From: Jakub Kicinski <kuba@kernel.org>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, edumazet@google.com, pabeni@redhat.com,
michael.chan@broadcom.com, shuah@kernel.org,
ecree.xilinx@gmail.com, przemyslaw.kitszel@intel.com,
ahmed.zaki@intel.com, andrew@lunn.ch, willemb@google.com,
pavan.chebbi@broadcom.com, petrm@nvidia.com,
Jakub Kicinski <kuba@kernel.org>
Subject: [PATCH net 4/5] ethtool: fix the state of additional contexts with old API
Date: Thu, 25 Jul 2024 15:23:52 -0700 [thread overview]
Message-ID: <20240725222353.2993687-5-kuba@kernel.org> (raw)
In-Reply-To: <20240725222353.2993687-1-kuba@kernel.org>
We expect drivers implementing the new create/modify/destroy
API to populate the defaults in struct ethtool_rxfh_context.
In legacy API ctx isn't even passed, and rxfh.indir / rxfh.key
are NULL so drivers can't give us defaults even if they want to.
Call get_rxfh() to fetch the values. We can reuse rxfh_dev
for the get_rxfh(), rxfh stores the input from the user.
This fixes IOCTL reporting 0s instead of the default key /
indir table for drivers using legacy API.
Add a check to try to catch drivers using the new API
but not populating the key.
Fixes: 7964e7884643 ("net: ethtool: use the tracking array for get_rxfh on custom RSS contexts")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
net/ethtool/ioctl.c | 38 ++++++++++++++++++++++++++++++--------
1 file changed, 30 insertions(+), 8 deletions(-)
diff --git a/net/ethtool/ioctl.c b/net/ethtool/ioctl.c
index a37ba113610a..8ca13208d240 100644
--- a/net/ethtool/ioctl.c
+++ b/net/ethtool/ioctl.c
@@ -1382,10 +1382,9 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
rxfh.input_xfrm == RXH_XFRM_NO_CHANGE))
return -EINVAL;
- if (rxfh.indir_size != ETH_RXFH_INDIR_NO_CHANGE)
- indir_bytes = dev_indir_size * sizeof(rxfh_dev.indir[0]);
+ indir_bytes = dev_indir_size * sizeof(rxfh_dev.indir[0]);
- rss_config = kzalloc(indir_bytes + rxfh.key_size, GFP_USER);
+ rss_config = kzalloc(indir_bytes + dev_key_size, GFP_USER);
if (!rss_config)
return -ENOMEM;
@@ -1475,16 +1474,21 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
rxfh_dev.input_xfrm = rxfh.input_xfrm;
if (rxfh.rss_context && ops->create_rxfh_context) {
- if (create)
+ if (create) {
ret = ops->create_rxfh_context(dev, ctx, &rxfh_dev,
extack);
- else if (rxfh_dev.rss_delete)
+ /* Make sure driver populates defaults */
+ WARN_ON_ONCE(!ret && !rxfh_dev.key &&
+ !memchr_inv(ethtool_rxfh_context_key(ctx),
+ 0, ctx->key_size));
+ } else if (rxfh_dev.rss_delete) {
ret = ops->remove_rxfh_context(dev, ctx,
rxfh.rss_context,
extack);
- else
+ } else {
ret = ops->modify_rxfh_context(dev, ctx, &rxfh_dev,
extack);
+ }
} else {
ret = ops->set_rxfh(dev, &rxfh_dev, extack);
}
@@ -1523,6 +1527,22 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
kfree(ctx);
goto out;
}
+
+ /* Fetch the defaults for the old API, in the new API drivers
+ * should write defaults into ctx themselves.
+ */
+ rxfh_dev.indir = (u32 *)rss_config;
+ rxfh_dev.indir_size = dev_indir_size;
+
+ rxfh_dev.key = rss_config + indir_bytes;
+ rxfh_dev.key_size = dev_key_size;
+
+ ret = ops->get_rxfh(dev, &rxfh_dev);
+ if (WARN_ON(ret)) {
+ xa_erase(&dev->ethtool->rss_ctx, rxfh.rss_context);
+ kfree(ctx);
+ goto out;
+ }
}
if (rxfh_dev.rss_delete) {
WARN_ON(xa_erase(&dev->ethtool->rss_ctx, rxfh.rss_context) != ctx);
@@ -1531,12 +1551,14 @@ static noinline_for_stack int ethtool_set_rxfh(struct net_device *dev,
if (rxfh_dev.indir) {
for (i = 0; i < dev_indir_size; i++)
ethtool_rxfh_context_indir(ctx)[i] = rxfh_dev.indir[i];
- ctx->indir_configured = 1;
+ ctx->indir_configured =
+ rxfh.indir_size &&
+ rxfh.indir_size != ETH_RXFH_INDIR_NO_CHANGE;
}
if (rxfh_dev.key) {
memcpy(ethtool_rxfh_context_key(ctx), rxfh_dev.key,
dev_key_size);
- ctx->key_configured = 1;
+ ctx->key_configured = !!rxfh.key_size;
}
if (rxfh_dev.hfunc != ETH_RSS_HASH_NO_CHANGE)
ctx->hfunc = rxfh_dev.hfunc;
--
2.45.2
next prev parent reply other threads:[~2024-07-25 22:24 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-07-25 22:23 [PATCH net 0/5] ethtool: more RSS fixes Jakub Kicinski
2024-07-25 22:23 ` [PATCH net 1/5] eth: bnxt: reject unsupported hash functions Jakub Kicinski
2024-07-26 6:16 ` Pavan Chebbi
2024-07-25 22:23 ` [PATCH net 2/5] eth: bnxt: populate defaults in the RSS context struct Jakub Kicinski
2024-07-26 6:15 ` Pavan Chebbi
2024-07-25 22:23 ` [PATCH net 3/5] ethtool: fix setting key and resetting indir at once Jakub Kicinski
2024-07-25 22:23 ` Jakub Kicinski [this message]
2024-07-26 5:06 ` [PATCH net 4/5] ethtool: fix the state of additional contexts with old API Edward Cree
2024-07-25 22:23 ` [PATCH net 5/5] selftests: drv-net: rss_ctx: check for all-zero keys Jakub Kicinski
2024-07-26 11:25 ` Petr Machata
2024-07-29 10:08 ` [PATCH net 0/5] ethtool: more RSS fixes patchwork-bot+netdevbpf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240725222353.2993687-5-kuba@kernel.org \
--to=kuba@kernel.org \
--cc=ahmed.zaki@intel.com \
--cc=andrew@lunn.ch \
--cc=davem@davemloft.net \
--cc=ecree.xilinx@gmail.com \
--cc=edumazet@google.com \
--cc=michael.chan@broadcom.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=pavan.chebbi@broadcom.com \
--cc=petrm@nvidia.com \
--cc=przemyslaw.kitszel@intel.com \
--cc=shuah@kernel.org \
--cc=willemb@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).