2019-07-09 03:32:51

by Jiangfeng Xiao

[permalink] [raw]
Subject: [PATCH v2 00/10] net: hisilicon: Add support for HI13X1 to hip04_eth

The main purpose of this patch series is to extend the
hip04_eth driver to support HI13X1_GMAC.

The offset and bitmap of some registers of HI13X1_GMAC
are different from hip04_eth common soc. In addition,
the definition of send descriptor and parsing descriptor
are different from hip04_eth common soc. So the macro
of the register offset is redefined to adapt the HI13X1_GMAC.

Clean up the sparse warning by the way.

Change since v1:
* Add a cover letter.

Jiangfeng Xiao (10):
net: hisilicon: Add support for HI13X1 to hip04_eth
net: hisilicon: Cleanup for got restricted __be32
net: hisilicon: Cleanup for cast to restricted __be32
net: hisilicon: HI13X1_GMAX skip write LOCAL_PAGE_REG
net: hisilicon: HI13X1_GMAX need dreq reset at first
net: hisilicon: dt-bindings: Add an field of port-handle
net: hisilicon: Add group field to adapt HI13X1_GMAC
net: hisilicon: Offset buf address to adapt HI13X1_GMAC
net: hisilicon: Add an rx_desc to adapt HI13X1_GMAC
net: hisilicon: Add an tx_desc to adapt HI13X1_GMAC

.../bindings/net/hisilicon-hip04-net.txt | 7 +-
drivers/net/ethernet/hisilicon/Kconfig | 10 ++
drivers/net/ethernet/hisilicon/hip04_eth.c | 142 ++++++++++++++++++---
3 files changed, 136 insertions(+), 23 deletions(-)

--
1.8.5.6


2019-07-09 03:33:20

by Jiangfeng Xiao

[permalink] [raw]
Subject: [PATCH v2 07/10] net: hisilicon: Add group field to adapt HI13X1_GMAC

In general, group is the same as the port, but some
boards specify a special group for better load
balancing of each processing unit.

Signed-off-by: Jiangfeng Xiao <[email protected]>
---
drivers/net/ethernet/hisilicon/hip04_eth.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
index 19d8cfd..5328219 100644
--- a/drivers/net/ethernet/hisilicon/hip04_eth.c
+++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
@@ -178,6 +178,7 @@ struct hip04_priv {
int phy_mode;
int chan;
unsigned int port;
+ unsigned int group;
unsigned int speed;
unsigned int duplex;
unsigned int reg_inten;
@@ -278,10 +279,10 @@ static void hip04_config_fifo(struct hip04_priv *priv)
val |= PPE_CFG_STS_RX_PKT_CNT_RC;
writel_relaxed(val, priv->base + PPE_CFG_STS_MODE);

- val = BIT(priv->port);
+ val = BIT(priv->group);
regmap_write(priv->map, priv->port * 4 + PPE_CFG_POOL_GRP, val);

- val = priv->port << PPE_CFG_QOS_VMID_GRP_SHIFT;
+ val = priv->group << PPE_CFG_QOS_VMID_GRP_SHIFT;
val |= PPE_CFG_QOS_VMID_MODE;
writel_relaxed(val, priv->base + PPE_CFG_QOS_VMID_GEN);

@@ -876,7 +877,7 @@ static int hip04_mac_probe(struct platform_device *pdev)
}
#endif

- ret = of_parse_phandle_with_fixed_args(node, "port-handle", 2, 0, &arg);
+ ret = of_parse_phandle_with_fixed_args(node, "port-handle", 3, 0, &arg);
if (ret < 0) {
dev_warn(d, "no port-handle\n");
goto init_fail;
@@ -884,6 +885,7 @@ static int hip04_mac_probe(struct platform_device *pdev)

priv->port = arg.args[0];
priv->chan = arg.args[1] * RX_DESC_NUM;
+ priv->group = arg.args[2];

hrtimer_init(&priv->tx_coalesce_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);

--
1.8.5.6

2019-07-09 03:33:32

by Jiangfeng Xiao

[permalink] [raw]
Subject: [PATCH v2 05/10] net: hisilicon: HI13X1_GMAX need dreq reset at first

HI13X1_GMAC delete request for soft reset at first,
otherwise, the subsequent initialization will not
take effect.

