Hi,
According to David's remark I split the original patchset into two. First
one, which I submit is a list of fixes, all stable-CC'ed.
On the occasion two new bugs were fixed - wrong define in one of registers
and, more important, an error path in mvneta_rx() after failing of skb
build.
I'm looking forward to the comments or remarks.
Best regards,
Marcin Wojtas
Changes from v1:
* update MBUS windows access protection register once, after whole loop
* add fixing value of MVNETA_RXQ_INTR_ENABLE_ALL_MASK
* add fixing error path for skb_build()
* add possibility of setting custom TX IP checksum limit in DT property
Marcin Wojtas (6):
net: mvneta: add configuration for MBUS windows access protection
net: mvneta: fix bit assignment in MVNETA_RXQ_CONFIG_REG
net: mvneta: fix bit assignment for RX packet irq enable
net: mvneta: fix error path for building skb
net: mvneta: enable setting custom TX IP checksum limit
net: mvneta: enable IP checksum with jumbo frames for Armada 38x on
Port0
.../bindings/net/marvell-armada-370-neta.txt | 6 ++++
arch/arm/boot/dts/armada-38x.dtsi | 1 +
drivers/net/ethernet/marvell/mvneta.c | 34 ++++++++++++++++++----
3 files changed, 35 insertions(+), 6 deletions(-)
--
1.8.3.1
This commit adds missing configuration of MBUS windows access protection
in mvneta_conf_mbus_windows function - a dedicated variable for that
purpose remained there unused since v3.8 initial mvneta support. Because
of that the register contents were inherited from the bootloader.
Signed-off-by: Marcin Wojtas <[email protected]>
Reviewed-by: Gregory CLEMENT <[email protected]>
Cc: <[email protected]> # v3.8+
Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
unit")
---
drivers/net/ethernet/marvell/mvneta.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index e84c7f2..c4327b9 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -62,6 +62,7 @@
#define MVNETA_WIN_SIZE(w) (0x2204 + ((w) << 3))
#define MVNETA_WIN_REMAP(w) (0x2280 + ((w) << 2))
#define MVNETA_BASE_ADDR_ENABLE 0x2290
+#define MVNETA_ACCESS_PROTECT_ENABLE 0x2294
#define MVNETA_PORT_CONFIG 0x2400
#define MVNETA_UNI_PROMISC_MODE BIT(0)
#define MVNETA_DEF_RXQ(q) ((q) << 1)
@@ -3188,9 +3189,11 @@ static void mvneta_conf_mbus_windows(struct mvneta_port *pp,
win_enable &= ~(1 << i);
win_protect |= 3 << (2 * i);
+
}
mvreg_write(pp, MVNETA_BASE_ADDR_ENABLE, win_enable);
+ mvreg_write(pp, MVNETA_ACCESS_PROTECT_ENABLE, win_protect);
}
/* Power up the port */
--
1.8.3.1
MVNETA_RXQ_HW_BUF_ALLOC bit which controls enabling hardware buffer
allocation was mistakenly set as BIT(1). This commit fixes the assignment.
Signed-off-by: Marcin Wojtas <[email protected]>
Reviewed-by: Gregory CLEMENT <[email protected]>
Cc: <[email protected]> # v3.8+
Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
unit")
---
drivers/net/ethernet/marvell/mvneta.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index c4327b9..d78edbb 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -36,7 +36,7 @@
/* Registers */
#define MVNETA_RXQ_CONFIG_REG(q) (0x1400 + ((q) << 2))
-#define MVNETA_RXQ_HW_BUF_ALLOC BIT(1)
+#define MVNETA_RXQ_HW_BUF_ALLOC BIT(0)
#define MVNETA_RXQ_PKT_OFFSET_ALL_MASK (0xf << 8)
#define MVNETA_RXQ_PKT_OFFSET_MASK(offs) ((offs) << 8)
#define MVNETA_RXQ_THRESHOLD_REG(q) (0x14c0 + ((q) << 2))
--
1.8.3.1
A value originally defined in the driver was inappropriate. Even though
the ingress was somehow working, writing MVNETA_RXQ_INTR_ENABLE_ALL_MASK
to MVNETA_INTR_ENABLE didn't make any effect, because the bits [31:16]
are reserved and read-only.
