@vger.kernel.org, Oros, Petr, Nguyen, Anthony L,
Kitszel, Przemyslaw, Andrew Lunn, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Richard Cochran,
Kubalewski, Arkadiusz, Loktionov, Aleksandr, Vecera, Ivan,
intel-wired-lan@lists.osuosl.org, linux-kernel@vger.kernel.org
In-Reply-To: <20260327074658.2963328-1-poros@redhat.com>
> -----Original Message-----
> From: Petr Oros <poros@redhat.com>
> Sent: Friday, March 27, 2026 8:47 AM
> To: netdev@vger.kernel.org
> Cc: stable@vger.kernel.org; Oros, Petr <poros@redhat.com>; Nguyen,
> Anthony L <anthony.l.nguyen@intel.com>; Kitszel, Przemyslaw
> <przemyslaw.kitszel@intel.com>; Andrew Lunn <andrew+netdev@lunn.ch>;
> David S. Miller <davem@davemloft.net>; Eric Dumazet
> <edumazet@google.com>; Jakub Kicinski <kuba@kernel.org>; Paolo Abeni
> <pabeni@redhat.com>; Richard Cochran <richardcochran@gmail.com>;
> Kubalewski, Arkadiusz <arkadiusz.kubalewski@intel.com>; Nitka, Grzegorz
> <grzegorz.nitka@intel.com>; Loktionov, Aleksandr
> <aleksandr.loktionov@intel.com>; Vecera, Ivan <ivecera@redhat.com>;
> intel-wired-lan@lists.osuosl.org; linux-kernel@vger.kernel.org
> Subject: [PATCH iwl-net] ice: fix PTP timestamping broken by SyncE code on
> E825C
>
> The E825C SyncE support added in commit ad1df4f2d591 ("ice: dpll:
> Support E825-C SyncE and dynamic pin discovery") introduced a SyncE
> reconfiguration block in ice_ptp_link_change() that prevents
> ice_ptp_port_phy_restart() from being called in several error paths.
> Without the PHY restart, PTP timestamps stop working after any link
> change event.
>
> There are three ways the PHY restart gets blocked:
>
> 1. When DPLL initialization fails (e.g. missing ACPI firmware node
> properties), ICE_FLAG_DPLL is not set and the function returns early
> before reaching the PHY restart.
>
> 2. When ice_tspll_bypass_mux_active_e825c() fails to read the CGU
> register, WARN_ON_ONCE fires and the function returns early.
>
> 3. When ice_tspll_cfg_synce_ethdiv_e825c() fails to configure the
> clock divider for an active pin, same early return.
>
> SyncE and PTP are independent features. SyncE reconfiguration failures
> must not prevent the PTP PHY restart that is essential for timestamp
> recovery after link changes.
>
> Fix by making the entire SyncE block conditional on ICE_FLAG_DPLL
> without an early return, and replacing the WARN_ON_ONCE + return error
> handling inside the loop with dev_err_once + break. The function always
> proceeds to ice_ptp_port_phy_restart() regardless of SyncE errors.
>
> Fixes: ad1df4f2d591 ("ice: dpll: Support E825-C SyncE and dynamic pin
> discovery")
> Signed-off-by: Petr Oros <poros@redhat.com>
Thanks Petr for catching and fixing this.
Tested it in the environment w/o ACPI support and, indeed, PTP stopped
working. This is a valid fix.
Thanks!
Reviewed-by: Grzegorz Nitka <grzegorz.nitka@intel.com>
> ---
> drivers/net/ethernet/intel/ice/ice_ptp.c | 22 ++++++++++++----------
> 1 file changed, 12 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c
> b/drivers/net/ethernet/intel/ice/ice_ptp.c
> index 094e96219f4565..60bc47099432a2 100644
> --- a/drivers/net/ethernet/intel/ice/ice_ptp.c
> +++ b/drivers/net/ethernet/intel/ice/ice_ptp.c
> @@ -1296,12 +1296,10 @@ void ice_ptp_link_change(struct ice_pf *pf, bool
> linkup)
> if (pf->hw.reset_ongoing)
> return;
>
> - if (hw->mac_type == ICE_MAC_GENERIC_3K_E825) {
> + if (hw->mac_type == ICE_MAC_GENERIC_3K_E825 &&
> + test_bit(ICE_FLAG_DPLL, pf->flags)) {
> int pin, err;
>
> - if (!test_bit(ICE_FLAG_DPLL, pf->flags))
> - return;
> -
> mutex_lock(&pf->dplls.lock);
> for (pin = 0; pin < ICE_SYNCE_CLK_NUM; pin++) {
> enum ice_synce_clk clk_pin;
> @@ -1314,15 +1312,19 @@ void ice_ptp_link_change(struct ice_pf *pf, bool
> linkup)
> port_num,
> &active,
> clk_pin);
> - if (WARN_ON_ONCE(err)) {
> - mutex_unlock(&pf->dplls.lock);
> - return;
> + if (err) {
> + dev_err_once(ice_pf_to_dev(pf),
> + "Failed to read SyncE bypass mux
> for pin %d, err %d\n",
> + pin, err);
> + break;
> }
>
> err = ice_tspll_cfg_synce_ethdiv_e825c(hw, clk_pin);
> - if (active && WARN_ON_ONCE(err)) {
> - mutex_unlock(&pf->dplls.lock);
> - return;
> + if (active && err) {
> + dev_err_once(ice_pf_to_dev(pf),
> + "Failed to configure SyncE ETH
> divider for pin %d, err %d\n",
> + pin, err);
> + break;
> }
> }
> mutex_unlock(&pf->dplls.lock);
> --
> 2.52.0
^ permalink raw reply
* [PATCH net-next] fbnic: Set Relaxed Ordering PCIe TLP attributes for DMA engines
From: Dimitri Daskalakis @ 2026-03-27 20:44 UTC (permalink / raw)
To: David S . Miller
Cc: Alexander Duyck, Jakub Kicinski, kernel-team, Andrew Lunn,
Eric Dumazet, Paolo Abeni, Simon Horman, Jacob Keller,
Mohsin Bashir, Lee Trager, Mike Marciniszyn, netdev
From: Alexander Duyck <alexanderduyck@fb.com>
Add ATTR CSR bit field definitions for the DMA engine TLP header
configuration registers:
AW_CFG: RDE_ATTR[17:15], RQM_ATTR[14:12], TQM_ATTR[11:9]
AR_CFG: TDE_ATTR[17:15], RQM_ATTR[14:12], TQM_ATTR[11:9]
These fields control the PCIe TLP attribute bits for outbound
transactions from the TQM, RQM, RDE (write path), and TDE (read path)
DMA engines. An enum is added with standard PCIe TLP attribute values:
NS (No Snoop), RO (Relaxed Ordering), and IDO (ID-based Ordering).
Read the PCIe Relaxed Ordering capability at probe time and store it in
fbnic_dev. Configure Relaxed Ordering on the PCIe TLP attributes in
fbnic_mbx_init_desc_ring when the capability is enabled. For the write
path (AW_CFG), set RO on RDE and TQM attributes. For the read path
(AR_CFG), set RO on all three attributes (TDE, RQM, TQM). This allows
the PCIe fabric to reorder these transactions for improved throughput.