Signed-off-by: Jiangfeng Xiao <[email protected]>
---
drivers/net/ethernet/hisilicon/hip04_eth.c | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)

diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
index fe61b01..19d8cfd 100644
--- a/drivers/net/ethernet/hisilicon/hip04_eth.c
+++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
@@ -16,6 +16,8 @@
#include <linux/mfd/syscon.h>
#include <linux/regmap.h>

+#define SC_PPE_RESET_DREQ 0x026C
+
#define PPE_CFG_RX_ADDR 0x100
#define PPE_CFG_POOL_GRP 0x300
#define PPE_CFG_RX_BUF_SIZE 0x400
@@ -61,6 +63,8 @@

#define PPE_HIS_RX_PKT_CNT 0x804

+#define RESET_DREQ_ALL 0xffffffff
+
/* REG_INTERRUPT */
#define RCV_INT BIT(10)
#define RCV_NOBUF BIT(8)
@@ -168,6 +172,9 @@ struct rx_desc {

struct hip04_priv {
void __iomem *base;
+#if defined(CONFIG_HI13X1_GMAC)
+ void __iomem *sysctrl_base;
+#endif
int phy_mode;
int chan;
unsigned int port;
@@ -244,6 +251,13 @@ static void hip04_config_port(struct net_device *ndev, u32 speed, u32 duplex)
writel_relaxed(val, priv->base + GE_MODE_CHANGE_REG);
}

+static void hip04_reset_dreq(struct hip04_priv *priv)
+{
+#if defined(CONFIG_HI13X1_GMAC)
+ writel_relaxed(RESET_DREQ_ALL, priv->sysctrl_base + SC_PPE_RESET_DREQ);
+#endif
+}
+
static void hip04_reset_ppe(struct hip04_priv *priv)
{
u32 val, tmp, timeout = 0;
@@ -853,6 +867,15 @@ static int hip04_mac_probe(struct platform_device *pdev)
goto init_fail;
}

+#if defined(CONFIG_HI13X1_GMAC)
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+ priv->sysctrl_base = devm_ioremap_resource(d, res);
+ if (IS_ERR(priv->sysctrl_base)) {
+ ret = PTR_ERR(priv->sysctrl_base);
+ goto init_fail;
+ }
+#endif
+
ret = of_parse_phandle_with_fixed_args(node, "port-handle", 2, 0, &arg);
if (ret < 0) {
dev_warn(d, "no port-handle\n");
@@ -921,6 +944,7 @@ static int hip04_mac_probe(struct platform_device *pdev)
ndev->irq = irq;
netif_napi_add(ndev, &priv->napi, hip04_rx_poll, NAPI_POLL_WEIGHT);

+ hip04_reset_dreq(priv);
hip04_reset_ppe(priv);
if (priv->phy_mode == PHY_INTERFACE_MODE_MII)
hip04_config_port(ndev, SPEED_100, DUPLEX_FULL);
--
1.8.5.6

2019-07-09 03:34:11

by Jiangfeng Xiao

[permalink] [raw]
Subject: [PATCH v2 04/10] net: hisilicon: HI13X1_GMAX skip write LOCAL_PAGE_REG

HI13X1_GMAC changed the offsets and bitmaps for
GE_TX_LOCAL_PAGE_REG registers in the same peripheral
device on different models of the hip04_eth. With the
default configuration, HI13X1_GMAC can also work without
any writes to the GE_TX_LOCAL_PAGE_REG register.