This commit updates MVNETA_RXQ_INTR_ENABLE_ALL_MASK to be compliant with
the controller's documentation.
Signed-off-by: Marcin Wojtas <[email protected]>
Cc: <[email protected]> # v3.8+
Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
unit")
---
drivers/net/ethernet/marvell/mvneta.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index d78edbb..0c3d923 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -160,7 +160,7 @@
#define MVNETA_INTR_ENABLE 0x25b8
#define MVNETA_TXQ_INTR_ENABLE_ALL_MASK 0x0000ff00
-#define MVNETA_RXQ_INTR_ENABLE_ALL_MASK 0xff000000 // note: neta says it's 0x000000FF
+#define MVNETA_RXQ_INTR_ENABLE_ALL_MASK 0x000000ff
#define MVNETA_RXQ_CMD 0x2680
#define MVNETA_RXQ_DISABLE_SHIFT 8
--
1.8.3.1
In the actual RX processing, there is same error path for both descriptor
ring refilling and building skb fails. This is not correct, because after
successful refill, the ring is already updated with newly allocated
buffer. Then, in case of build_skb() fail, hitherto code left the original
buffer unmapped.
This patch fixes above situation by swapping error check of skb build with
DMA-unmap of original buffer.
Signed-off-by: Marcin Wojtas <[email protected]>
Cc: <[email protected]> # v4.2+
Fixes a84e32894191 ("net: mvneta: fix refilling for Rx DMA buffers")
---
drivers/net/ethernet/marvell/mvneta.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index 0c3d923..62cf971 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -1580,12 +1580,16 @@ static int mvneta_rx(struct mvneta_port *pp, int rx_todo,
}
skb = build_skb(data, pp->frag_size > PAGE_SIZE ? 0 : pp->frag_size);
- if (!skb)
- goto err_drop_frame;
+ /* After refill old buffer has to be unmapped regardless
+ * the skb is successfully built or not.
+ */
dma_unmap_single(dev->dev.parent, phys_addr,
MVNETA_RX_BUF_SIZE(pp->pkt_size), DMA_FROM_DEVICE);
+ if (!skb)
+ goto err_drop_frame;
+
rcvd_pkts++;
rcvd_bytes += rx_bytes;
--
1.8.3.1
Since Armada 38x SoC can support IP checksum for jumbo frames only on
a single port, it means that this feature should be enabled per-port,
rather than for the whole SoC.
This patch enables setting custom TX IP checksum limit by adding new
optional property to the mvneta device tree node. If not used, by
default 1600B is set for "marvell,armada-370-neta" and 9800B for other
strings, which ensures backward compatibility. Binding documentation
is updated accordingly.
Signed-off-by: Marcin Wojtas <[email protected]>
Cc: <[email protected]> # v3.18+
---
.../bindings/net/marvell-armada-370-neta.txt | 6 ++++++
drivers/net/ethernet/marvell/mvneta.c | 19 +++++++++++++++++--
2 files changed, 23 insertions(+), 2 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/marvell-armada-370-neta.txt b/Documentation/devicetree/bindings/net/marvell-armada-370-neta.txt
index f5a8ca2..aeea50c 100644
--- a/Documentation/devicetree/bindings/net/marvell-armada-370-neta.txt
+++ b/Documentation/devicetree/bindings/net/marvell-armada-370-neta.txt
@@ -8,6 +8,11 @@ Required properties:
- phy-mode: See ethernet.txt file in the same directory
- clocks: a pointer to the reference clock for this device.
+Optional properties:
+- tx-csum-limit: maximum mtu supported by port that allow TX checksum.
+ Value is presented in bytes. If not used, by default 1600B is set for
+ "marvell,armada-370-neta" and 9800B for others.