Signed-off-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Dimitri Daskalakis <daskald@meta.com>
---
drivers/net/ethernet/meta/fbnic/fbnic.h | 1 +
drivers/net/ethernet/meta/fbnic/fbnic_csr.h | 12 ++++++++++++
drivers/net/ethernet/meta/fbnic/fbnic_devlink.c | 1 +
drivers/net/ethernet/meta/fbnic/fbnic_fw.c | 15 +++++++++++++--
4 files changed, 27 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/meta/fbnic/fbnic.h b/drivers/net/ethernet/meta/fbnic/fbnic.h
index f7df5302e91a..d0715695c43e 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic.h
+++ b/drivers/net/ethernet/meta/fbnic/fbnic.h
@@ -57,6 +57,7 @@ struct fbnic_dev {
u64 dsn;
u32 mps;
u32 readrq;
+ u8 relaxed_ord;
/* Local copy of the devices TCAM */
struct fbnic_act_tcam act_tcam[FBNIC_RPC_TCAM_ACT_NUM_ENTRIES];
diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_csr.h b/drivers/net/ethernet/meta/fbnic/fbnic_csr.h
index 43de522af172..81794bd326e1 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_csr.h
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_csr.h
@@ -975,9 +975,21 @@ enum {
#define FBNIC_PUL_OB_TLP_HDR_AW_CFG 0x3103d /* 0xc40f4 */
#define FBNIC_PUL_OB_TLP_HDR_AW_CFG_FLUSH CSR_BIT(19)
#define FBNIC_PUL_OB_TLP_HDR_AW_CFG_BME CSR_BIT(18)
+#define FBNIC_PUL_OB_TLP_HDR_AW_CFG_RDE_ATTR CSR_GENMASK(17, 15)
+#define FBNIC_PUL_OB_TLP_HDR_AW_CFG_RQM_ATTR CSR_GENMASK(14, 12)
+#define FBNIC_PUL_OB_TLP_HDR_AW_CFG_TQM_ATTR CSR_GENMASK(11, 9)
+enum {
+ FBNIC_TLP_ATTR_NS = 1,
+ FBNIC_TLP_ATTR_RO = 2,
+ FBNIC_TLP_ATTR_IDO = 4,
+};
+
#define FBNIC_PUL_OB_TLP_HDR_AR_CFG 0x3103e /* 0xc40f8 */
#define FBNIC_PUL_OB_TLP_HDR_AR_CFG_FLUSH CSR_BIT(19)
#define FBNIC_PUL_OB_TLP_HDR_AR_CFG_BME CSR_BIT(18)
+#define FBNIC_PUL_OB_TLP_HDR_AR_CFG_TDE_ATTR CSR_GENMASK(17, 15)
+#define FBNIC_PUL_OB_TLP_HDR_AR_CFG_RQM_ATTR CSR_GENMASK(14, 12)
+#define FBNIC_PUL_OB_TLP_HDR_AR_CFG_TQM_ATTR CSR_GENMASK(11, 9)
#define FBNIC_PUL_USER_OB_RD_TLP_CNT_31_0 \
0x3106e /* 0xc41b8 */
#define FBNIC_PUL_USER_OB_RD_DWORD_CNT_31_0 \
diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_devlink.c b/drivers/net/ethernet/meta/fbnic/fbnic_devlink.c
index f1c992f5fe94..546e1c12d287 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_devlink.c
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_devlink.c
@@ -647,6 +647,7 @@ struct fbnic_dev *fbnic_devlink_alloc(struct pci_dev *pdev)
fbd->dsn = pci_get_dsn(pdev);
fbd->mps = pcie_get_mps(pdev);
fbd->readrq = pcie_get_readrq(pdev);
+ fbd->relaxed_ord = pcie_relaxed_ordering_enabled(pdev);
fbd->mac_addr_boundary = FBNIC_RPC_TCAM_MACDA_DEFAULT_BOUNDARY;
diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_fw.c b/drivers/net/ethernet/meta/fbnic/fbnic_fw.c
index c2bad51bdde6..0c6812fcf185 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_fw.c
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_fw.c
@@ -448,6 +448,7 @@ static int fbnic_fw_xmit_simple_msg(struct fbnic_dev *fbd, u32 msg_type)
static int fbnic_mbx_init_desc_ring(struct fbnic_dev *fbd, int mbx_idx)
{
+ u8 tlp_attr = fbd->relaxed_ord ? FBNIC_TLP_ATTR_RO : 0;
struct fbnic_fw_mbx *mbx = &fbd->mbx[mbx_idx];
mbx->ready = true;
@@ -456,14 +457,24 @@ static int fbnic_mbx_init_desc_ring(struct fbnic_dev *fbd, int mbx_idx)
case FBNIC_IPC_MBX_RX_IDX:
/* Enable DMA writes from the device */
wr32(fbd, FBNIC_PUL_OB_TLP_HDR_AW_CFG,
- FBNIC_PUL_OB_TLP_HDR_AW_CFG_BME);
+ FBNIC_PUL_OB_TLP_HDR_AW_CFG_BME |
+ FIELD_PREP(FBNIC_PUL_OB_TLP_HDR_AW_CFG_RDE_ATTR,
+ tlp_attr) |
+ FIELD_PREP(FBNIC_PUL_OB_TLP_HDR_AW_CFG_TQM_ATTR,
+ tlp_attr));
/* Make sure we have a page for the FW to write to */
return fbnic_mbx_alloc_rx_msgs(fbd);
case FBNIC_IPC_MBX_TX_IDX:
/* Enable DMA reads from the device */
wr32(fbd, FBNIC_PUL_OB_TLP_HDR_AR_CFG,
- FBNIC_PUL_OB_TLP_HDR_AR_CFG_BME);
+ FBNIC_PUL_OB_TLP_HDR_AR_CFG_BME |
+ FIELD_PREP(FBNIC_PUL_OB_TLP_HDR_AR_CFG_TDE_ATTR,
+ tlp_attr) |
+ FIELD_PREP(FBNIC_PUL_OB_TLP_HDR_AR_CFG_RQM_ATTR,
+ tlp_attr) |
+ FIELD_PREP(FBNIC_PUL_OB_TLP_HDR_AR_CFG_TQM_ATTR,
+ tlp_attr));
break;
}
--
2.52.0
^ permalink raw reply related
* Re: [PATCH bpf-next 2/3] bpf: Disallow freplace on kprobe with mismatched kprobe_write_ctx values
From: Jiri Olsa @ 2026-03-27 20:41 UTC (permalink / raw)
To: Leon Hwang
Cc: bpf, Alexei Starovoitov, Daniel Borkmann, John Fastabend,
Andrii Nakryiko, Martin KaFai Lau, Eduard Zingerman, Song Liu,
Yonghong Song, KP Singh, Stanislav Fomichev, Hao Luo, Shuah Khan,
David S . Miller, Jakub Kicinski, Jesper Dangaard Brouer,
Toke Hoiland-Jorgensen, Lorenzo Bianconi, linux-kernel,
linux-kselftest, netdev, kernel-patches-bot
In-Reply-To: <20260324150444.68166-3-leon.hwang@linux.dev>
On Tue, Mar 24, 2026 at 11:04:43PM +0800, Leon Hwang wrote:
> uprobe programs are allowed to modify struct pt_regs.
>
> Since the actual program type of uprobe is KPROBE, it can be abused to
> modify struct pt_regs via kprobe+freplace when the kprobe attaches to
> kernel functions.
>
> For example,
>
> SEC("?kprobe")
> int kprobe(struct pt_regs *regs)
> {
> return 0;
> }
>
> SEC("?freplace")
> int freplace_kprobe(struct pt_regs *regs)
> {
> regs->di = 0;
> return 0;
> }
>
> freplace_kprobe prog will attach to kprobe prog.
> kprobe prog will attach to a kernel function.
>
> Without this patch, when the kernel function runs, its first arg will
> always be set as 0 via the freplace_kprobe prog.
>
> To avoid the abuse of kprobe_write_ctx=true via kprobe+freplace, disallow
> freplace on kprobe programs with mismatched kprobe_write_ctx values.
>
> Fixes: 7384893d970e ("bpf: Allow uprobe program to change context registers")
> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> ---
> kernel/bpf/verifier.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 12330466d58b..f8257bae6081 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -6404,6 +6404,14 @@ static int check_ctx_access(struct bpf_verifier_env *env, int insn_idx, int off,
> /* remember the offset of last byte accessed in ctx */
> if (env->prog->aux->max_ctx_offset < off + size)
> env->prog->aux->max_ctx_offset = off + size;
> + if (env->prog->type == BPF_PROG_TYPE_EXT) {
> + struct bpf_prog *dst_prog = env->prog->aux->dst_prog;
> +
> + if (env->prog->aux->kprobe_write_ctx != dst_prog->aux->kprobe_write_ctx) {
> + verbose(env, "Extension program cannot have different kprobe_write_ctx value with target prog\n");
> + return -EINVAL;
> + }
> + }
would it be more clear to check this in bpf_check_attach_target,
instead of depending on actual write to context
jirka
^ permalink raw reply
* Re: [PATCH v6 phy-next 09/28] scsi: ufs: exynos: stop poking into struct phy guts
From: Peter Griffin @ 2026-03-27 20:23 UTC (permalink / raw)
To: Vladimir Oltean
Cc: linux-phy, Vinod Koul, Neil Armstrong, dri-devel, freedreno,
linux-arm-kernel, linux-arm-msm, linux-can, linux-gpio, linux-ide,
linux-kernel, linux-media, linux-pci, linux-renesas-soc,
linux-riscv, linux-rockchip, linux-samsung-soc, linux-scsi,
linux-sunxi, linux-tegra, linux-usb, netdev, spacemit,
UNGLinuxDriver, Bart Van Assche, Alim Akhtar,
James E.J. Bottomley, Martin K. Petersen, Krzysztof Kozlowski,
Chanho Park
In-Reply-To: <20260327184706.1600329-10-vladimir.oltean@nxp.com>
On Fri, 27 Mar 2026 at 18:48, Vladimir Oltean <vladimir.oltean@nxp.com> wrote:
>
> The Exynos host controller driver is clearly a PHY consumer (gets the
> ufs->phy using devm_phy_get()), but pokes into the guts of struct phy
> to get the generic_phy->power_count.
>
> The UFS core (specifically ufshcd_link_startup()) may call the variant
> operation exynos_ufs_pre_link() -> exynos_ufs_phy_init() multiple times
> if the link startup fails and needs to be retried.
>
> However ufs-exynos shouldn't be doing what it's doing, i.e. looking at
> the generic_phy->power_count, because in the general sense of the API, a
> single Generic PHY may have multiple consumers. If ufs-exynos looks at
> generic_phy->power_count, there's no guarantee that this ufs-exynos
> instance is the one who previously bumped that power count. So it may be
> powering down the PHY on behalf of another consumer.
>
> The correct way in which this should be handled is ufs-exynos should
> *remember* whether it has initialized and powered up the PHY before, and
> power it down during link retries. Not rely on the power_count (which,
> btw, on the writer side is modified under &phy->mutex, but on the reader
> side is accessed unlocked). This is a discouraged pattern even if here
> it doesn't cause functional problems.
