txgbe register clk which name is i2c_designware.pci_dev_id(),
clk_name will be stored in clk_lookup_alloc. If PCIe bus number
is larger than 0x39, clk_name size will be larger than 20 bytes.
It exceeds clk_lookup_alloc MAX_DEV_ID limits. So the driver
shortened clk_name.
Fixes: b63f20485e43 ("net: txgbe: Register fixed rate clock")
Signed-off-by: Duanqiang Wen <[email protected]>
---
drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
index 93295916b1d2..5b5d5e4310d1 100644
--- a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
+++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
@@ -571,7 +571,7 @@ static int txgbe_clock_register(struct txgbe *txgbe)
char clk_name[32];
struct clk *clk;
- snprintf(clk_name, sizeof(clk_name), "i2c_designware.%d",
+ snprintf(clk_name, sizeof(clk_name), "i2c_dw.%d",
pci_dev_id(pdev));
clk = clk_register_fixed_rate(NULL, clk_name, NULL, 0, 156250000);
--
2.27.0
On Wed, Mar 13, 2024 at 04:06:34PM +0800, Duanqiang Wen wrote:
> txgbe register clk which name is i2c_designware.pci_dev_id(),
> clk_name will be stored in clk_lookup_alloc. If PCIe bus number
> is larger than 0x39, clk_name size will be larger than 20 bytes.
> It exceeds clk_lookup_alloc MAX_DEV_ID limits. So the driver
> shortened clk_name.
>
> Fixes: b63f20485e43 ("net: txgbe: Register fixed rate clock")
> Signed-off-by: Duanqiang Wen <[email protected]>
> ---
> drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
> index 93295916b1d2..5b5d5e4310d1 100644
> --- a/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
> +++ b/drivers/net/ethernet/wangxun/txgbe/txgbe_phy.c
> @@ -571,7 +571,7 @@ static int txgbe_clock_register(struct txgbe *txgbe)
> char clk_name[32];
> struct clk *clk;
>
> - snprintf(clk_name, sizeof(clk_name), "i2c_designware.%d",
> + snprintf(clk_name, sizeof(clk_name), "i2c_dw.%d",
> pci_dev_id(pdev));
>
> clk = clk_register_fixed_rate(NULL, clk_name, NULL, 0, 156250000);
> --
> 2.27.0
>
>
I believe the patch is targeted to the "net" tree?
If so, the fix looks OK to me.
Thanks,
Reviewed-by: Michal Kubiak <[email protected]>
On Wed, 2024-03-13 at 16:06 +0800, Duanqiang Wen wrote:
> txgbe register clk which name is i2c_designware.pci_dev_id(),
> clk_name will be stored in clk_lookup_alloc. If PCIe bus number
> is larger than 0x39, clk_name size will be larger than 20 bytes.
> It exceeds clk_lookup_alloc MAX_DEV_ID limits. So the driver
> shortened clk_name.
>
> Fixes: b63f20485e43 ("net: txgbe: Register fixed rate clock")
> Signed-off-by: Duanqiang Wen <[email protected]>
Unrelated to this patch, but more related to the self-test burst this
patch landed into: it looks like that since roughly the net-next PR was
merged back forwarding debug variant of the vxlan-bridge-1q-ipv6-sh and
vxlan-bridge-1d-port-8472-ipv6-sh tests become less stable.
IDK if that was due to some glitch in the AWS hosts or what else. Just
an head-up.
Cheers,
Paolo
Hello:
This patch was applied to netdev/net.git (main)
by Paolo Abeni <[email protected]>:
On Wed, 13 Mar 2024 16:06:34 +0800 you wrote:
> txgbe register clk which name is i2c_designware.pci_dev_id(),
> clk_name will be stored in clk_lookup_alloc. If PCIe bus number
> is larger than 0x39, clk_name size will be larger than 20 bytes.
> It exceeds clk_lookup_alloc MAX_DEV_ID limits. So the driver
> shortened clk_name.
>
> Fixes: b63f20485e43 ("net: txgbe: Register fixed rate clock")
> Signed-off-by: Duanqiang Wen <[email protected]>
>
> [...]
Here is the summary with links:
- net: txgbe: fix clk_name exceed MAX_DEV_ID limits
https://git.kernel.org/netdev/net/c/e30cef001da2
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
On Thu, Mar 14, 2024 at 01:46:01PM +0100, Paolo Abeni wrote:
> Unrelated to this patch, but more related to the self-test burst this
> patch landed into: it looks like that since roughly the net-next PR was
> merged back forwarding debug variant of the vxlan-bridge-1q-ipv6-sh and
> vxlan-bridge-1d-port-8472-ipv6-sh tests become less stable.
>
> IDK if that was due to some glitch in the AWS hosts or what else. Just
> an head-up.
Looked into it. The test is sending 100 pings with an interval of 0.1
seconds and using a timeout of 11 seconds which is borderline,
especially with a debug kernel. Will bump it to 20 seconds for good
measures. It won't prolong the test unless something is actually wrong,
in which case the test will rightfully fail.