Signed-off-by: Jiangfeng Xiao <[email protected]>
---
drivers/net/ethernet/hisilicon/hip04_eth.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
index d8f0619..fe61b01 100644
--- a/drivers/net/ethernet/hisilicon/hip04_eth.c
+++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
@@ -308,8 +308,10 @@ static void hip04_config_fifo(struct hip04_priv *priv)
val |= GE_RX_STRIP_PAD | GE_RX_PAD_EN;
writel_relaxed(val, priv->base + GE_RECV_CONTROL_REG);

+#ifndef CONFIG_HI13X1_GMAC
val = GE_AUTO_NEG_CTL;
writel_relaxed(val, priv->base + GE_TX_LOCAL_PAGE_REG);
+#endif
}

static void hip04_mac_enable(struct net_device *ndev)
--
1.8.5.6

2019-07-09 03:34:41

by Jiangfeng Xiao

[permalink] [raw]
Subject: [PATCH v2 06/10] net: hisilicon: dt-bindings: Add an field of port-handle

In general, group is the same as the port, but some
boards specify a special group for better load
balancing of each processing unit.

Signed-off-by: Jiangfeng Xiao <[email protected]>
---
Documentation/devicetree/bindings/net/hisilicon-hip04-net.txt | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/hisilicon-hip04-net.txt b/Documentation/devicetree/bindings/net/hisilicon-hip04-net.txt
index d1df8a0..464c0da 100644
--- a/Documentation/devicetree/bindings/net/hisilicon-hip04-net.txt
+++ b/Documentation/devicetree/bindings/net/hisilicon-hip04-net.txt
@@ -10,6 +10,7 @@ Required properties:
phandle, specifies a reference to the syscon ppe node
port, port number connected to the controller
channel, recv channel start from channel * number (RX_DESC_NUM)
+ group, field in the pkg desc, in general, it is the same as the port.
- phy-mode: see ethernet.txt [1].

Optional properties:
@@ -66,7 +67,7 @@ Example:
reg = <0x28b0000 0x10000>;
interrupts = <0 413 4>;
phy-mode = "mii";
- port-handle = <&ppe 31 0>;
+ port-handle = <&ppe 31 0 31>;
};

ge0: ethernet@2800000 {
@@ -74,7 +75,7 @@ Example:
reg = <0x2800000 0x10000>;
interrupts = <0 402 4>;
phy-mode = "sgmii";
- port-handle = <&ppe 0 1>;
+ port-handle = <&ppe 0 1 0>;
phy-handle = <&phy0>;
};

@@ -83,6 +84,6 @@ Example:
reg = <0x2880000 0x10000>;
interrupts = <0 410 4>;
phy-mode = "sgmii";
- port-handle = <&ppe 8 2>;
+ port-handle = <&ppe 8 2 8>;
phy-handle = <&phy1>;
};
--
1.8.5.6

2019-07-09 09:37:54

by Sergei Shtylyov

[permalink] [raw]
Subject: Re: [PATCH v2 05/10] net: hisilicon: HI13X1_GMAX need dreq reset at first

Hello!

On 09.07.2019 6:31, Jiangfeng Xiao wrote:

> HI13X1_GMAC delete request for soft reset at first,
> otherwise, the subsequent initialization will not
> take effect.
>
> Signed-off-by: Jiangfeng Xiao <[email protected]>
> ---
> drivers/net/ethernet/hisilicon/hip04_eth.c | 24 ++++++++++++++++++++++++
> 1 file changed, 24 insertions(+)
>
> diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
> index fe61b01..19d8cfd 100644
> --- a/drivers/net/ethernet/hisilicon/hip04_eth.c
> +++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
[...]
> @@ -853,6 +867,15 @@ static int hip04_mac_probe(struct platform_device *pdev)
> goto init_fail;
> }
>
> +#if defined(CONFIG_HI13X1_GMAC)
> + res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
> + priv->sysctrl_base = devm_ioremap_resource(d, res);

There's devm_platform_ioremap_resource() now.

> + if (IS_ERR(priv->sysctrl_base)) {
> + ret = PTR_ERR(priv->sysctrl_base);
> + goto init_fail;
> + }
> +#endif
> +
> ret = of_parse_phandle_with_fixed_args(node, "port-handle", 2, 0, &arg);
> if (ret < 0) {
> dev_warn(d, "no port-handle\n");
[...]