>
> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
> Reviewed-by: Bart Van Assche <bvanassche@acm.org>
> Acked-by: Alim Akhtar <alim.akhtar@samsung.com>
> Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
> ---
Reviewed-by: Peter Griffin <peter.griffin@linaro.org>
> Cc: Alim Akhtar <alim.akhtar@samsung.com>
> Cc: Peter Griffin <peter.griffin@linaro.org>
> Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>
> Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
> Cc: Krzysztof Kozlowski <krzk@kernel.org>
> Cc: Chanho Park <chanho61.park@samsung.com>
>
> v5->v6: collect tags from Alim Akhtar
> v4->v5: collect tag, add "scsi: " prefix to commit title
> v3->v4: none
> v2->v3:
> - add Cc Chanho Park, author of commit 3d73b200f989 ("scsi: ufs:
> ufs-exynos: Change ufs phy control sequence")
> v1->v2:
> - add better ufs->phy_powered_on handling in exynos_ufs_exit(),
> exynos_ufs_suspend() and exynos_ufs_resume() which ensures we won't
> enter a phy->power_count underrun condition
> ---
> drivers/ufs/host/ufs-exynos.c | 24 ++++++++++++++++++++----
> drivers/ufs/host/ufs-exynos.h | 1 +
> 2 files changed, 21 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/ufs/host/ufs-exynos.c b/drivers/ufs/host/ufs-exynos.c
> index 76fee3a79c77..274e53833571 100644
> --- a/drivers/ufs/host/ufs-exynos.c
> +++ b/drivers/ufs/host/ufs-exynos.c
> @@ -963,9 +963,10 @@ static int exynos_ufs_phy_init(struct exynos_ufs *ufs)
>
> phy_set_bus_width(generic_phy, ufs->avail_ln_rx);
>
> - if (generic_phy->power_count) {
> + if (ufs->phy_powered_on) {
> phy_power_off(generic_phy);
> phy_exit(generic_phy);
> + ufs->phy_powered_on = false;
> }
>
> ret = phy_init(generic_phy);
> @@ -979,6 +980,8 @@ static int exynos_ufs_phy_init(struct exynos_ufs *ufs)
> if (ret)
> goto out_exit_phy;
>
> + ufs->phy_powered_on = true;
> +
> return 0;
>
> out_exit_phy:
> @@ -1527,6 +1530,9 @@ static void exynos_ufs_exit(struct ufs_hba *hba)
> {
> struct exynos_ufs *ufs = ufshcd_get_variant(hba);
>
> + if (!ufs->phy_powered_on)
> + return;
> +
> phy_power_off(ufs->phy);
> phy_exit(ufs->phy);
> }
> @@ -1728,8 +1734,10 @@ static int exynos_ufs_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op,
> if (ufs->drv_data->suspend)
> ufs->drv_data->suspend(ufs);
>
> - if (!ufshcd_is_link_active(hba))
> + if (!ufshcd_is_link_active(hba) && ufs->phy_powered_on) {
> phy_power_off(ufs->phy);
> + ufs->phy_powered_on = false;
> + }
>
> return 0;
> }
> @@ -1737,9 +1745,17 @@ static int exynos_ufs_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op,
> static int exynos_ufs_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)
> {
> struct exynos_ufs *ufs = ufshcd_get_variant(hba);
> + int err;
>
> - if (!ufshcd_is_link_active(hba))
> - phy_power_on(ufs->phy);
> + if (!ufshcd_is_link_active(hba) && !ufs->phy_powered_on) {
> + err = phy_power_on(ufs->phy);
> + if (err) {
> + dev_err(hba->dev, "Failed to power on PHY: %pe\n",
> + ERR_PTR(err));
> + } else {
> + ufs->phy_powered_on = true;
> + }
> + }
>
> exynos_ufs_config_smu(ufs);
> exynos_ufs_fmp_resume(hba);
> diff --git a/drivers/ufs/host/ufs-exynos.h b/drivers/ufs/host/ufs-exynos.h
> index abe7e472759e..683b9150e2ba 100644
> --- a/drivers/ufs/host/ufs-exynos.h
> +++ b/drivers/ufs/host/ufs-exynos.h
> @@ -227,6 +227,7 @@ struct exynos_ufs {
> int avail_ln_rx;
> int avail_ln_tx;
> int rx_sel_idx;
> + bool phy_powered_on;
> struct ufs_pa_layer_attr dev_req_params;
> struct ufs_phy_time_cfg t_cfg;
> ktime_t entry_hibern8_t;
> --
> 2.43.0
>
^ permalink raw reply
* [PATCH net-next v2] net: phy: realtek: support MDI swapping for RTL8226-CG
From: Jan Hoffmann @ 2026-03-27 19:32 UTC (permalink / raw)
To: Andrew Lunn, Heiner Kallweit, Russell King, David S. Miller,
Eric Dumazet, Jakub Kicinski, Paolo Abeni, Daniel Golle,
Markus Stockhausen, Damien Dejean
Cc: netdev, linux-kernel, Jan Hoffmann
Add support for configuring swapping of MDI pairs (ABCD->DCBA) when the
property "enet-phy-pair-order" is specified.
Unfortunately, no documentation about this feature is available, but
this implementation still tries to avoid magic numbers and raw register
numbers where it seems clear what is going on. As it is unknown whether
the patching step can be safely reversed, only enabling MDI swapping is
supported.
Some other Realtek PHYs also support similar mechanisms:
- RTL8221B-VB-CG allows to configure MDI swapping via the same register,
but does not need the additional patching step. However, it is unclear
whether a driver implementation for that PHY is necessary, as it is
known to support configuration via strapping pins (which is working
fine at least in Zyxel XGS1210-12 rev B1).
- The patching step seems to match the one for the integrated PHYs of
some Realtek PCIe/USB NICs (see for example the r8152 driver).
For now, only implement this for the RTL8226-CG PHY, where it is needed
for the switches Zyxel XGS1010-12 rev A1 and XGS1210-12 rev A1.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
---
v2:
- Added register/field definitions where possible. For some registers,
it seems hard to find a proper name, but I still defined them where
some of the fields are known.
- Introduced new helper function to swap RG_LPF_CAP and RG_LPF_CAP_XG
values, in order to reduce duplication.
drivers/net/phy/realtek/realtek_main.c | 149 ++++++++++++++++++++++++-
1 file changed, 148 insertions(+), 1 deletion(-)
diff --git a/drivers/net/phy/realtek/realtek_main.c b/drivers/net/phy/realtek/realtek_main.c
index 023e47ad605b..2f007dc8ec8c 100644
--- a/drivers/net/phy/realtek/realtek_main.c
+++ b/drivers/net/phy/realtek/realtek_main.c
@@ -174,6 +174,20 @@
#define RTL8224_VND1_MDI_PAIR_SWAP 0xa90
#define RTL8224_VND1_MDI_POLARITY_SWAP 0xa94
+#define RTL8226_VND1_UNKNOWN_6A21 0x6a21
+#define RTL8226_VND1_UNKNOWN_6A21_MDI_SWAP_EN BIT(5)
+
+#define RTL8226_VND2_UNKNOWN_D068 0xd068
+#define RTL8226_VND2_UNKNOWN_D068_PAIR_SEL GENMASK(4, 3)
+#define RTL8226_VND2_ADCCAL_OFFSET 0xd06a
+
+#define RTL8226_VND2_RG_LPF_CAP_XG_P0_P1 0xbd5a
+#define RTL8226_VND2_RG_LPF_CAP_XG_P2_P3 0xbd5c
+#define RTL8226_VND2_RG_LPF_CAP_P0_P1 0xbc18
+#define RTL8226_VND2_RG_LPF_CAP_P2_P3 0xbc1a
+#define RTL8226_RG_LPF_CAP_PAIR_A_MASK GENMASK(4, 0)
+#define RTL8226_RG_LPF_CAP_PAIR_B_MASK GENMASK(12, 8)
+
#define RTL8366RB_POWER_SAVE 0x15
#define RTL8366RB_POWER_SAVE_ON BIT(12)
@@ -1447,6 +1461,139 @@ static unsigned int rtl822x_inband_caps(struct phy_device *phydev,
}
}
+static int rtl8226_set_mdi_swap(struct phy_device *phydev, bool swap_enable)
+{
+ u16 val = swap_enable ? RTL8226_VND1_UNKNOWN_6A21_MDI_SWAP_EN : 0;
+
+ return phy_modify_mmd(phydev, MDIO_MMD_VEND1, RTL8226_VND1_UNKNOWN_6A21,
+ RTL8226_VND1_UNKNOWN_6A21_MDI_SWAP_EN, val);
+}
+
+static int rtl8226_swap_rg_lpf_cap(struct phy_device *phydev, u32 reg_p0_p1, u32 reg_p2_p3)
+{
+ u16 val_p0, val_p1, val_p2, val_p3;
+ int ret;
+
+ ret = phy_read_mmd(phydev, MDIO_MMD_VEND2, reg_p0_p1);
+ if (ret < 0)
+ return ret;
+
+ val_p0 = FIELD_GET(RTL8226_RG_LPF_CAP_PAIR_A_MASK, ret);
+ val_p1 = FIELD_GET(RTL8226_RG_LPF_CAP_PAIR_B_MASK, ret);
+
+ ret = phy_read_mmd(phydev, MDIO_MMD_VEND2, reg_p2_p3);
+ if (ret < 0)
+ return ret;
+
+ val_p2 = FIELD_GET(RTL8226_RG_LPF_CAP_PAIR_A_MASK, ret);
+ val_p3 = FIELD_GET(RTL8226_RG_LPF_CAP_PAIR_B_MASK, ret);
+
+ ret = phy_modify_mmd(phydev, MDIO_MMD_VEND2, reg_p0_p1,
+ RTL8226_RG_LPF_CAP_PAIR_A_MASK | RTL8226_RG_LPF_CAP_PAIR_B_MASK,
+ FIELD_PREP(RTL8226_RG_LPF_CAP_PAIR_A_MASK, val_p3) |
+ FIELD_PREP(RTL8226_RG_LPF_CAP_PAIR_B_MASK, val_p2));
+ if (ret < 0)
+ return ret;
+
+ return phy_modify_mmd(phydev, MDIO_MMD_VEND2, reg_p2_p3,
+ RTL8226_RG_LPF_CAP_PAIR_A_MASK | RTL8226_RG_LPF_CAP_PAIR_B_MASK,
+ FIELD_PREP(RTL8226_RG_LPF_CAP_PAIR_A_MASK, val_p1) |
+ FIELD_PREP(RTL8226_RG_LPF_CAP_PAIR_B_MASK, val_p0));
+}
+
+static int rtl8226_patch_mdi_swap(struct phy_device *phydev)
+{
+ u16 adccal_offset[4];
+ int ret;
+
+ ret = phy_read_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_UNKNOWN_D068);
+ if (ret < 0)
+ return ret;
+
+ if (!(ret & BIT(1))) {
+ /* already swapped */
+ return 0;
+ }
+
+ /* The exact meaning of these bits is unknown. We only know that bit 1
+ * is used as a flag that swapping is already done.