+
Example:
ethernet@d0070000 {
@@ -15,6 +20,7 @@ ethernet@d0070000 {
reg = <0xd0070000 0x2500>;
interrupts = <8>;
clocks = <&gate_clk 4>;
+ tx-csum-limit = <9800>
status = "okay";
phy = <&phy0>;
phy-mode = "rgmii-id";
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index 62cf971..32d6f28 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -243,6 +243,7 @@
#define MVNETA_VLAN_TAG_LEN 4
#define MVNETA_CPU_D_CACHE_LINE_SIZE 32
+#define MVNETA_TX_CSUM_DEF_SIZE 1600
#define MVNETA_TX_CSUM_MAX_SIZE 9800
#define MVNETA_ACC_MODE_EXT 1
@@ -3257,6 +3258,7 @@ static int mvneta_probe(struct platform_device *pdev)
char hw_mac_addr[ETH_ALEN];
const char *mac_from;
const char *managed;
+ int tx_csum_limit;
int phy_mode;
int err;
int cpu;
@@ -3357,8 +3359,21 @@ static int mvneta_probe(struct platform_device *pdev)
}
}
- if (of_device_is_compatible(dn, "marvell,armada-370-neta"))
- pp->tx_csum_limit = 1600;
+ if (!of_property_read_u32(dn, "tx-csum-limit", &tx_csum_limit)) {
+ if (tx_csum_limit < 0 ||
+ tx_csum_limit > MVNETA_TX_CSUM_MAX_SIZE) {
+ tx_csum_limit = MVNETA_TX_CSUM_DEF_SIZE;
+ dev_info(&pdev->dev,
+ "Wrong TX csum limit in DT, set to %dB\n",
+ MVNETA_TX_CSUM_DEF_SIZE);
+ }
+ } else if (of_device_is_compatible(dn, "marvell,armada-370-neta")) {
+ tx_csum_limit = MVNETA_TX_CSUM_DEF_SIZE;
+ } else {
+ tx_csum_limit = MVNETA_TX_CSUM_MAX_SIZE;
+ }
+
+ pp->tx_csum_limit = tx_csum_limit;
pp->tx_ring_size = MVNETA_MAX_TXD;
pp->rx_ring_size = MVNETA_MAX_RXD;
--
1.8.3.1
The Ethernet controller found in the Armada 38x SoC's family support
TCP/IP checksumming with frame sizes larger than 1600 bytes, however
only on port 0.
This commit enables it by setting 'tx-csum-limit' to 9800B in
'ethernet@70000' node.
Signed-off-by: Marcin Wojtas <[email protected]>
Cc: <[email protected]> # v3.18+
---
arch/arm/boot/dts/armada-38x.dtsi | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/arm/boot/dts/armada-38x.dtsi b/arch/arm/boot/dts/armada-38x.dtsi
index c6a0e9d..e8b7f67 100644
--- a/arch/arm/boot/dts/armada-38x.dtsi
+++ b/arch/arm/boot/dts/armada-38x.dtsi
@@ -498,6 +498,7 @@
reg = <0x70000 0x4000>;
interrupts-extended = <&mpic 8>;
clocks = <&gateclk 4>;
+ tx-csum-limit = <9800>;
status = "disabled";
};
--
1.8.3.1
Marcin,
On Thu, 26 Nov 2015 19:08:08 +0100, Marcin Wojtas wrote:
> This commit adds missing configuration of MBUS windows access protection
> in mvneta_conf_mbus_windows function - a dedicated variable for that
> purpose remained there unused since v3.8 initial mvneta support. Because
> of that the register contents were inherited from the bootloader.
>
> Signed-off-by: Marcin Wojtas <[email protected]>
> Reviewed-by: Gregory CLEMENT <[email protected]>
> Cc: <[email protected]> # v3.8+
> Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
> unit")
Is this actually fixing a user visible problem? If not, then I am not
sure it qualifies for stable.
> diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
> index e84c7f2..c4327b9 100644
> --- a/drivers/net/ethernet/marvell/mvneta.c
> +++ b/drivers/net/ethernet/marvell/mvneta.c
> @@ -62,6 +62,7 @@
> #define MVNETA_WIN_SIZE(w) (0x2204 + ((w) << 3))
> #define MVNETA_WIN_REMAP(w) (0x2280 + ((w) << 2))
> #define MVNETA_BASE_ADDR_ENABLE 0x2290
> +#define MVNETA_ACCESS_PROTECT_ENABLE 0x2294
Spaces are used for indentation in the definitions of all other
constants in this driver, so I think you should stick with this
convention, at least to remain consistent.
> #define MVNETA_PORT_CONFIG 0x2400
> #define MVNETA_UNI_PROMISC_MODE BIT(0)
> #define MVNETA_DEF_RXQ(q) ((q) << 1)
> @@ -3188,9 +3189,11 @@ static void mvneta_conf_mbus_windows(struct mvneta_port *pp,
>
> win_enable &= ~(1 << i);
> win_protect |= 3 << (2 * i);
> +
Unneeded change.
> }
>
> mvreg_write(pp, MVNETA_BASE_ADDR_ENABLE, win_enable);
> + mvreg_write(pp, MVNETA_ACCESS_PROTECT_ENABLE, win_protect);
> }
Thanks,
Thomas
--
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
Dear Marcin Wojtas,
On Thu, 26 Nov 2015 19:08:09 +0100, Marcin Wojtas wrote:
> MVNETA_RXQ_HW_BUF_ALLOC bit which controls enabling hardware buffer
> allocation was mistakenly set as BIT(1). This commit fixes the assignment.
>
> Signed-off-by: Marcin Wojtas <[email protected]>
> Reviewed-by: Gregory CLEMENT <[email protected]>
> Cc: <[email protected]> # v3.8+
> Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
> unit")
> ---
> drivers/net/ethernet/marvell/mvneta.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
What is the user-visible impact of this bug? Indeed, the code is wrong,
but things were working fine until now.
So I am not questioning that this patch should be merged, but only
questioning whether the stable tag is appropriate.
Same questions for your patch 3/6.
Thomas
--
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
Hi Thomas,
2015-11-26 19:25 GMT+01:00 Thomas Petazzoni
<[email protected]>:
> Marcin,
>
> On Thu, 26 Nov 2015 19:08:08 +0100, Marcin Wojtas wrote:
>> This commit adds missing configuration of MBUS windows access protection
>> in mvneta_conf_mbus_windows function - a dedicated variable for that
>> purpose remained there unused since v3.8 initial mvneta support. Because
>> of that the register contents were inherited from the bootloader.
>>
>> Signed-off-by: Marcin Wojtas <[email protected]>
>> Reviewed-by: Gregory CLEMENT <[email protected]>
>> Cc: <[email protected]> # v3.8+
>> Fixes: c5aff18204da ("net: mvneta: driver for Marvell Armada 370/XP network
>> unit")
>
> Is this actually fixing a user visible problem? If not, then I am not
> sure it qualifies for stable.
Indeed, it's not fixing a visible thing, unless you resume from
suspend to ram. However, the original driver declares and updates
'win_protect ' variable, which have remained unused for past 3 years:)
If you think it's not worth cc'ing stable despite that, I will remove
the notification.
>
>> diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
>> index e84c7f2..c4327b9 100644
>> --- a/drivers/net/ethernet/marvell/mvneta.c
>> +++ b/drivers/net/ethernet/marvell/mvneta.c
>> @@ -62,6 +62,7 @@
>> #define MVNETA_WIN_SIZE(w) (0x2204 + ((w) << 3))
>> #define MVNETA_WIN_REMAP(w) (0x2280 + ((w) << 2))
>> #define MVNETA_BASE_ADDR_ENABLE 0x2290
>> +#define MVNETA_ACCESS_PROTECT_ENABLE 0x2294
>
> Spaces are used for indentation in the definitions of all other
> constants in this driver, so I think you should stick with this
> convention, at least to remain consistent.
