netdev - Re: [RFC net-next 1/8] pkt_cls: make tc_can_offload

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20180124082844.GA2234@nanopsycho>
Date:   Wed, 24 Jan 2018 09:28:44 +0100
From:   Jiri Pirko <jiri@...nulli.us>
To:     Jakub Kicinski <jakub.kicinski@...ronome.com>
Cc:     netdev@...r.kernel.org, oss-drivers@...ronome.com,
        dsahern@...il.com
Subject: Re: [RFC net-next 1/8] pkt_cls: make tc_can_offload_extack() check
 chain index

Tue, Jan 23, 2018 at 10:33:33PM CET, jakub.kicinski@...ronome.com wrote:
>No upstream drivers seem to allow offload of chains other than 0.

mlxsw does. And I know that Intel was talking about adding the support
to i40e iirc.


>Save driver developers typing and make tc_can_offload_extack()
>check for that condition as well.  Rename the function to
>tc_can_offload_cls() to better represent its application.
>
>Signed-off-by: Jakub Kicinski <jakub.kicinski@...ronome.com>
>---

[...]


>--- a/include/net/pkt_cls.h
>+++ b/include/net/pkt_cls.h
>@@ -645,15 +645,21 @@ static inline bool tc_can_offload(const struct net_device *dev)
> 	return dev->features & NETIF_F_HW_TC;
> }
> 
>-static inline bool tc_can_offload_extack(const struct net_device *dev,
>-					 struct netlink_ext_ack *extack)
>+static inline bool tc_can_offload_cls(const struct net_device *dev,
>+				      struct tc_cls_common_offload *common)
> {
>-	bool can = tc_can_offload(dev);
>-
>-	if (!can)
>-		NL_SET_ERR_MSG(extack, "TC offload is disabled on net device");
>+	if (common->chain_index) {

This check here is wrong. It is up to the driver if it supports
multichain offload or not, should not be checked in a generic code along
with tc_can_offload.


>+		NL_SET_ERR_MSG(common->extack,
>+			       "Driver supports only offload of chain 0");
>+		return false;
>+	}
>+	if (!tc_can_offload(dev)) {
>+		NL_SET_ERR_MSG(common->extack,
>+			       "TC offload is disabled on net device");
>+		return false;
>+	}
> 
>-	return can;
>+	return true;
> }
> 
> static inline bool tc_skip_hw(u32 flags)
>-- 
>2.15.1
>