+ */
+ ret = phy_modify_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_UNKNOWN_D068, 0x7, 0x1);
+ if (ret < 0)
+ return ret;
+
+ for (int i = 0; i < 4; i++) {
+ ret = phy_modify_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_UNKNOWN_D068,
+ RTL8226_VND2_UNKNOWN_D068_PAIR_SEL,
+ FIELD_PREP(RTL8226_VND2_UNKNOWN_D068_PAIR_SEL, i));
+ if (ret < 0)
+ return ret;
+
+ ret = phy_read_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_ADCCAL_OFFSET);
+ if (ret < 0)
+ return ret;
+
+ adccal_offset[i] = ret;
+ }
+
+ for (int i = 0; i < 4; i++) {
+ ret = phy_modify_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_UNKNOWN_D068,
+ RTL8226_VND2_UNKNOWN_D068_PAIR_SEL,
+ FIELD_PREP(RTL8226_VND2_UNKNOWN_D068_PAIR_SEL, i));
+ if (ret < 0)
+ return ret;
+
+ ret = phy_write_mmd(phydev, MDIO_MMD_VEND2, RTL8226_VND2_ADCCAL_OFFSET,
+ adccal_offset[3 - i]);
+ if (ret < 0)
+ return ret;
+ }
+
+ ret = rtl8226_swap_rg_lpf_cap(phydev, RTL8226_VND2_RG_LPF_CAP_XG_P0_P1,
+ RTL8226_VND2_RG_LPF_CAP_XG_P2_P3);
+ if (ret < 0)
+ return ret;
+
+ return rtl8226_swap_rg_lpf_cap(phydev, RTL8226_VND2_RG_LPF_CAP_P0_P1,
+ RTL8226_VND2_RG_LPF_CAP_P2_P3);
+}
+
+static int rtl8226_config_mdi_order(struct phy_device *phydev)
+{
+ u32 order;
+ int ret;
+
+ ret = of_property_read_u32(phydev->mdio.dev.of_node, "enet-phy-pair-order", &order);
+
+ /* Property not present, nothing to do */
+ if (ret == -EINVAL || ret == -ENOSYS)
+ return 0;
+
+ if (ret)
+ return ret;
+
+ /* Only enabling MDI swapping is supported */
+ if (order != 1)
+ return -EINVAL;
+
+ ret = rtl8226_set_mdi_swap(phydev, true);
+ if (ret)
+ return ret;
+
+ return rtl8226_patch_mdi_swap(phydev);
+}
+
+static int rtl8226_config_init(struct phy_device *phydev)
+{
+ int ret;
+
+ ret = rtl8226_config_mdi_order(phydev);
+ if (ret)
+ return ret;
+
+ return rtl822x_config_init(phydev);
+}
+
static int rtl822xb_get_rate_matching(struct phy_device *phydev,
phy_interface_t iface)
{
@@ -2384,7 +2531,7 @@ static struct phy_driver realtek_drvs[] = {
.soft_reset = rtl822x_c45_soft_reset,
.get_features = rtl822x_c45_get_features,
.config_aneg = rtl822x_c45_config_aneg,
- .config_init = rtl822x_config_init,
+ .config_init = rtl8226_config_init,
.inband_caps = rtl822x_inband_caps,
.config_inband = rtl822x_config_inband,
.read_status = rtl822xb_c45_read_status,
--
2.53.0
^ permalink raw reply related
* Re: [PATCH v6 phy-next 11/28] scsi: ufs: qcom: include missing <linux/interrupt.h>
From: Martin K. Petersen @ 2026-03-27 20:20 UTC (permalink / raw)
To: Vladimir Oltean
Cc: linux-phy, Vinod Koul, Neil Armstrong, dri-devel, freedreno,
linux-arm-kernel, linux-arm-msm, linux-can, linux-gpio, linux-ide,
linux-kernel, linux-media, linux-pci, linux-renesas-soc,
linux-riscv, linux-rockchip, linux-samsung-soc, linux-scsi,
linux-sunxi, linux-tegra, linux-usb, netdev, spacemit,
UNGLinuxDriver, Manivannan Sadhasivam, James E.J. Bottomley,
Martin K. Petersen
In-Reply-To: <20260327184706.1600329-12-vladimir.oltean@nxp.com>
Vladimir,
> The point is that <linux/phy/phy.h> will stop providing
> <linux/regulator/consumer.h>, and this would break the transitive
> include chain on armv7.
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
--
Martin K. Petersen
^ permalink raw reply
* Re: [PATCH v6 phy-next 09/28] scsi: ufs: exynos: stop poking into struct phy guts
From: Martin K. Petersen @ 2026-03-27 20:19 UTC (permalink / raw)
To: Vladimir Oltean
Cc: linux-phy, Vinod Koul, Neil Armstrong, dri-devel, freedreno,
linux-arm-kernel, linux-arm-msm, linux-can, linux-gpio, linux-ide,
linux-kernel, linux-media, linux-pci, linux-renesas-soc,
linux-riscv, linux-rockchip, linux-samsung-soc, linux-scsi,
linux-sunxi, linux-tegra, linux-usb, netdev, spacemit,
UNGLinuxDriver, Bart Van Assche, Alim Akhtar, Peter Griffin,
James E.J. Bottomley, Martin K. Petersen, Krzysztof Kozlowski,
Chanho Park
In-Reply-To: <20260327184706.1600329-10-vladimir.oltean@nxp.com>
Vladimir,
> The Exynos host controller driver is clearly a PHY consumer (gets the
> ufs->phy using devm_phy_get()), but pokes into the guts of struct phy
> to get the generic_phy->power_count.
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
--
Martin K. Petersen
^ permalink raw reply
* [PATCH net-next v10 10/10] bng_en: add support for ethtool -S stats display
From: Bhargava Marreddy @ 2026-03-27 20:13 UTC (permalink / raw)
To: davem, edumazet, kuba, pabeni, andrew+netdev, horms
Cc: netdev, linux-kernel, michael.chan, pavan.chebbi,
vsrama-krishna.nemani, vikas.gupta, Bhargava Marreddy
In-Reply-To: <20260327201343.16146-1-bhargava.marreddy@broadcom.com>
Implement the legacy ethtool statistics interface (get_sset_count,
get_strings, get_ethtool_stats) to expose hardware counters not
available through standard kernel stats APIs.
Ex:
a) Per-queue ring stats
rxq0_ucast_packets: 2
rxq0_mcast_packets: 0
rxq0_bcast_packets: 15
rxq0_ucast_bytes: 120
rxq0_mcast_bytes: 0
rxq0_bcast_bytes: 900
txq0_ucast_packets: 0
txq0_mcast_packets: 0
txq0_bcast_packets: 0
txq0_ucast_bytes: 0
txq0_mcast_bytes: 0
txq0_bcast_bytes: 0
b) Per-queue TPA(LRO/GRO) stats
rxq4_tpa_eligible_pkt: 0
rxq4_tpa_eligible_bytes: 0
rxq4_tpa_pkt: 0
rxq4_tpa_bytes: 0
rxq4_tpa_errors: 0
rxq4_tpa_events: 0
c) Port level stats
rxp_good_vlan_frames: 0
rxp_mtu_err_frames: 0
rxp_tagged_frames: 0
rxp_double_tagged_frames: 0
rxp_pfc_ena_frames_pri0: 0
rxp_pfc_ena_frames_pri1: 0
rxp_pfc_ena_frames_pri2: 0
rxp_pfc_ena_frames_pri3: 0
rxp_pfc_ena_frames_pri4: 0
rxp_pfc_ena_frames_pri5: 0
rxp_pfc_ena_frames_pri6: 0
rxp_pfc_ena_frames_pri7: 0
rxp_eee_lpi_events: 0
rxp_eee_lpi_duration: 0
rxp_runt_bytes: 0
rxp_runt_frames: 0
txp_good_vlan_frames: 0
txp_jabber_frames: 0
txp_fcs_err_frames: 0
txp_pfc_ena_frames_pri0: 0
txp_pfc_ena_frames_pri1: 0
txp_pfc_ena_frames_pri2: 0
txp_pfc_ena_frames_pri3: 0
txp_pfc_ena_frames_pri4: 0
txp_pfc_ena_frames_pri5: 0
txp_pfc_ena_frames_pri6: 0
txp_pfc_ena_frames_pri7: 0
txp_eee_lpi_events: 0
txp_eee_lpi_duration: 0
txp_xthol_frames: 0
d) Per-priority stats
rx_bytes_pri0: 4182650
rx_bytes_pri1: 4182650
rx_bytes_pri2: 4182650
rx_bytes_pri3: 4182650
rx_bytes_pri4: 4182650
rx_bytes_pri5: 4182650
rx_bytes_pri6: 4182650
rx_bytes_pri7: 4182650
Signed-off-by: Bhargava Marreddy <bhargava.marreddy@broadcom.com>
Reviewed-by: Vikas Gupta <vikas.gupta@broadcom.com>
---
.../net/ethernet/broadcom/bnge/bnge_ethtool.c | 481 ++++++++++++++++++
1 file changed, 481 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_ethtool.c b/drivers/net/ethernet/broadcom/bnge/bnge_ethtool.c
index 87226d5bf15c..2467e44de291 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_ethtool.c
@@ -35,6 +35,271 @@ static int bnge_nway_reset(struct net_device *dev)
return rc;
}
+static const char * const bnge_ring_q_stats_str[] = {
+ "ucast_packets",
+ "mcast_packets",
+ "bcast_packets",
+ "ucast_bytes",
+ "mcast_bytes",
+ "bcast_bytes",
+};
+
+static const char * const bnge_ring_tpa2_stats_str[] = {
+ "tpa_eligible_pkt",
+ "tpa_eligible_bytes",
+ "tpa_pkt",
+ "tpa_bytes",
+ "tpa_errors",
+ "tpa_events",
+};
+
+#define BNGE_RX_PORT_STATS_ENTRY(suffix) \
+ { BNGE_RX_STATS_OFFSET(rx_##suffix), "rxp_" __stringify(suffix) }
+
+#define BNGE_TX_PORT_STATS_ENTRY(suffix) \
+ { BNGE_TX_STATS_OFFSET(tx_##suffix), "txp_" __stringify(suffix) }
+
+#define BNGE_RX_STATS_EXT_ENTRY(counter) \
+ { BNGE_RX_STATS_EXT_OFFSET(counter), __stringify(counter) }
+
+#define BNGE_TX_STATS_EXT_ENTRY(counter) \
+ { BNGE_TX_STATS_EXT_OFFSET(counter), __stringify(counter) }
+
+#define BNGE_RX_STATS_EXT_PFC_ENTRY(n) \
+ BNGE_RX_STATS_EXT_ENTRY(pfc_pri##n##_rx_duration_us), \
+ BNGE_RX_STATS_EXT_ENTRY(pfc_pri##n##_rx_transitions)
+
+#define BNGE_TX_STATS_EXT_PFC_ENTRY(n) \
+ BNGE_TX_STATS_EXT_ENTRY(pfc_pri##n##_tx_duration_us), \
+ BNGE_TX_STATS_EXT_ENTRY(pfc_pri##n##_tx_transitions)
+