>
Ok.
>> #define MVNETA_PORT_CONFIG 0x2400
>> #define MVNETA_UNI_PROMISC_MODE BIT(0)
>> #define MVNETA_DEF_RXQ(q) ((q) << 1)
>> @@ -3188,9 +3189,11 @@ static void mvneta_conf_mbus_windows(struct mvneta_port *pp,
>>
>> win_enable &= ~(1 << i);
>> win_protect |= 3 << (2 * i);
>> +
>
> Unneeded change.
>
Indeed, my mistake.
Thanks,
Marcin
Marcin,
On Thu, 26 Nov 2015 19:08:13 +0100, Marcin Wojtas wrote:
> The Ethernet controller found in the Armada 38x SoC's family support
> TCP/IP checksumming with frame sizes larger than 1600 bytes, however
> only on port 0.
>
> This commit enables it by setting 'tx-csum-limit' to 9800B in
> 'ethernet@70000' node.
>
> Signed-off-by: Marcin Wojtas <[email protected]>
> Cc: <[email protected]> # v3.18+
This and the previous patch is not a fix, and has no business going in
stable. It is enabling a better functionality, but it is clearly not a
fix.
Thomas
--
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
On Thu, Nov 26, 2015 at 07:08:11PM +0100, Marcin Wojtas wrote:
> In the actual RX processing, there is same error path for both descriptor
> ring refilling and building skb fails. This is not correct, because after
> successful refill, the ring is already updated with newly allocated
> buffer. Then, in case of build_skb() fail, hitherto code left the original
> buffer unmapped.
>
> This patch fixes above situation by swapping error check of skb build with
> DMA-unmap of original buffer.
>
> Signed-off-by: Marcin Wojtas <[email protected]>
> Cc: <[email protected]> # v4.2+
> Fixes a84e32894191 ("net: mvneta: fix refilling for Rx DMA buffers")
> ---
> drivers/net/ethernet/marvell/mvneta.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
Acked-by: Simon Guinot <[email protected]>
Thanks,
Simon
> diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
> index 0c3d923..62cf971 100644
> --- a/drivers/net/ethernet/marvell/mvneta.c
> +++ b/drivers/net/ethernet/marvell/mvneta.c
> @@ -1580,12 +1580,16 @@ static int mvneta_rx(struct mvneta_port *pp, int rx_todo,
> }
>
> skb = build_skb(data, pp->frag_size > PAGE_SIZE ? 0 : pp->frag_size);
> - if (!skb)
> - goto err_drop_frame;
>
> + /* After refill old buffer has to be unmapped regardless
> + * the skb is successfully built or not.
> + */
> dma_unmap_single(dev->dev.parent, phys_addr,
> MVNETA_RX_BUF_SIZE(pp->pkt_size), DMA_FROM_DEVICE);
>
> + if (!skb)
> + goto err_drop_frame;
> +
> rcvd_pkts++;
> rcvd_bytes += rx_bytes;
>
> --
> 1.8.3.1
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Thomas,
2015-11-26 19:33 GMT+01:00 Thomas Petazzoni
<[email protected]>:
> Marcin,
>
> On Thu, 26 Nov 2015 19:08:13 +0100, Marcin Wojtas wrote:
>> The Ethernet controller found in the Armada 38x SoC's family support
>> TCP/IP checksumming with frame sizes larger than 1600 bytes, however
>> only on port 0.
>>
>> This commit enables it by setting 'tx-csum-limit' to 9800B in
>> 'ethernet@70000' node.
>>
>> Signed-off-by: Marcin Wojtas <[email protected]>
>> Cc: <[email protected]> # v3.18+
>
> This and the previous patch is not a fix, and has no business going in
> stable. It is enabling a better functionality, but it is clearly not a
> fix.
>
Ok. I thought that patches enabling tx_csum_limit were adding a sort
of regression, but I can agree to your interpretation. In such case I
will leave stable notification only for patch #4.
Thanks,
Marcin