MBR, Sergei

2019-07-09 13:49:59

by Jiangfeng Xiao

[permalink] [raw]
Subject: Re: [PATCH v2 05/10] net: hisilicon: HI13X1_GMAX need dreq reset at first



On 2019/7/9 17:35, Sergei Shtylyov wrote:
> Hello!
>
> On 09.07.2019 6:31, Jiangfeng Xiao wrote:
>
>> HI13X1_GMAC delete request for soft reset at first,
>> otherwise, the subsequent initialization will not
>> take effect.
>>
>> Signed-off-by: Jiangfeng Xiao <[email protected]>
>> ---
>> drivers/net/ethernet/hisilicon/hip04_eth.c | 24 ++++++++++++++++++++++++
>> 1 file changed, 24 insertions(+)
>>
>> diff --git a/drivers/net/ethernet/hisilicon/hip04_eth.c b/drivers/net/ethernet/hisilicon/hip04_eth.c
>> index fe61b01..19d8cfd 100644
>> --- a/drivers/net/ethernet/hisilicon/hip04_eth.c
>> +++ b/drivers/net/ethernet/hisilicon/hip04_eth.c
> [...]
>> @@ -853,6 +867,15 @@ static int hip04_mac_probe(struct platform_device *pdev)
>> goto init_fail;
>> }
>> +#if defined(CONFIG_HI13X1_GMAC)
>> + res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
>> + priv->sysctrl_base = devm_ioremap_resource(d, res);
>
> There's devm_platform_ioremap_resource() now.

Thank you for your review, Great issue, which makes my code more concise.

I will fix it in v3. Or submit a patch to modify it separately, if maintainer
applies this patch series.

2019-07-09 21:30:41

by David Miller

[permalink] [raw]
Subject: Re: [PATCH v2 00/10] net: hisilicon: Add support for HI13X1 to hip04_eth

From: Jiangfeng Xiao <[email protected]>
Date: Tue, 9 Jul 2019 11:31:01 +0800

> The main purpose of this patch series is to extend the
> hip04_eth driver to support HI13X1_GMAC.
>
> The offset and bitmap of some registers of HI13X1_GMAC
> are different from hip04_eth common soc. In addition,
> the definition of send descriptor and parsing descriptor
> are different from hip04_eth common soc. So the macro
> of the register offset is redefined to adapt the HI13X1_GMAC.
>
> Clean up the sparse warning by the way.
>
> Change since v1:
> * Add a cover letter.

Series applied, thanks.

2019-07-10 05:06:39

by Jiangfeng Xiao

[permalink] [raw]
Subject: Re: [PATCH v2 05/10] net: hisilicon: HI13X1_GMAX need dreq reset at first



On 2019/7/9 21:48, Jiangfeng Xiao wrote:
>
>
> On 2019/7/9 17:35, Sergei Shtylyov wrote:
>> Hello!
>>
>> On 09.07.2019 6:31, Jiangfeng Xiao wrote:
>>
[...]
>>> @@ -853,6 +867,15 @@ static int hip04_mac_probe(struct platform_device *pdev)
>>> goto init_fail;
>>> }
>>> +#if defined(CONFIG_HI13X1_GMAC)
>>> + res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
>>> + priv->sysctrl_base = devm_ioremap_resource(d, res);
>>
>> There's devm_platform_ioremap_resource() now.
>
> Thank you for your review, Great issue, which makes my code more concise.
>
> I will fix it in v3. Or submit a patch to modify it separately, if maintainer
> applies this patch series.
>
I decided to wait for this series of patches to sync to the mainline
and then fix this based on the mainline.

Because the mainline does not currently have this part of the code,
if I submit the changes, and the patch is accidentally merged into
another branch or another maintainer to handle, a conflict will occur.

As we all know, maintianer has to deal with many commits every day,
I don't want to increase the burden of maintainer.

So I decided to wait until the patch is synced to the mainline
and then modify it, which is more safe.