+#define BNGE_RX_STATS_EXT_PFC_ENTRIES \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(0), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(1), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(2), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(3), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(4), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(5), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(6), \
+ BNGE_RX_STATS_EXT_PFC_ENTRY(7)
+
+#define BNGE_TX_STATS_EXT_PFC_ENTRIES \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(0), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(1), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(2), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(3), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(4), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(5), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(6), \
+ BNGE_TX_STATS_EXT_PFC_ENTRY(7)
+
+#define BNGE_RX_STATS_EXT_COS_ENTRY(n) \
+ BNGE_RX_STATS_EXT_ENTRY(rx_bytes_cos##n), \
+ BNGE_RX_STATS_EXT_ENTRY(rx_packets_cos##n)
+
+#define BNGE_TX_STATS_EXT_COS_ENTRY(n) \
+ BNGE_TX_STATS_EXT_ENTRY(tx_bytes_cos##n), \
+ BNGE_TX_STATS_EXT_ENTRY(tx_packets_cos##n)
+
+#define BNGE_RX_STATS_EXT_COS_ENTRIES \
+ BNGE_RX_STATS_EXT_COS_ENTRY(0), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(1), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(2), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(3), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(4), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(5), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(6), \
+ BNGE_RX_STATS_EXT_COS_ENTRY(7) \
+
+#define BNGE_TX_STATS_EXT_COS_ENTRIES \
+ BNGE_TX_STATS_EXT_COS_ENTRY(0), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(1), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(2), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(3), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(4), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(5), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(6), \
+ BNGE_TX_STATS_EXT_COS_ENTRY(7) \
+
+#define BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(n) \
+ BNGE_RX_STATS_EXT_ENTRY(rx_discard_bytes_cos##n), \
+ BNGE_RX_STATS_EXT_ENTRY(rx_discard_packets_cos##n)
+
+#define BNGE_RX_STATS_EXT_DISCARD_COS_ENTRIES \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(0), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(1), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(2), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(3), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(4), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(5), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(6), \
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRY(7)
+
+#define BNGE_RX_STATS_PRI_ENTRY(counter, n) \
+ { BNGE_RX_STATS_EXT_OFFSET(counter##_cos0), \
+ __stringify(counter##_pri##n) }
+
+#define BNGE_TX_STATS_PRI_ENTRY(counter, n) \
+ { BNGE_TX_STATS_EXT_OFFSET(counter##_cos0), \
+ __stringify(counter##_pri##n) }
+
+#define BNGE_RX_STATS_PRI_ENTRIES(counter) \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 0), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 1), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 2), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 3), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 4), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 5), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 6), \
+ BNGE_RX_STATS_PRI_ENTRY(counter, 7)
+
+#define BNGE_TX_STATS_PRI_ENTRIES(counter) \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 0), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 1), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 2), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 3), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 4), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 5), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 6), \
+ BNGE_TX_STATS_PRI_ENTRY(counter, 7)
+
+#define NUM_RING_Q_HW_STATS ARRAY_SIZE(bnge_ring_q_stats_str)
+
+static const struct {
+ long offset;
+ char string[ETH_GSTRING_LEN];
+} bnge_tx_port_stats_ext_arr[] = {
+ BNGE_TX_STATS_EXT_COS_ENTRIES,
+ BNGE_TX_STATS_EXT_PFC_ENTRIES,
+};
+
+static const struct {
+ long base_off;
+ char string[ETH_GSTRING_LEN];
+} bnge_rx_bytes_pri_arr[] = {
+ BNGE_RX_STATS_PRI_ENTRIES(rx_bytes),
+};
+
+static const struct {
+ long base_off;
+ char string[ETH_GSTRING_LEN];
+} bnge_rx_pkts_pri_arr[] = {
+ BNGE_RX_STATS_PRI_ENTRIES(rx_packets),
+};
+
+static const struct {
+ long base_off;
+ char string[ETH_GSTRING_LEN];
+} bnge_tx_bytes_pri_arr[] = {
+ BNGE_TX_STATS_PRI_ENTRIES(tx_bytes),
+};
+
+static const struct {
+ long base_off;
+ char string[ETH_GSTRING_LEN];
+} bnge_tx_pkts_pri_arr[] = {
+ BNGE_TX_STATS_PRI_ENTRIES(tx_packets),
+};
+
+static const struct {
+ long offset;
+ char string[ETH_GSTRING_LEN];
+} bnge_port_stats_arr[] = {
+ BNGE_RX_PORT_STATS_ENTRY(good_vlan_frames),
+ BNGE_RX_PORT_STATS_ENTRY(mtu_err_frames),
+ BNGE_RX_PORT_STATS_ENTRY(tagged_frames),
+ BNGE_RX_PORT_STATS_ENTRY(double_tagged_frames),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri0),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri1),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri2),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri3),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri4),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri5),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri6),
+ BNGE_RX_PORT_STATS_ENTRY(pfc_ena_frames_pri7),
+ BNGE_RX_PORT_STATS_ENTRY(eee_lpi_events),
+ BNGE_RX_PORT_STATS_ENTRY(eee_lpi_duration),
+ BNGE_RX_PORT_STATS_ENTRY(runt_bytes),
+ BNGE_RX_PORT_STATS_ENTRY(runt_frames),
+
+ BNGE_TX_PORT_STATS_ENTRY(good_vlan_frames),
+ BNGE_TX_PORT_STATS_ENTRY(jabber_frames),
+ BNGE_TX_PORT_STATS_ENTRY(fcs_err_frames),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri0),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri1),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri2),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri3),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri4),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri5),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri6),
+ BNGE_TX_PORT_STATS_ENTRY(pfc_ena_frames_pri7),
+ BNGE_TX_PORT_STATS_ENTRY(eee_lpi_events),
+ BNGE_TX_PORT_STATS_ENTRY(eee_lpi_duration),
+ BNGE_TX_PORT_STATS_ENTRY(xthol_frames),
+};
+
+static const struct {
+ long offset;
+ char string[ETH_GSTRING_LEN];
+} bnge_port_stats_ext_arr[] = {
+ BNGE_RX_STATS_EXT_ENTRY(continuous_pause_events),
+ BNGE_RX_STATS_EXT_ENTRY(resume_pause_events),
+ BNGE_RX_STATS_EXT_ENTRY(continuous_roce_pause_events),
+ BNGE_RX_STATS_EXT_ENTRY(resume_roce_pause_events),
+ BNGE_RX_STATS_EXT_COS_ENTRIES,
+ BNGE_RX_STATS_EXT_PFC_ENTRIES,
+ BNGE_RX_STATS_EXT_ENTRY(rx_bits),
+ BNGE_RX_STATS_EXT_ENTRY(rx_buffer_passed_threshold),
+ BNGE_RX_STATS_EXT_DISCARD_COS_ENTRIES,
+ BNGE_RX_STATS_EXT_ENTRY(rx_filter_miss),
+};
+
+static int bnge_get_num_tpa_ring_stats(struct bnge_dev *bd)
+{
+ if (BNGE_SUPPORTS_TPA(bd))
+ return BNGE_NUM_TPA_RING_STATS;
+ return 0;
+}
+
+#define BNGE_NUM_PORT_STATS ARRAY_SIZE(bnge_port_stats_arr)
+#define BNGE_NUM_STATS_PRI \
+ (ARRAY_SIZE(bnge_rx_bytes_pri_arr) + \
+ ARRAY_SIZE(bnge_rx_pkts_pri_arr) + \
+ ARRAY_SIZE(bnge_tx_bytes_pri_arr) + \
+ ARRAY_SIZE(bnge_tx_pkts_pri_arr))
+
+static int bnge_get_num_ring_stats(struct bnge_dev *bd)
+{
+ int rx, tx;
+
+ rx = NUM_RING_Q_HW_STATS + bnge_get_num_tpa_ring_stats(bd);
+ tx = NUM_RING_Q_HW_STATS;
+ return rx * bd->rx_nr_rings +
+ tx * bd->tx_nr_rings_per_tc;
+}
+
+static u32 bnge_get_num_stats(struct bnge_net *bn)
+{
