From: Guangbin Huang <huangguangbin2@huawei.com>
To: <davem@davemloft.net>, <kuba@kernel.org>
Cc: <netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<lipeng321@huawei.com>, <huangguangbin2@huawei.com>,
<chenhao288@hisilicon.com>
Subject: [PATCH net 1/6] net: hns3: fix ethtool tx copybreak buf size indicating not aligned issue
Date: Sat, 26 Mar 2022 17:51:00 +0800 [thread overview]
Message-ID: <20220326095105.54075-2-huangguangbin2@huawei.com> (raw)
In-Reply-To: <20220326095105.54075-1-huangguangbin2@huawei.com>
From: Hao Chen <chenhao288@hisilicon.com>
When use ethtoool set tx copybreak buf size to a large value
which causes order exceeding 10 or memory is not enough,
it causes allocating tx copybreak buffer failed and print
"the active tx spare buf is 0, not enabled tx spare buffer",
however, use --get-tunable parameter query tx copybreak buf
size and it indicates setting value not 0.
So, it's necessary to change the print value from setting
value to 0.
Set kinfo.tx_spare_buf_size to 0 when set tx copybreak buf size failed.
Fixes: e445f08af2b1 ("net: hns3: add support to set/get tx copybreak buf size via ethtool for hns3 driver")
Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
---
.../net/ethernet/hisilicon/hns3/hns3_enet.c | 20 +++++++++++--------
.../ethernet/hisilicon/hns3/hns3_ethtool.c | 3 ++-
2 files changed, 14 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index 0b8a73c40b12..16137238ddbf 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -1028,13 +1028,12 @@ static bool hns3_can_use_tx_sgl(struct hns3_enet_ring *ring,
static void hns3_init_tx_spare_buffer(struct hns3_enet_ring *ring)
{
+ u32 alloc_size = ring->tqp->handle->kinfo.tx_spare_buf_size;
struct hns3_tx_spare *tx_spare;
struct page *page;
- u32 alloc_size;
dma_addr_t dma;
int order;
- alloc_size = ring->tqp->handle->kinfo.tx_spare_buf_size;
if (!alloc_size)
return;
@@ -1044,30 +1043,35 @@ static void hns3_init_tx_spare_buffer(struct hns3_enet_ring *ring)
if (!tx_spare) {
/* The driver still work without the tx spare buffer */
dev_warn(ring_to_dev(ring), "failed to allocate hns3_tx_spare\n");
- return;
+ goto devm_kzalloc_error;
}
page = alloc_pages_node(dev_to_node(ring_to_dev(ring)),
GFP_KERNEL, order);
if (!page) {
dev_warn(ring_to_dev(ring), "failed to allocate tx spare pages\n");
- devm_kfree(ring_to_dev(ring), tx_spare);
- return;
+ goto alloc_pages_error;
}
dma = dma_map_page(ring_to_dev(ring), page, 0,
PAGE_SIZE << order, DMA_TO_DEVICE);
if (dma_mapping_error(ring_to_dev(ring), dma)) {
dev_warn(ring_to_dev(ring), "failed to map pages for tx spare\n");
- put_page(page);
- devm_kfree(ring_to_dev(ring), tx_spare);
- return;
+ goto dma_mapping_error;
}
tx_spare->dma = dma;
tx_spare->buf = page_address(page);
tx_spare->len = PAGE_SIZE << order;
ring->tx_spare = tx_spare;
+ return;
+
+dma_mapping_error:
+ put_page(page);
+alloc_pages_error:
+ devm_kfree(ring_to_dev(ring), tx_spare);
+devm_kzalloc_error:
+ ring->tqp->handle->kinfo.tx_spare_buf_size = 0;
}
/* Use hns3_tx_spare_space() to make sure there is enough buffer
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
index 6469238ae090..ae30dbe7ef52 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
@@ -1818,7 +1818,8 @@ static int hns3_set_tunable(struct net_device *netdev,
old_tx_spare_buf_size = h->kinfo.tx_spare_buf_size;
new_tx_spare_buf_size = *(u32 *)data;
ret = hns3_set_tx_spare_buf_size(netdev, new_tx_spare_buf_size);
- if (ret) {
+ if (ret ||
+ (!priv->ring->tx_spare && new_tx_spare_buf_size != 0)) {
int ret1;
netdev_warn(netdev,
--
2.33.0
next prev parent reply other threads:[~2022-03-26 9:56 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-26 9:50 [PATCH net 0/6] net: hns3: add some fixes for -net Guangbin Huang
2022-03-26 9:51 ` Guangbin Huang [this message]
2022-03-26 9:51 ` [PATCH net 2/6] net: hns3: add max order judgement for tx spare buffer Guangbin Huang
2022-03-26 9:51 ` [PATCH net 3/6] net: hns3: clean residual vf config after disable sriov Guangbin Huang
2022-03-26 9:51 ` [PATCH net 4/6] net: hns3: add netdev reset check for hns3_set_tunable() Guangbin Huang
2022-03-26 9:51 ` [PATCH net 5/6] net: hns3: add NULL pointer check for hns3_set/get_ringparam() Guangbin Huang
2022-03-26 9:51 ` [PATCH net 6/6] net: hns3: fix phy can not link up when autoneg off and reset Guangbin Huang
2022-03-26 16:20 ` [PATCH net 0/6] net: hns3: add some fixes for -net 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=20220326095105.54075-2-huangguangbin2@huawei.com \
--to=huangguangbin2@huawei.com \
--cc=chenhao288@hisilicon.com \
--cc=davem@davemloft.net \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lipeng321@huawei.com \
--cc=netdev@vger.kernel.org \
/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
all inboxes | Powered by JetHome®