+ u32 num_stats = bnge_get_num_ring_stats(bn->bd);
+ u32 len;
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS)
+ num_stats += BNGE_NUM_PORT_STATS;
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS_EXT) {
+ len = min_t(u32, bn->fw_rx_stats_ext_size,
+ ARRAY_SIZE(bnge_port_stats_ext_arr));
+ num_stats += len;
+ len = min_t(u32, bn->fw_tx_stats_ext_size,
+ ARRAY_SIZE(bnge_tx_port_stats_ext_arr));
+ num_stats += len;
+ if (bn->pri2cos_valid)
+ num_stats += BNGE_NUM_STATS_PRI;
+ }
+
+ return num_stats;
+}
+
static void bnge_get_drvinfo(struct net_device *dev,
struct ethtool_drvinfo *info)
{
@@ -46,6 +311,219 @@ static void bnge_get_drvinfo(struct net_device *dev,
strscpy(info->bus_info, pci_name(bd->pdev), sizeof(info->bus_info));
}
+static int bnge_get_sset_count(struct net_device *dev, int sset)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+
+ switch (sset) {
+ case ETH_SS_STATS:
+ return bnge_get_num_stats(bn);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
+static bool is_rx_ring(struct bnge_dev *bd, u16 ring_num)
+{
+ return ring_num < bd->rx_nr_rings;
+}
+
+static bool is_tx_ring(struct bnge_dev *bd, u16 ring_num)
+{
+ u16 tx_base = 0;
+
+ if (!(bd->flags & BNGE_EN_SHARED_CHNL))
+ tx_base = bd->rx_nr_rings;
+
+ return ring_num >= tx_base && ring_num < (tx_base + bd->tx_nr_rings_per_tc);
+}
+
+static void bnge_get_ethtool_stats(struct net_device *dev,
+ struct ethtool_stats *stats, u64 *buf)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+ struct bnge_dev *bd = bn->bd;
+ u32 tpa_stats;
+ u32 i, j = 0;
+
+ if (!bn->bnapi) {
+ j += bnge_get_num_ring_stats(bd);
+ goto skip_ring_stats;
+ }
+
+ tpa_stats = bnge_get_num_tpa_ring_stats(bd);
+ for (i = 0; i < bd->nq_nr_rings; i++) {
+ struct bnge_napi *bnapi = bn->bnapi[i];
+ struct bnge_nq_ring_info *nqr;
+ u64 *sw_stats;
+ int k;
+
+ nqr = &bnapi->nq_ring;
+ sw_stats = nqr->stats.sw_stats;
+
+ if (is_rx_ring(bd, i)) {
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_ucast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_mcast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_bcast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_ucast_bytes);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_mcast_bytes);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, rx_bcast_bytes);
+ }
+ if (is_tx_ring(bd, i)) {
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_ucast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_mcast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_bcast_pkts);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_ucast_bytes);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_mcast_bytes);
+ buf[j++] = BNGE_GET_RING_STATS64(sw_stats, tx_bcast_bytes);
+ }
+ if (!tpa_stats || !is_rx_ring(bd, i))
+ continue;
+
+ k = BNGE_NUM_RX_RING_STATS + BNGE_NUM_TX_RING_STATS;
+ for (; k < BNGE_NUM_RX_RING_STATS + BNGE_NUM_TX_RING_STATS +
+ tpa_stats; j++, k++)
+ buf[j] = sw_stats[k];
+ }
+
+skip_ring_stats:
+ if (bn->flags & BNGE_FLAG_PORT_STATS) {
+ u64 *port_stats = bn->port_stats.sw_stats;
+
+ for (i = 0; i < BNGE_NUM_PORT_STATS; i++, j++)
+ buf[j] = *(port_stats + bnge_port_stats_arr[i].offset);
+ }
+ if (bn->flags & BNGE_FLAG_PORT_STATS_EXT) {
+ u64 *rx_port_stats_ext = bn->rx_port_stats_ext.sw_stats;
+ u64 *tx_port_stats_ext = bn->tx_port_stats_ext.sw_stats;
+ u32 len;
+
+ len = min_t(u32, bn->fw_rx_stats_ext_size,
+ ARRAY_SIZE(bnge_port_stats_ext_arr));
+ for (i = 0; i < len; i++, j++) {
+ buf[j] = *(rx_port_stats_ext +
+ bnge_port_stats_ext_arr[i].offset);
+ }
+ len = min_t(u32, bn->fw_tx_stats_ext_size,
+ ARRAY_SIZE(bnge_tx_port_stats_ext_arr));
+ for (i = 0; i < len; i++, j++) {
+ buf[j] = *(tx_port_stats_ext +
+ bnge_tx_port_stats_ext_arr[i].offset);
+ }
+ if (bn->pri2cos_valid) {
+ for (i = 0; i < 8; i++, j++) {
+ long n = bnge_rx_bytes_pri_arr[i].base_off +
+ bn->pri2cos_idx[i];
+
+ buf[j] = *(rx_port_stats_ext + n);
+ }
+ for (i = 0; i < 8; i++, j++) {
+ long n = bnge_rx_pkts_pri_arr[i].base_off +
+ bn->pri2cos_idx[i];
+
+ buf[j] = *(rx_port_stats_ext + n);
+ }
+ for (i = 0; i < 8; i++, j++) {
+ long n = bnge_tx_bytes_pri_arr[i].base_off +
+ bn->pri2cos_idx[i];
+
+ buf[j] = *(tx_port_stats_ext + n);
+ }
+ for (i = 0; i < 8; i++, j++) {
+ long n = bnge_tx_pkts_pri_arr[i].base_off +
+ bn->pri2cos_idx[i];
+
+ buf[j] = *(tx_port_stats_ext + n);
+ }
+ }
+ }
+}
+
+static void bnge_get_strings(struct net_device *dev, u32 stringset, u8 *buf)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+ struct bnge_dev *bd = bn->bd;
+ u32 i, j, num_str;
+ const char *str;
+
+ switch (stringset) {
+ case ETH_SS_STATS:
+ for (i = 0; i < bd->nq_nr_rings; i++) {
+ if (is_rx_ring(bd, i))
+ for (j = 0; j < NUM_RING_Q_HW_STATS; j++) {
+ str = bnge_ring_q_stats_str[j];
+ ethtool_sprintf(&buf, "rxq%d_%s", i,
+ str);
+ }
+ if (is_tx_ring(bd, i))
+ for (j = 0; j < NUM_RING_Q_HW_STATS; j++) {
+ str = bnge_ring_q_stats_str[j];
+ ethtool_sprintf(&buf, "txq%d_%s", i,
+ str);
+ }
+ num_str = bnge_get_num_tpa_ring_stats(bd);
+ if (!num_str || !is_rx_ring(bd, i))
+ continue;
+
+ for (j = 0; j < num_str; j++) {
+ str = bnge_ring_tpa2_stats_str[j];
+ ethtool_sprintf(&buf, "rxq%d_%s", i, str);
+ }
+ }
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS)
+ for (i = 0; i < BNGE_NUM_PORT_STATS; i++) {
+ str = bnge_port_stats_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS_EXT) {
+ u32 len;
+
+ len = min_t(u32, bn->fw_rx_stats_ext_size,
+ ARRAY_SIZE(bnge_port_stats_ext_arr));
+ for (i = 0; i < len; i++) {
+ str = bnge_port_stats_ext_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ len = min_t(u32, bn->fw_tx_stats_ext_size,
+ ARRAY_SIZE(bnge_tx_port_stats_ext_arr));
+ for (i = 0; i < len; i++) {
+ str = bnge_tx_port_stats_ext_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ if (bn->pri2cos_valid) {
+ for (i = 0; i < 8; i++) {
+ str = bnge_rx_bytes_pri_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ for (i = 0; i < 8; i++) {
+ str = bnge_rx_pkts_pri_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ for (i = 0; i < 8; i++) {
+ str = bnge_tx_bytes_pri_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+
+ for (i = 0; i < 8; i++) {
+ str = bnge_tx_pkts_pri_arr[i].string;
+ ethtool_puts(&buf, str);
+ }
+ }
+ }
+ break;
+ default:
+ netdev_err(bd->netdev, "%s invalid request %x\n",
+ __func__, stringset);
+ break;
+ }
+}
+
static void bnge_get_eth_phy_stats(struct net_device *dev,
struct ethtool_eth_phy_stats *phy_stats)
{
@@ -271,6 +749,9 @@ static const struct ethtool_ops bnge_ethtool_ops = {
.nway_reset = bnge_nway_reset,
.get_pauseparam = bnge_get_pauseparam,
.set_pauseparam = bnge_set_pauseparam,
+ .get_sset_count = bnge_get_sset_count,
+ .get_strings = bnge_get_strings,
+ .get_ethtool_stats = bnge_get_ethtool_stats,
.get_eth_phy_stats = bnge_get_eth_phy_stats,
.get_eth_mac_stats = bnge_get_eth_mac_stats,
.get_eth_ctrl_stats = bnge_get_eth_ctrl_stats,
--
2.47.3
^ permalink raw reply related
* [PATCH net-next v10 09/10] bng_en: implement netdev_stat_ops
From: Bhargava Marreddy @ 2026-03-27 20:13 UTC (permalink / raw)
To: davem, edumazet, kuba, pabeni, andrew+netdev, horms
Cc: netdev, linux-kernel, michael.chan, pavan.chebbi,
vsrama-krishna.nemani, vikas.gupta, Bhargava Marreddy
In-Reply-To: <20260327201343.16146-1-bhargava.marreddy@broadcom.com>
Implement netdev_stat_ops to provide standardized per-queue
statistics via the Netlink API.
Below is the description of the hardware drop counters:
rx-hw-drop-overruns: Packets dropped by HW due to resource limitations
(e.g., no BDs available in the host ring).
rx-hw-drops: Total packets dropped by HW (sum of overruns and error
drops).
tx-hw-drop-errors: Packets dropped by HW because they were invalid or
malformed.
tx-hw-drops: Total packets dropped by HW (sum of resource limitations
and error drops).
The implementation was verified using the ynl tool:
./tools/net/ynl/pyynl/cli.py --spec \
Documentation/netlink/specs/netdev.yaml --dump qstats-get --json \
'{"ifindex":14, "scope":"queue"}'
[{'ifindex': 14, 'queue-id': 0, 'queue-type': 'rx', 'rx-bytes': 758,
'rx-hw-drop-overruns': 0, 'rx-hw-drops': 0, 'rx-packets': 11},
{'ifindex': 14, 'queue-id': 1, 'queue-type': 'rx', 'rx-bytes': 0,
'rx-hw-drop-overruns': 0, 'rx-hw-drops': 0, 'rx-packets': 0},
{'ifindex': 14, 'queue-id': 0, 'queue-type': 'tx', 'tx-bytes': 0,
'tx-hw-drop-errors': 0, 'tx-hw-drops': 0, 'tx-packets': 0},
{'ifindex': 14, 'queue-id': 1, 'queue-type': 'tx', 'tx-bytes': 0,
'tx-hw-drop-errors': 0, 'tx-hw-drops': 0, 'tx-packets': 0},
{'ifindex': 14, 'queue-id': 2, 'queue-type': 'tx', 'tx-bytes': 810,
'tx-hw-drop-errors': 0, 'tx-hw-drops': 0, 'tx-packets': 10},]
Signed-off-by: Bhargava Marreddy <bhargava.marreddy@broadcom.com>
Reviewed-by: Vikas Gupta <vikas.gupta@broadcom.com>
---
.../net/ethernet/broadcom/bnge/bnge_netdev.c | 76 +++++++++++++++++++
1 file changed, 76 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
index b5cc96d1ee0e..7ffd8e8e70fb 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
@@ -3041,6 +3041,81 @@ static int bnge_close(struct net_device *dev)
return 0;
}
+static void bnge_get_queue_stats_rx(struct net_device *dev, int i,
+ struct netdev_queue_stats_rx *stats)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+ struct bnge_nq_ring_info *nqr;
+ u64 *sw;
+
+ if (!bn->bnapi)
+ return;
+
+ nqr = &bn->bnapi[i]->nq_ring;
+ sw = nqr->stats.sw_stats;
+
+ stats->packets = 0;
+ stats->packets += BNGE_GET_RING_STATS64(sw, rx_ucast_pkts);
+ stats->packets += BNGE_GET_RING_STATS64(sw, rx_mcast_pkts);
+ stats->packets += BNGE_GET_RING_STATS64(sw, rx_bcast_pkts);
+
+ stats->bytes = 0;
+ stats->bytes += BNGE_GET_RING_STATS64(sw, rx_ucast_bytes);
+ stats->bytes += BNGE_GET_RING_STATS64(sw, rx_mcast_bytes);
+ stats->bytes += BNGE_GET_RING_STATS64(sw, rx_bcast_bytes);
+
+ stats->hw_drop_overruns = BNGE_GET_RING_STATS64(sw, rx_discard_pkts);
+ stats->hw_drops = BNGE_GET_RING_STATS64(sw, rx_error_pkts) +
+ stats->hw_drop_overruns;
+}
+
+static void bnge_get_queue_stats_tx(struct net_device *dev, int i,
+ struct netdev_queue_stats_tx *stats)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+ struct bnge_napi *bnapi;
+ u64 *sw;
+
+ if (!bn->tx_ring)
+ return;
+
+ bnapi = bn->tx_ring[bn->tx_ring_map[i]].bnapi;
+ sw = bnapi->nq_ring.stats.sw_stats;
+
+ stats->packets = 0;
+ stats->packets += BNGE_GET_RING_STATS64(sw, tx_ucast_pkts);
+ stats->packets += BNGE_GET_RING_STATS64(sw, tx_mcast_pkts);
+ stats->packets += BNGE_GET_RING_STATS64(sw, tx_bcast_pkts);
+
+ stats->bytes = 0;
+ stats->bytes += BNGE_GET_RING_STATS64(sw, tx_ucast_bytes);
+ stats->bytes += BNGE_GET_RING_STATS64(sw, tx_mcast_bytes);
+ stats->bytes += BNGE_GET_RING_STATS64(sw, tx_bcast_bytes);
+
+ stats->hw_drop_errors = BNGE_GET_RING_STATS64(sw, tx_error_pkts);
+ stats->hw_drops = BNGE_GET_RING_STATS64(sw, tx_discard_pkts) +
+ stats->hw_drop_errors;
+}
+
+static void bnge_get_base_stats(struct net_device *dev,
+ struct netdev_queue_stats_rx *rx,
+ struct netdev_queue_stats_tx *tx)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+
+ rx->packets = bn->net_stats_prev.rx_packets;
+ rx->bytes = bn->net_stats_prev.rx_bytes;
+
+ tx->packets = bn->net_stats_prev.tx_packets;
+ tx->bytes = bn->net_stats_prev.tx_bytes;
+}
+
+static const struct netdev_stat_ops bnge_stat_ops = {
+ .get_queue_stats_rx = bnge_get_queue_stats_rx,
+ .get_queue_stats_tx = bnge_get_queue_stats_tx,
+ .get_base_stats = bnge_get_base_stats,
+};
+
static const struct net_device_ops bnge_netdev_ops = {
.ndo_open = bnge_open,
.ndo_stop = bnge_close,
@@ -3204,6 +3279,7 @@ int bnge_netdev_alloc(struct bnge_dev *bd, int max_irqs)
bd->netdev = netdev;
netdev->netdev_ops = &bnge_netdev_ops;
+ netdev->stat_ops = &bnge_stat_ops;
bnge_set_ethtool_ops(netdev);
--
2.47.3
^ permalink raw reply related
* [PATCH net-next v10 08/10] bng_en: implement ndo_get_stats64
From: Bhargava Marreddy @ 2026-03-27 20:13 UTC (permalink / raw)
To: davem, edumazet, kuba, pabeni, andrew+netdev, horms
Cc: netdev, linux-kernel, michael.chan, pavan.chebbi,
vsrama-krishna.nemani, vikas.gupta, Bhargava Marreddy
In-Reply-To: <20260327201343.16146-1-bhargava.marreddy@broadcom.com>
Implement the ndo_get_stats64 callback to report aggregate network
statistics. The driver gathers these by accumulating the per-ring
counters into the provided rtnl_link_stats64 structure.
Signed-off-by: Bhargava Marreddy <bhargava.marreddy@broadcom.com>
Reviewed-by: Vikas Gupta <vikas.gupta@broadcom.com>
---
.../net/ethernet/broadcom/bnge/bnge_netdev.c | 103 ++++++++++++++++++
.../net/ethernet/broadcom/bnge/bnge_netdev.h | 3 +
2 files changed, 106 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
index 3b9bcfbe6fdc..b5cc96d1ee0e 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
@@ -2872,6 +2872,7 @@ static int bnge_open_core(struct bnge_net *bn)
}
set_bit(BNGE_STATE_OPEN, &bd->state);
+ set_bit(BNGE_STATE_STATS_ENABLE, &bn->state);
bnge_enable_int(bn);
@@ -2910,6 +2911,99 @@ static int bnge_shutdown_nic(struct bnge_net *bn)
return 0;
}
+static void bnge_get_ring_stats(struct bnge_dev *bd,
+ struct rtnl_link_stats64 *stats)
+{
+ struct bnge_net *bn = netdev_priv(bd->netdev);
+ int i;
+
+ for (i = 0; i < bd->nq_nr_rings; i++) {
+ struct bnge_napi *bnapi = bn->bnapi[i];
+ struct bnge_nq_ring_info *nqr;
+ u64 *sw;
+
+ nqr = &bnapi->nq_ring;
+ sw = nqr->stats.sw_stats;
+
+ stats->rx_packets += BNGE_GET_RING_STATS64(sw, rx_ucast_pkts);
+ stats->rx_packets += BNGE_GET_RING_STATS64(sw, rx_mcast_pkts);
+ stats->rx_packets += BNGE_GET_RING_STATS64(sw, rx_bcast_pkts);
+
+ stats->tx_packets += BNGE_GET_RING_STATS64(sw, tx_ucast_pkts);
+ stats->tx_packets += BNGE_GET_RING_STATS64(sw, tx_mcast_pkts);
+ stats->tx_packets += BNGE_GET_RING_STATS64(sw, tx_bcast_pkts);
+
+ stats->rx_bytes += BNGE_GET_RING_STATS64(sw, rx_ucast_bytes);
+ stats->rx_bytes += BNGE_GET_RING_STATS64(sw, rx_mcast_bytes);
+ stats->rx_bytes += BNGE_GET_RING_STATS64(sw, rx_bcast_bytes);
+
+ stats->tx_bytes += BNGE_GET_RING_STATS64(sw, tx_ucast_bytes);
+ stats->tx_bytes += BNGE_GET_RING_STATS64(sw, tx_mcast_bytes);
+ stats->tx_bytes += BNGE_GET_RING_STATS64(sw, tx_bcast_bytes);
+
+ stats->rx_missed_errors +=
+ BNGE_GET_RING_STATS64(sw, rx_discard_pkts);
+
+ stats->multicast += BNGE_GET_RING_STATS64(sw, rx_mcast_pkts);
+ stats->tx_dropped += BNGE_GET_RING_STATS64(sw, tx_error_pkts);
+ }
+}
+
+static void bnge_add_prev_stats(struct bnge_net *bn,
+ struct rtnl_link_stats64 *stats)
+{
+ struct rtnl_link_stats64 *prev_stats = &bn->net_stats_prev;
+
+ stats->rx_packets += prev_stats->rx_packets;
+ stats->tx_packets += prev_stats->tx_packets;
+ stats->rx_bytes += prev_stats->rx_bytes;
+ stats->tx_bytes += prev_stats->tx_bytes;
+ stats->rx_missed_errors += prev_stats->rx_missed_errors;
+ stats->multicast += prev_stats->multicast;
+ stats->tx_dropped += prev_stats->tx_dropped;
+}
+
+static void bnge_get_stats64(struct net_device *dev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct bnge_net *bn = netdev_priv(dev);
+ struct bnge_dev *bd = bn->bd;
+
+ rcu_read_lock();
+ if (!test_bit(BNGE_STATE_STATS_ENABLE, &bn->state)) {
+ rcu_read_unlock();
+ *stats = bn->net_stats_prev;
+ return;
+ }
+
+ bnge_get_ring_stats(bd, stats);
+ bnge_add_prev_stats(bn, stats);
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS) {
+ u64 *tx = bn->port_stats.sw_stats +
+ BNGE_TX_PORT_STATS_BYTE_OFFSET / 8;
+ u64 *rx = bn->port_stats.sw_stats;
+
+ stats->rx_crc_errors =
+ BNGE_GET_RX_PORT_STATS64(rx, rx_fcs_err_frames);
+ stats->rx_frame_errors =
+ BNGE_GET_RX_PORT_STATS64(rx, rx_align_err_frames);
+ stats->rx_length_errors =
+ BNGE_GET_RX_PORT_STATS64(rx, rx_undrsz_frames) +
+ BNGE_GET_RX_PORT_STATS64(rx, rx_ovrsz_frames) +
+ BNGE_GET_RX_PORT_STATS64(rx, rx_runt_frames);
+ stats->rx_errors =
+ BNGE_GET_RX_PORT_STATS64(rx, rx_false_carrier_frames) +
+ BNGE_GET_RX_PORT_STATS64(rx, rx_jbr_frames);
+ stats->collisions =
+ BNGE_GET_TX_PORT_STATS64(tx, tx_total_collisions);
+ stats->tx_fifo_errors =
+ BNGE_GET_TX_PORT_STATS64(tx, tx_fifo_underruns);
+ stats->tx_errors = BNGE_GET_TX_PORT_STATS64(tx, tx_err);
+ }
+ rcu_read_unlock();
+}
+
static void bnge_close_core(struct bnge_net *bn)
{
struct bnge_dev *bd = bn->bd;
@@ -2921,6 +3015,14 @@ static void bnge_close_core(struct bnge_net *bn)
timer_delete_sync(&bn->timer);
bnge_shutdown_nic(bn);
bnge_disable_napi(bn);
+
+ /* Save ring stats before shutdown */
+ if (bn->bnapi)
+ bnge_get_ring_stats(bd, &bn->net_stats_prev);
+
+ clear_bit(BNGE_STATE_STATS_ENABLE, &bn->state);
+ synchronize_rcu();
+
bnge_free_all_rings_bufs(bn);
bnge_free_irq(bn);
bnge_del_napi(bn);
@@ -2943,6 +3045,7 @@ static const struct net_device_ops bnge_netdev_ops = {
.ndo_open = bnge_open,
.ndo_stop = bnge_close,
.ndo_start_xmit = bnge_start_xmit,
+ .ndo_get_stats64 = bnge_get_stats64,
.ndo_features_check = bnge_features_check,
};
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
index 548c5a77e76a..c972567882bf 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
@@ -262,6 +262,7 @@ struct bnge_stats_mem {
enum bnge_net_state {
BNGE_STATE_NAPI_DISABLED,
+ BNGE_STATE_STATS_ENABLE,
};
#define BNGE_TIMER_INTERVAL HZ
@@ -353,6 +354,8 @@ struct bnge_net {
u64 flags;
+ struct rtnl_link_stats64 net_stats_prev;
+
struct bnge_stats_mem port_stats;
struct bnge_stats_mem rx_port_stats_ext;
struct bnge_stats_mem tx_port_stats_ext;
--
2.47.3
^ permalink raw reply related
* [PATCH net-next v10 07/10] bng_en: periodically fetch and accumulate hardware statistics
From: Bhargava Marreddy @ 2026-03-27 20:13 UTC (permalink / raw)
To: davem, edumazet, kuba, pabeni, andrew+netdev, horms
Cc: netdev, linux-kernel, michael.chan, pavan.chebbi,
vsrama-krishna.nemani, vikas.gupta, Bhargava Marreddy,
Rahul Gupta, Ajit Kumar Khaparde
In-Reply-To: <20260327201343.16146-1-bhargava.marreddy@broadcom.com>
Use the timer to schedule periodic stats collection via
the workqueue when the link is up. Fetch fresh counters from
hardware via DMA and accumulate them into 64-bit software
shadows, handling wrap-around for counters narrower than
64 bits.
Signed-off-by: Bhargava Marreddy <bhargava.marreddy@broadcom.com>
Reviewed-by: Vikas Gupta <vikas.gupta@broadcom.com>
Reviewed-by: Rahul Gupta <rahul-rg.gupta@broadcom.com>
Reviewed-by: Ajit Kumar Khaparde <ajit.khaparde@broadcom.com>
---
.../net/ethernet/broadcom/bnge/bnge_netdev.c | 90 +++++++++++++++++++
.../net/ethernet/broadcom/bnge/bnge_netdev.h | 1 +
2 files changed, 91 insertions(+)
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
index 23a07f640847..3b9bcfbe6fdc 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.c
@@ -313,9 +313,93 @@ static void bnge_timer(struct timer_list *t)
}
}
+ if (BNGE_LINK_IS_UP(bd) && bn->stats_coal_ticks)
+ bnge_queue_sp_work(bn, BNGE_PERIODIC_STATS_SP_EVENT);
+
mod_timer(&bn->timer, jiffies + bn->current_interval);
}
+static void bnge_add_one_ctr(u64 hw, u64 *sw, u64 mask)
+{
+ u64 sw_tmp, sw_val;
+
+ hw &= mask;
+ sw_val = READ_ONCE(*sw);
+ sw_tmp = (sw_val & ~mask) | hw;
+ if (hw < (sw_val & mask))
+ sw_tmp += mask + 1;
+ WRITE_ONCE(*sw, sw_tmp);
+}
+
+static void __bnge_accumulate_stats(__le64 *hw_stats, u64 *sw_stats, u64 *masks,
+ int count)
+{
+ int i;
+
+ for (i = 0; i < count; i++) {
+ u64 hw = le64_to_cpu(READ_ONCE(hw_stats[i]));
+
+ if (masks[i] == -1ULL)
+ WRITE_ONCE(sw_stats[i], hw);
+ else
+ bnge_add_one_ctr(hw, &sw_stats[i], masks[i]);
+ }
+}
+
+static void bnge_accumulate_stats(struct bnge_stats_mem *stats)
+{
+ if (!stats->hw_stats)
+ return;
+
+ __bnge_accumulate_stats(stats->hw_stats, stats->sw_stats,
+ stats->hw_masks, stats->len / 8);
+}
+
+static void bnge_accumulate_all_stats(struct bnge_dev *bd)
+{
+ struct bnge_net *bn = netdev_priv(bd->netdev);
+ struct bnge_stats_mem *ring0_stats = NULL;
+ int i;
+
+ for (i = 0; i < bd->nq_nr_rings; i++) {
+ struct bnge_napi *bnapi = bn->bnapi[i];
+ struct bnge_nq_ring_info *nqr;
+ struct bnge_stats_mem *stats;
+
+ nqr = &bnapi->nq_ring;
+ stats = &nqr->stats;
+
+ if (!ring0_stats)
+ ring0_stats = &bn->bnapi[0]->nq_ring.stats;
+
+ __bnge_accumulate_stats(stats->hw_stats, stats->sw_stats,
+ ring0_stats->hw_masks,
+ ring0_stats->len / 8);
+ }
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS) {
+ struct bnge_stats_mem *stats = &bn->port_stats;
+ __le64 *hw_stats = stats->hw_stats;
+ u64 *sw_stats = stats->sw_stats;
+ u64 *masks = stats->hw_masks;
+ u16 cnt;
+
+ cnt = sizeof(struct rx_port_stats) / 8;
+ __bnge_accumulate_stats(hw_stats, sw_stats, masks, cnt);
+
+ hw_stats += BNGE_TX_PORT_STATS_BYTE_OFFSET / 8;
+ sw_stats += BNGE_TX_PORT_STATS_BYTE_OFFSET / 8;
+ masks += BNGE_TX_PORT_STATS_BYTE_OFFSET / 8;
+ cnt = sizeof(struct tx_port_stats) / 8;
+ __bnge_accumulate_stats(hw_stats, sw_stats, masks, cnt);
+ }
+
+ if (bn->flags & BNGE_FLAG_PORT_STATS_EXT) {
+ bnge_accumulate_stats(&bn->rx_port_stats_ext);
+ bnge_accumulate_stats(&bn->tx_port_stats_ext);
+ }
+}
+
static void bnge_sp_task(struct work_struct *work)
{
struct bnge_net *bn = container_of(work, struct bnge_net, sp_task);
@@ -328,6 +412,12 @@ static void bnge_sp_task(struct work_struct *work)
return;
}
+ if (test_and_clear_bit(BNGE_PERIODIC_STATS_SP_EVENT, &bn->sp_event)) {
+ bnge_hwrm_port_qstats(bd, 0);
+ bnge_hwrm_port_qstats_ext(bd, 0);
+ bnge_accumulate_all_stats(bd);
+ }
+
if (test_and_clear_bit(BNGE_UPDATE_PHY_SP_EVENT, &bn->sp_event)) {
int rc;
diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
index 74b08d492c53..548c5a77e76a 100644
--- a/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
+++ b/drivers/net/ethernet/broadcom/bnge/bnge_netdev.h
@@ -276,6 +276,7 @@ enum bnge_sp_event {
BNGE_LINK_SPEED_CHNG_SP_EVENT,
BNGE_LINK_CFG_CHANGE_SP_EVENT,
BNGE_UPDATE_PHY_SP_EVENT,
+ BNGE_PERIODIC_STATS_SP_EVENT,
};
struct bnge_net {
--
2.47.3
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox