* Re: [PATCH] tcp: tcp_replace_ts_recent() should not be called from tcp_validate_incoming()
From: Eric Dumazet @ 2012-11-13 13:28 UTC (permalink / raw)
To: Romain Francoise
Cc: David Miller, netdev, Neal Cardwell, Yuchung Cheng,
Nandita Dukkipati, H.K. Jerry Chu
In-Reply-To: <87k3tp6783.fsf@silenus.orebokech.com>
On Tue, 2012-11-13 at 13:02 +0100, Romain Francoise wrote:
> Eric Dumazet <eric.dumazet@gmail.com> writes:
>
> > We added support for RFC 5962 in latest kernels but TCP fails
> > to perform exhaustive check of ACK sequence.
>
> You probably meant 5961...
Oops you're right ;)
^ permalink raw reply
* [PATCH v3 00/11] rename random32 to prandom and introduce prandom_bytes()
From: Akinobu Mita @ 2012-11-13 13:36 UTC (permalink / raw)
To: linux-kernel, akpm
Cc: Akinobu Mita, Theodore Ts'o, Artem Bityutskiy, Adrian Hunter,
David Woodhouse, linux-mtd, Eilon Greenstein, netdev, Robert Love,
devel, Michel Lespinasse
This patchset introduces new functions into random32 library for
getting the requested number of pseudo-random bytes.
Before introducing these new functions into random32 library,
rename all random32 functions to have 'prandom_' prefix. As a result
the function prototypes are as follows:
void prandom_seed(u32 seed); /* rename from srandom32() */
u32 prandom_u32(void); /* rename from random32() */
void prandom_bytes(void *buf, int nbytes);
void prandom_seed_state(struct rnd_state *state, u64 seed);
/* rename from prandom32_seed() */
u32 prandom_u32_state(struct rnd_state *state);
/* rename from prandom32() */
void prandom_bytes_state(struct rnd_state *state, void *buf, int nbytes);
The purpose of this renaming is to prevent some kernel developers
from assuming that prandom32() and random32() might imply that only
prandom32() was the one using a pseudo-random number generator by
prandom32's "p", and the result may be a very embarassing security
exposure. This concern was expressed by Theodore Ts'o.
Changelog
* v3
- change common prefix from 'random32_' to 'prandom_'
- ensure prandom_bytes_state() generates same bytes with same rnd_state
* v2
- rename prandom32 to random32_state
- dropped lib/uuid.c patch
- add bnx2 and mtd_stresstest patches
Akinobu Mita (11):
random32: rename random32 to prandom
prandom: introduce prandom_bytes() and prandom_bytes_state()
bnx2x: use prandom_bytes()
mtd: nandsim: use prandom_bytes
ubifs: use prandom_bytes
mtd: mtd_nandecctest: use prandom_bytes instead of get_random_bytes()
mtd: mtd_oobtest: convert to use prandom library
mtd: mtd_pagetest: convert to use prandom library
mtd: mtd_speedtest: use prandom_bytes
mtd: mtd_subpagetest: convert to use prandom library
mtd: mtd_stresstest: use prandom_bytes()
drivers/mtd/nand/nandsim.c | 5 +-
drivers/mtd/tests/mtd_nandecctest.c | 2 +-
drivers/mtd/tests/mtd_oobtest.c | 49 ++++---------
drivers/mtd/tests/mtd_pagetest.c | 43 ++++-------
drivers/mtd/tests/mtd_speedtest.c | 9 +--
drivers/mtd/tests/mtd_stresstest.c | 3 +-
drivers/mtd/tests/mtd_subpagetest.c | 42 +++--------
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 5 +-
drivers/scsi/fcoe/fcoe_ctlr.c | 4 +-
fs/ubifs/debug.c | 8 +-
include/linux/random.h | 19 +++--
lib/interval_tree_test_main.c | 7 +-
lib/random32.c | 97 +++++++++++++++++++------
lib/rbtree_test.c | 6 +-
14 files changed, 145 insertions(+), 154 deletions(-)
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Artem Bityutskiy <dedekind1@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: linux-mtd@lists.infradead.org
Cc: Eilon Greenstein <eilong@broadcom.com>
Cc: netdev@vger.kernel.org
Cc: Robert Love <robert.w.love@intel.com>
Cc: devel@open-fcoe.org
Cc: Michel Lespinasse <walken@google.com>
--
1.7.11.7
^ permalink raw reply
* [PATCH v3 02/11] prandom: introduce prandom_bytes() and prandom_bytes_state()
From: Akinobu Mita @ 2012-11-13 13:37 UTC (permalink / raw)
To: linux-kernel, akpm
Cc: Akinobu Mita, Theodore Ts'o, Artem Bityutskiy, Adrian Hunter,
David Woodhouse, linux-mtd, Eilon Greenstein, netdev
In-Reply-To: <1352813830-4624-1-git-send-email-akinobu.mita@gmail.com>
Add functions to get the requested number of pseudo-random bytes.
The difference from get_random_bytes() is that it generates pseudo-random
numbers by prandom_u32(). It doesn't consume the entropy pool, and the
sequence is reproducible if the same rnd_state is used. So it is suitable
for generating random bytes for testing.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Artem Bityutskiy <dedekind1@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: linux-mtd@lists.infradead.org
Cc: Eilon Greenstein <eilong@broadcom.com>
Cc: netdev@vger.kernel.org
---
* v3
- rename random32_get_bytes_state to prandom_bytes_state
- ensure prandom_bytes_state() generates same bytes with same rnd_state
* v2
- rename prandom32_get_bytes to random32_get_bytes_state
include/linux/random.h | 2 ++
lib/random32.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 51 insertions(+)
diff --git a/include/linux/random.h b/include/linux/random.h
index db6debc..d984608 100644
--- a/include/linux/random.h
+++ b/include/linux/random.h
@@ -26,6 +26,7 @@ unsigned int get_random_int(void);
unsigned long randomize_range(unsigned long start, unsigned long end, unsigned long len);
u32 prandom_u32(void);
+void prandom_bytes(void *buf, int nbytes);
void prandom_seed(u32 seed);
/*
@@ -36,6 +37,7 @@ void prandom_seed(u32 seed);
#define srandom32(seed) prandom_seed(seed)
u32 prandom_u32_state(struct rnd_state *);
+void prandom_bytes_state(struct rnd_state *state, void *buf, int nbytes);
/*
* Handle minimum values for seeds
diff --git a/lib/random32.c b/lib/random32.c
index d1830fa..52280d5 100644
--- a/lib/random32.c
+++ b/lib/random32.c
@@ -77,6 +77,55 @@ u32 prandom_u32(void)
}
EXPORT_SYMBOL(prandom_u32);
+/*
+ * prandom_bytes_state - get the requested number of pseudo-random bytes
+ *
+ * @state: pointer to state structure holding seeded state.
+ * @buf: where to copy the pseudo-random bytes to
+ * @bytes: the requested number of bytes
+ *
+ * This is used for pseudo-randomness with no outside seeding.
+ * For more random results, use prandom_bytes().
+ */
+void prandom_bytes_state(struct rnd_state *state, void *buf, int bytes)
+{
+ unsigned char *p = buf;
+ int i;
+
+ for (i = 0; i < round_down(bytes, sizeof(u32)); i += sizeof(u32)) {
+ u32 random = prandom_u32_state(state);
+ int j;
+
+ for (j = 0; j < sizeof(u32); j++) {
+ p[i + j] = random;
+ random >>= BITS_PER_BYTE;
+ }
+ }
+ if (i < bytes) {
+ u32 random = prandom_u32_state(state);
+
+ for (; i < bytes; i++) {
+ p[i] = random;
+ random >>= BITS_PER_BYTE;
+ }
+ }
+}
+EXPORT_SYMBOL(prandom_bytes_state);
+
+/**
+ * prandom_bytes - get the requested number of pseudo-random bytes
+ * @buf: where to copy the pseudo-random bytes to
+ * @bytes: the requested number of bytes
+ */
+void prandom_bytes(void *buf, int bytes)
+{
+ struct rnd_state *state = &get_cpu_var(net_rand_state);
+
+ prandom_bytes_state(state, buf, bytes);
+ put_cpu_var(state);
+}
+EXPORT_SYMBOL(prandom_bytes);
+
/**
* prandom_seed - add entropy to pseudo random number generator
* @seed: seed value
--
1.7.11.7
^ permalink raw reply related
* [PATCH v3 03/11] bnx2x: use prandom_bytes()
From: Akinobu Mita @ 2012-11-13 13:37 UTC (permalink / raw)
To: linux-kernel, akpm; +Cc: Akinobu Mita, Eilon Greenstein, netdev
In-Reply-To: <1352813830-4624-1-git-send-email-akinobu.mita@gmail.com>
Use prandom_bytes() to fill rss key with pseudo-random bytes.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Eilon Greenstein <eilong@broadcom.com>
Cc: netdev@vger.kernel.org
---
* v3
- rename random32_get_bytes to prandom_bytes
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 4833b6a..257d38b 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -1741,7 +1741,6 @@ int bnx2x_config_rss_pf(struct bnx2x *bp, struct bnx2x_rss_config_obj *rss_obj,
bool config_hash)
{
struct bnx2x_config_rss_params params = {NULL};
- int i;
/* Although RSS is meaningless when there is a single HW queue we
* still need it enabled in order to have HW Rx hash generated.
@@ -1773,9 +1772,7 @@ int bnx2x_config_rss_pf(struct bnx2x *bp, struct bnx2x_rss_config_obj *rss_obj,
if (config_hash) {
/* RSS keys */
- for (i = 0; i < sizeof(params.rss_key) / 4; i++)
- params.rss_key[i] = random32();
-
+ prandom_bytes(params.rss_key, sizeof(params.rss_key));
__set_bit(BNX2X_RSS_SET_SRCH, ¶ms.rss_flags);
}
--
1.7.11.7
^ permalink raw reply related
* Re: SR-IOV problem with Intel 82599EB (not enough MMIO resources for SR-IOV)
From: Jason Gao @ 2012-11-13 13:38 UTC (permalink / raw)
To: bhelgaas, Rose, Gregory V, sibai.li
Cc: ddutile, Kirsher, Jeffrey T, linux-kernel, netdev, kvm,
e1000-devel@lists.sourceforge.net, linux-pci, Yinghai Lu
In-Reply-To: <CAErSpo4xn9t=4yXGLdPQMtUJF=n38tYrMutAoj7NCsLUzQEbcg@mail.gmail.com>
I'm very sorry for delayed reply.now SR-IOV works for me in Centos
6.3,thank all of you.
On Fri, Nov 9, 2012 at 11:26 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> Linux normally uses the resource assignments done by the BIOS, but it
> is possible for the kernel to reassign those. We don't have good
> automatic support for that yet, but on a recent upstream kernel, you
> can try "pci=realloc". I doubt this option is in CentOS 6.3, though
Thank you very much,I try "pci=realloc" in Centos 6.3,and now it works for me.
On Sat, Nov 10, 2012 at 2:08 AM, Li, Sibai <sibai.li@intel.com> wrote:
> DellR710 with the latest BIOS should work fine for SR-IOV. My BIOS is v.6.3.0 and release date is 07/24/2012
> Please check if you configured intel_iommu=on in the grub.conf file.
> If you did, check your kernel .config file under Device Drivers-> IOMMU Hardware support->enable Support for Intel IOMMU using DMA remapping Devices, enable Intel DMA Remapping Devices by Default, enable Support for Interrupt Remapping.
thank you Sibai,Our server "Dell R710",its BIOS version is just
v.6.3.0 and release date is 07/24/2012,and I also configured
intel_iommu=on in the grub.conf file,but I can't find these IOMMU
options in "Device Drivers" in my kernel(2.6.32-279) .config file ,
btw my os is Centos 6.3(RHEL6.3),although the problem solved,I'd like
to know what's your os version ,kernel version?
^ permalink raw reply
* [RFC PATCH v1 3.7.0-rc4 0/2] dt:net/stmmac add full device tree support
From: Srinivas KANDAGATLA @ 2012-11-13 13:50 UTC (permalink / raw)
To: peppe.cavallaro-qxv4g6HH51o
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ,
davem-fT/PcQaiUtIeIZ0/mPfg9Q
From: Srinivas Kandagatla <srinivas.kandagatla-qxv4g6HH51o@public.gmane.org>
This patch series attempts to add full device tree support to stmmac driver.
Existing device bindings do not cover all the possible use cases of stmmac driver,
so new properties are added.
And also stmmac phy reset callback support via devicetrees is added.
Srinivas Kandagatla (2):
dt:net/stmmac: Add complete device tree support
dt:net/stmmac: Add dt specific phy reset callback support.
Documentation/devicetree/bindings/net/stmmac.txt | 51 +++++++++++++++++-
drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c | 50 ++++++++++++++---
.../net/ethernet/stmicro/stmmac/stmmac_platform.c | 58 ++++++++++++++------
3 files changed, 131 insertions(+), 28 deletions(-)
^ permalink raw reply
* [RFC PATCH v1 3.7.0-rc4 1/2] dt:net/stmmac: Add complete device tree support
From: Srinivas KANDAGATLA @ 2012-11-13 13:50 UTC (permalink / raw)
To: peppe.cavallaro; +Cc: netdev, devicetree-discuss, srinivas.kandagatla, davem
From: Srinivas Kandagatla <srinivas.kandagatla@st.com>
This patch attempts to add full device tree support to stmmac driver,
previously support to few optional properties were missed in both
bindings and driver.
With this patch DT support should be complete for stmmac driver.
Also all the vendor specific properties are prefixed with snps.
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@st.com>
---
Documentation/devicetree/bindings/net/stmmac.txt | 41 +++++++++++++-
.../net/ethernet/stmicro/stmmac/stmmac_platform.c | 58 ++++++++++++++------
2 files changed, 80 insertions(+), 19 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/stmmac.txt b/Documentation/devicetree/bindings/net/stmmac.txt
index 060bbf0..b55d369 100644
--- a/Documentation/devicetree/bindings/net/stmmac.txt
+++ b/Documentation/devicetree/bindings/net/stmmac.txt
@@ -14,16 +14,53 @@ Required properties:
Supported values are: "mii", "rmii", "gmii", "rgmii".
Optional properties:
-- mac-address: 6 bytes, mac address
+- mac-address 6 bytes, mac address
+- phy-handle Should be a phandle to the PHY.
+- snps,clk-csr Fixed CSR Clock range selection.
+- snps,enh-desc Boolean flag to indicate if mac support
+ enhanced descriptors.
+- snps,tx-coe flag for Tx Checksum Offload engine presence
+- snps,rx-coe flag fo Rx Checksum Offload engine presence.
+- snps,bugged-jumbo some HWs are not able to perform the csum in HW for
+ over-sized frames due to limited buffer sizes.
+ Setting this flag the csum will be done in SW on
+ JUMBO frames.
+- snps,force-sf-dma-mode force DMA to use the Store and Forward mode
+ instead of the Threshold.
+- snps,pbl Programmable Burst Length
+- snps,fixed-burst Program the DMA to use the fixed burst mode
+- snps,burst-len This is the value we put in the register
+ supported values are provided as macros in
+ linux/stmmac.h header file.
+- snps,phy-bus-name Name of the mdio bus to connect. if not specified
+ mac attempts to connect to stmmac mdio bus.
+- snps,phy-bus-id Mdio bus number to connect. if not specified 0 is used.
+- snps,phy-addr phy address to connect to.
+
Examples:
gmac0: ethernet@e0800000 {
- compatible = "st,spear600-gmac";
+ compatible = "snps,dwmac";
reg = <0xe0800000 0x8000>;
interrupt-parent = <&vic1>;
interrupts = <24 23>;
interrupt-names = "macirq", "eth_wake_irq";
mac-address = [000000000000]; /* Filled in by U-Boot */
phy-mode = "gmii";
+
+ snps,clk-csr = <0x0>;
+ snps,enh-desc;
+ snps,tx-coe;
+ snps,rx-coe;
+ snps,bugged-jumbo;
+ snps,force-sf-dma-mode;
+ snps,pbl = <32>;
+ snps,burst-len = <32>;
+ snps,fixed-burst;
+
+ /* PHY specific */
+ snps,phy-bus-name = "stmmac";
+ snps,phy-bus-id = <0>;
+ snps,phy-addr = <0x9>;
};
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
index ed112b5..6ee7548 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
@@ -29,26 +29,31 @@
#include "stmmac.h"
#ifdef CONFIG_OF
+
+static u64 stmmac_dma_mask = DMA_BIT_MASK(32);
static int __devinit stmmac_probe_config_dt(struct platform_device *pdev,
struct plat_stmmacenet_data *plat,
const char **mac)
{
struct device_node *np = pdev->dev.of_node;
-
+ struct stmmac_dma_cfg *dma_cfg;
+ const char **phy_bus_name = (const char **)&plat->phy_bus_name;
if (!np)
return -ENODEV;
*mac = of_get_mac_address(np);
plat->interface = of_get_phy_mode(np);
- plat->mdio_bus_data = devm_kzalloc(&pdev->dev,
- sizeof(struct stmmac_mdio_bus_data),
- GFP_KERNEL);
+ of_property_read_string(np, "snps,phy-bus-name", phy_bus_name);
+ of_property_read_u32(np, "snps,phy-bus-id", &plat->bus_id);
+ of_property_read_u32(np, "snps,phy-addr", &plat->phy_addr);
+
+ if (*phy_bus_name && strcmp(*phy_bus_name, "stmmac"))
+ plat->mdio_bus_data = NULL;
+ else
+ plat->mdio_bus_data = devm_kzalloc(&pdev->dev,
+ sizeof(struct stmmac_mdio_bus_data),
+ GFP_KERNEL);
- /*
- * Currently only the properties needed on SPEAr600
- * are provided. All other properties should be added
- * once needed on other platforms.
- */
if (of_device_is_compatible(np, "st,spear600-gmac") ||
of_device_is_compatible(np, "snps,dwmac-3.70a") ||
of_device_is_compatible(np, "snps,dwmac")) {
@@ -56,6 +61,24 @@ static int __devinit stmmac_probe_config_dt(struct platform_device *pdev,
plat->pmt = 1;
}
+ of_property_read_u32(np, "snps,clk-csr", &plat->clk_csr);
+
+ plat->enh_desc = of_property_read_bool(np, "snps,enh-desc");
+ plat->tx_coe = of_property_read_bool(np, "snps,tx-coe");
+ plat->rx_coe = of_property_read_bool(np, "snps,rx-coe");
+ plat->bugged_jumbo = of_property_read_bool(np, "snps,bugged-jumbo");
+ plat->force_sf_dma_mode = of_property_read_bool(np,
+ "snps,force-sf-dma-mode");
+
+ dma_cfg = devm_kzalloc(&pdev->dev, sizeof(*dma_cfg), GFP_KERNEL);
+ plat->dma_cfg = dma_cfg;
+ of_property_read_u32(np, "snps,burst-len", &dma_cfg->burst_len);
+ of_property_read_u32(np, "snps,pbl", &dma_cfg->pbl);
+ dma_cfg->fixed_burst = of_property_read_bool(np, "snps,fixed-burst");
+ dma_cfg->mixed_burst = of_property_read_bool(np, "snps,mixed-burst");
+
+ if (!pdev->dev.dma_mask)
+ pdev->dev.dma_mask = &stmmac_dma_mask;
return 0;
}
#else
@@ -93,14 +116,17 @@ static int __devinit stmmac_pltfr_probe(struct platform_device *pdev)
pr_err("%s: ERROR: memory mapping failed", __func__);
return -ENOMEM;
}
-
+ plat_dat = pdev->dev.platform_data;
if (pdev->dev.of_node) {
- plat_dat = devm_kzalloc(&pdev->dev,
+ if (!plat_dat) {
+ /* no data from OF_DEV_AUXDATA */
+ plat_dat = devm_kzalloc(&pdev->dev,
sizeof(struct plat_stmmacenet_data),
GFP_KERNEL);
- if (!plat_dat) {
- pr_err("%s: ERROR: no memory", __func__);
- return -ENOMEM;
+ if (!plat_dat) {
+ pr_err("%s: ERROR: no memory", __func__);
+ return -ENOMEM;
+ }
}
ret = stmmac_probe_config_dt(pdev, plat_dat, &mac);
@@ -108,8 +134,6 @@ static int __devinit stmmac_pltfr_probe(struct platform_device *pdev)
pr_err("%s: main dt probe failed", __func__);
return ret;
}
- } else {
- plat_dat = pdev->dev.platform_data;
}
/* Custom initialisation (if needed)*/
--
1.7.0.4
^ permalink raw reply related
* [RFC PATCH v1 3.7.0-rc4 2/2] dt:net/stmmac: Add dt specific phy reset callback support.
From: Srinivas KANDAGATLA @ 2012-11-13 13:50 UTC (permalink / raw)
To: peppe.cavallaro; +Cc: netdev, devicetree-discuss, srinivas.kandagatla, davem
From: Srinivas Kandagatla <srinivas.kandagatla@st.com>
This patch adds phy reset callback support for stmmac driver via device
trees. It adds three new properties to gmac device tree bindings to
define the reset signal via gpio.
With this patch users can conveniently pass reset gpio number with pre,
pulse and post delay in micro secs via DTs.
active low:
_________ ____________
<pre-delay> | <pulse-delay>|<post-delay>
| |
|______________|
active high:
________________
<pre-delay> |<pulse-delay> |<post-delay>
| |
_________| |___________
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@st.com>
---
Documentation/devicetree/bindings/net/stmmac.txt | 10 ++++
drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c | 50 +++++++++++++++++----
2 files changed, 51 insertions(+), 9 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/stmmac.txt b/Documentation/devicetree/bindings/net/stmmac.txt
index b55d369..cdc536c 100644
--- a/Documentation/devicetree/bindings/net/stmmac.txt
+++ b/Documentation/devicetree/bindings/net/stmmac.txt
@@ -37,6 +37,12 @@ Optional properties:
- snps,phy-bus-id Mdio bus number to connect. if not specified 0 is used.
- snps,phy-addr phy address to connect to.
+- snps,reset-gpio gpio number for phy reset.
+- snps,reset-active-low boolean flag to indicate if phy reset is active low.
+- snps,reset-delays-us is triplet of delays
+ The 1st cell is reset pre-delay in micro seconds.
+ The 2nd cell is reset pulse in micro seconds.
+ The 3rd cell is reset post-delay in micro seconds.
Examples:
@@ -63,4 +69,8 @@ Examples:
snps,phy-bus-name = "stmmac";
snps,phy-bus-id = <0>;
snps,phy-addr = <0x9>;
+
+ snps,reset-gpio = <&gpio_bank4 7>;
+ snps,reset-active-low;
+ snps,reset-delays-us = <0 10000 10000>;
};
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
index 0376a5e..f965e83 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
@@ -27,6 +27,9 @@
#include <linux/mii.h>
#include <linux/phy.h>
#include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+
#include <asm/io.h>
#include "stmmac.h"
@@ -131,17 +134,46 @@ static int stmmac_mdio_reset(struct mii_bus *bus)
struct net_device *ndev = bus->priv;
struct stmmac_priv *priv = netdev_priv(ndev);
unsigned int mii_address = priv->hw->mii.addr;
+ struct device *dev = priv->device;
- if (priv->plat->mdio_bus_data->phy_reset) {
- pr_debug("stmmac_mdio_reset: calling phy_reset\n");
- priv->plat->mdio_bus_data->phy_reset(priv->plat->bsp_priv);
- }
+ if (!dev->of_node) {
+ if (priv->plat->mdio_bus_data->phy_reset) {
+ pr_debug("stmmac_mdio_reset: calling phy_reset\n");
+ priv->plat->mdio_bus_data->phy_reset(
+ priv->plat->bsp_priv);
+ }
+
+ /* This is a workaround for problems with the STE101P PHY.
+ * It doesn't complete its reset until at least one clock cycle
+ * on MDC, so perform a dummy mdio read.
+ */
+ writel(0, priv->ioaddr + mii_address);
+ } else {
+ int reset_gpio, active_low;
+ struct device_node *np = dev->of_node;
+ u32 delays[3] = {0,};
+
+ if (!np)
+ return 0;
+
+ reset_gpio = of_get_named_gpio(np, "snps,reset-gpio", 0);
+ if (reset_gpio < 0)
+ return 0;
- /* This is a workaround for problems with the STE101P PHY.
- * It doesn't complete its reset until at least one clock cycle
- * on MDC, so perform a dummy mdio read.
- */
- writel(0, priv->ioaddr + mii_address);
+ active_low = of_get_property(np,
+ "snps,reset-active-low", NULL) ? 1 : 0;
+ of_property_read_u32_array(np,
+ "snps,reset-delays-us", delays, 3);
+
+ gpio_request(reset_gpio, "mdio-reset");
+ gpio_direction_output(reset_gpio, active_low ? 1 : 0);
+ udelay(delays[0]);
+ gpio_set_value(reset_gpio, active_low ? 0 : 1);
+ udelay(delays[1]);
+ gpio_set_value(reset_gpio, active_low ? 1 : 0);
+ udelay(delays[2]);
+ gpio_free(reset_gpio);
+ }
#endif
return 0;
}
--
1.7.0.4
^ permalink raw reply related
* e1000e on DH55HC stalling and kernel panic in 3.6.6
From: Denys Fedoryshchenko @ 2012-11-13 13:58 UTC (permalink / raw)
To: jeffrey.t.kirsher, jesse.brandeburg, bruce.w.allan,
carolyn.wyborny, donald.c.skidmore, gregory.v.rose,
peter.p.waskiewicz.jr, alexander.h.duyck, john.ronciak,
e1000-devel, netdev, linux-kernel
Hi
I just tried to run latest kernel on my DH55HC motherboard latest
kernel 3.6.6 and got various network problems, such as network traffic
are stopping, and sometimes i am getting kernel panic.
When traffic are stopping, ethtool -r eth0 sometimes helps.
When i do ethtool -G eth0 rx NNN , sometimes it will give kernel panic,
but it is hard to reproduce.
I tried to capture panic on pictures, so will try to decode on what i
got photo, it is a nightmare, but sadly i dont have serial in hands to
get data over it.
skbuff: skb_over_panic: text:f86fc769 len:25807 put:25807 head:c1da1800
data:c1da1840 tail:0xc1da7d0f end:c1da1f40 dev:eth0
kernel BUG at net/core/skbuff.c:127
opcode: 0000 [#1] SMP
Pid: 0 comm: swapper/6 Not tained 3.6.6-build-0063 #23
EIP: 0060:[<c02e7980>] EFLAGS: 00010296 CPU:6
EIP is at skb_put+0x83/0x8e
There is registers and stack, let me know if you need specific fields
Call trace:
f86fc769 ? e1000_clean_rx_irq+0x1e1/0x2af [e1000e]
f86fc769 e1000_clean_rx_irq+0x1e1/0x2af [e1000e]
f86fcc73 e1000e_poll+0x6a/0x209 [e1000e]
c02f1630 net_rx_action+0x90/0x15d
c01302d5 __do_softirq+0x8a/-x13b
c013024b ? local_bh_enable+0xd/0xd
<IRQ>
c0130504 irq_exit+0x41/0x91
c0102c37 do_IRQ+0x79/0x8d
There is also more data, let me know if you need it.
---
Denys Fedoryshchenko, Network Engineer, Virtual ISP S.A.L.
^ permalink raw reply
* [net-next 01/11] ixgbe: Do not use DCA to prefetch the entire packet into the cache
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
The way the code was previously written it was causing DCA to prefetch the
entire packet into the cache when it was enabled. That is excessive as we
only really need the headers.
We are now prefetching the headers via software so doing this from DCA would
be redundant anyway. So clear the bit that was causing us to prefetch the
packet data and instead only use DCA for the descriptor rings.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 690535a..38fc186 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1015,7 +1015,6 @@ static void ixgbe_update_rx_dca(struct ixgbe_adapter *adapter,
* which will cause the DCA tag to be cleared.
*/
rxctrl |= IXGBE_DCA_RXCTRL_DESC_RRO_EN |
- IXGBE_DCA_RXCTRL_DATA_DCA_EN |
IXGBE_DCA_RXCTRL_DESC_DCA_EN;
IXGBE_WRITE_REG(hw, IXGBE_DCA_RXCTRL(reg_idx), rxctrl);
--
1.7.11.7
^ permalink raw reply related
* [net-next 00/11][pull request] Intel Wired LAN Driver Updates
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, gospo, sassmann
This series contains updates to ixgbe, ixgbevf and igb.
The following are changes since commit 9fafd65ad407d4e0c96919a325f568dd95d032af:
ipv6 ndisc: Use pre-defined in6addr_linklocal_allnodes.
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Akeem G. Abodunrin (2):
igb: Support for modifying UDP RSS flow hashing
igb: Ethtool support to enable and disable EEE
Alexander Duyck (4):
ixgbe: Do not use DCA to prefetch the entire packet into the cache
igb: Make TSO check for CHECKSUM_PARTIAL to avoid skb_is_gso check
igb: Update igb Tx flags to improve code efficiency
igb: Improve performance and reduce size of igb_tx_map
Carolyn Wyborny (1):
igb: Clear Go Link Disconnect for 82580 and later devices
Emil Tantilov (1):
ixgbevf: fix possible use of uninitialized variable
Greg Rose (2):
ixgbevf: Add flag to indicate when rx is in net poll
ixgbevf: Reduce size of maximum rx buffer
Jakub Kicinski (1):
ixgbevf: make sure probe fails on MSI-X enable error
drivers/net/ethernet/intel/igb/e1000_82575.c | 8 +
drivers/net/ethernet/intel/igb/e1000_defines.h | 1 +
drivers/net/ethernet/intel/igb/e1000_phy.h | 1 +
drivers/net/ethernet/intel/igb/igb.h | 31 ++-
drivers/net/ethernet/intel/igb/igb_ethtool.c | 281 ++++++++++++++++++++++
drivers/net/ethernet/intel/igb/igb_main.c | 124 +++++-----
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 1 -
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 9 +-
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 54 +++--
9 files changed, 412 insertions(+), 98 deletions(-)
--
1.7.11.7
^ permalink raw reply
* [net-next 02/11] ixgbevf: make sure probe fails on MSI-X enable error
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Jakub Kicinski, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jakub Kicinski <jakub.kicinski@intel.com>
This driver cannot work without MSI-X interrupts
so there is no mechanism to fall back to.
Signed-off-by: Jakub Kicinski <jakub.kicinski@intel.com>
Acked-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 26 +++++++++++------------
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index f3d3947..a8fbe6c 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1721,8 +1721,8 @@ void ixgbevf_reset(struct ixgbevf_adapter *adapter)
}
}
-static void ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
- int vectors)
+static int ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
+ int vectors)
{
int err, vector_threshold;
@@ -1740,21 +1740,18 @@ static void ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
while (vectors >= vector_threshold) {
err = pci_enable_msix(adapter->pdev, adapter->msix_entries,
vectors);
- if (!err) /* Success in acquiring all requested vectors. */
+ if (!err || err < 0) /* Success or a nasty failure. */
break;
- else if (err < 0)
- vectors = 0; /* Nasty failure, quit now */
else /* err == number of vectors we should try again with */
vectors = err;
}
- if (vectors < vector_threshold) {
- /* Can't allocate enough MSI-X interrupts? Oh well.
- * This just means we'll go with either a single MSI
- * vector or fall back to legacy interrupts.
- */
- hw_dbg(&adapter->hw,
- "Unable to allocate MSI-X interrupts\n");
+ if (vectors < vector_threshold)
+ err = -ENOMEM;
+
+ if (err) {
+ dev_err(&adapter->pdev->dev,
+ "Unable to allocate MSI-X interrupts\n");
kfree(adapter->msix_entries);
adapter->msix_entries = NULL;
} else {
@@ -1765,6 +1762,7 @@ static void ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
*/
adapter->num_msix_vectors = vectors;
}
+ return err;
}
/**
@@ -1868,7 +1866,9 @@ static int ixgbevf_set_interrupt_capability(struct ixgbevf_adapter *adapter)
for (vector = 0; vector < v_budget; vector++)
adapter->msix_entries[vector].entry = vector;
- ixgbevf_acquire_msix_vectors(adapter, v_budget);
+ err = ixgbevf_acquire_msix_vectors(adapter, v_budget);
+ if (err)
+ goto out;
err = netif_set_real_num_tx_queues(netdev, adapter->num_tx_queues);
if (err)
--
1.7.11.7
^ permalink raw reply related
* [net-next 03/11] ixgbevf: fix possible use of uninitialized variable
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
This patch resolves the following warning:
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c: In function ‘ixgbevf_probe’:
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c:1742:290: warning: ‘err’ may be used uninitialized in this function [-Wuninitialized]
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c:1717:6: note: ‘err’ was declared here
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index a8fbe6c..ee5ff0e 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1724,7 +1724,8 @@ void ixgbevf_reset(struct ixgbevf_adapter *adapter)
static int ixgbevf_acquire_msix_vectors(struct ixgbevf_adapter *adapter,
int vectors)
{
- int err, vector_threshold;
+ int err = 0;
+ int vector_threshold;
/* We'll want at least 2 (vector_threshold):
* 1) TxQ[0] + RxQ[0] handler
--
1.7.11.7
^ permalink raw reply related
* [net-next 04/11] ixgbevf: Add flag to indicate when rx is in net poll
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
napi_gro_receive shouldn't be called from netpoll context. Doing
so was causing kernel panics when jumbo frames larger than 2K were set.
Add a flag to check if the Rx ring processing is occurring from interrupt
context or from netpoll context and call netif_rx() if in the polling
context.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 1 +
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 7 ++++++-
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index 2323ccd..9faaf54 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -229,6 +229,7 @@ struct ixgbevf_adapter {
*/
u32 flags;
#define IXGBE_FLAG_IN_WATCHDOG_TASK (u32)(1)
+#define IXGBE_FLAG_IN_NETPOLL (u32)(1 << 1)
/* OS defined structs */
struct net_device *netdev;
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index ee5ff0e..00f9698 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -288,7 +288,10 @@ static void ixgbevf_receive_skb(struct ixgbevf_q_vector *q_vector,
if (is_vlan && test_bit(tag & VLAN_VID_MASK, adapter->active_vlans))
__vlan_hwaccel_put_tag(skb, tag);
- napi_gro_receive(&q_vector->napi, skb);
+ if (!(adapter->flags & IXGBE_FLAG_IN_NETPOLL))
+ napi_gro_receive(&q_vector->napi, skb);
+ else
+ netif_rx(skb);
}
/**
@@ -550,9 +553,11 @@ static int ixgbevf_poll(struct napi_struct *napi, int budget)
else
per_ring_budget = budget;
+ adapter->flags |= IXGBE_FLAG_IN_NETPOLL;
ixgbevf_for_each_ring(ring, q_vector->rx)
clean_complete &= ixgbevf_clean_rx_irq(q_vector, ring,
per_ring_budget);
+ adapter->flags &= ~IXGBE_FLAG_IN_NETPOLL;
/* If all work not completed, return budget and keep polling */
if (!clean_complete)
--
1.7.11.7
^ permalink raw reply related
* [net-next 06/11] igb: Clear Go Link Disconnect for 82580 and later devices
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Carolyn Wyborny, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
Customers are requesting that the hw prevents PHY from establishing link
until the driver loads. This patch clears the Go Link Disconnect bit which
provides the requested behavior on parts 82580 and later.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/e1000_82575.c | 8 ++++++++
drivers/net/ethernet/intel/igb/e1000_phy.h | 1 +
2 files changed, 9 insertions(+)
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index deb0597..8c12dbd 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -1278,12 +1278,20 @@ static s32 igb_setup_copper_link_82575(struct e1000_hw *hw)
{
u32 ctrl;
s32 ret_val;
+ u32 phpm_reg;
ctrl = rd32(E1000_CTRL);
ctrl |= E1000_CTRL_SLU;
ctrl &= ~(E1000_CTRL_FRCSPD | E1000_CTRL_FRCDPX);
wr32(E1000_CTRL, ctrl);
+ /* Clear Go Link Disconnect bit */
+ if (hw->mac.type >= e1000_82580) {
+ phpm_reg = rd32(E1000_82580_PHY_POWER_MGMT);
+ phpm_reg &= ~E1000_82580_PM_GO_LINKD;
+ wr32(E1000_82580_PHY_POWER_MGMT, phpm_reg);
+ }
+
ret_val = igb_setup_serdes_link_82575(hw);
if (ret_val)
goto out;
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.h b/drivers/net/ethernet/intel/igb/e1000_phy.h
index 6ac3299..ed282f8 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.h
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.h
@@ -124,6 +124,7 @@ s32 igb_check_polarity_m88(struct e1000_hw *hw);
#define E1000_82580_PM_SPD 0x0001 /* Smart Power Down */
#define E1000_82580_PM_D0_LPLU 0x0002 /* For D0a states */
#define E1000_82580_PM_D3_LPLU 0x0004 /* For all other states */
+#define E1000_82580_PM_GO_LINKD 0x0020 /* Go Link Disconnect */
/* Enable flexible speed on link-up */
#define IGP02E1000_PM_D0_LPLU 0x0002 /* For D0a states */
--
1.7.11.7
^ permalink raw reply related
* [net-next 05/11] ixgbevf: Reduce size of maximum rx buffer
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Greg Rose, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Greg Rose <gregory.v.rose@intel.com>
There's no need to support up to 15k buffers since the HW is limited to
9.5k in SR-IOV mode. Instead, allocate buffers that fit and align inside
of a 32K memory buffer.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/ixgbevf/ixgbevf.h | 8 ++++----
drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 18 +++++++++---------
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index 9faaf54..1211fa0 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -101,10 +101,10 @@ struct ixgbevf_ring {
/* Supported Rx Buffer Sizes */
#define IXGBEVF_RXBUFFER_256 256 /* Used for packet split */
-#define IXGBEVF_RXBUFFER_3K 3072
-#define IXGBEVF_RXBUFFER_7K 7168
-#define IXGBEVF_RXBUFFER_15K 15360
-#define IXGBEVF_MAX_RXBUFFER 16384 /* largest size for single descriptor */
+#define IXGBEVF_RXBUFFER_2K 2048
+#define IXGBEVF_RXBUFFER_4K 4096
+#define IXGBEVF_RXBUFFER_8K 8192
+#define IXGBEVF_RXBUFFER_10K 10240
#define IXGBEVF_RX_HDR_SIZE IXGBEVF_RXBUFFER_256
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 00f9698..9d88153 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -1087,20 +1087,20 @@ static void ixgbevf_set_rx_buffer_len(struct ixgbevf_adapter *adapter)
max_frame += VLAN_HLEN;
/*
- * Make best use of allocation by using all but 1K of a
- * power of 2 allocation that will be used for skb->head.
+ * Allocate buffer sizes that fit well into 32K and
+ * take into account max frame size of 9.5K
*/
if ((hw->mac.type == ixgbe_mac_X540_vf) &&
(max_frame <= MAXIMUM_ETHERNET_VLAN_SIZE))
rx_buf_len = MAXIMUM_ETHERNET_VLAN_SIZE;
- else if (max_frame <= IXGBEVF_RXBUFFER_3K)
- rx_buf_len = IXGBEVF_RXBUFFER_3K;
- else if (max_frame <= IXGBEVF_RXBUFFER_7K)
- rx_buf_len = IXGBEVF_RXBUFFER_7K;
- else if (max_frame <= IXGBEVF_RXBUFFER_15K)
- rx_buf_len = IXGBEVF_RXBUFFER_15K;
+ else if (max_frame <= IXGBEVF_RXBUFFER_2K)
+ rx_buf_len = IXGBEVF_RXBUFFER_2K;
+ else if (max_frame <= IXGBEVF_RXBUFFER_4K)
+ rx_buf_len = IXGBEVF_RXBUFFER_4K;
+ else if (max_frame <= IXGBEVF_RXBUFFER_8K)
+ rx_buf_len = IXGBEVF_RXBUFFER_8K;
else
- rx_buf_len = IXGBEVF_MAX_RXBUFFER;
+ rx_buf_len = IXGBEVF_RXBUFFER_10K;
for (i = 0; i < adapter->num_rx_queues; i++)
adapter->rx_ring[i].rx_buf_len = rx_buf_len;
--
1.7.11.7
^ permalink raw reply related
* [net-next 07/11] igb: Support for modifying UDP RSS flow hashing
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Akeem G. Abodunrin, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: "Akeem G. Abodunrin" <akeem.g.abodunrin@intel.com>
This patch provides ability to enable or disable UDP RSS hashing. It gives
users option of generating RSS hash based on the UDP source and destination
ports numbers. Currently, UDP flow hash is always disabled in igb-driver.
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/igb.h | 14 ++-
drivers/net/ethernet/intel/igb/igb_ethtool.c | 181 +++++++++++++++++++++++++++
drivers/net/ethernet/intel/igb/igb_main.c | 13 +-
3 files changed, 197 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index 796db53..b23a667 100644
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -386,12 +386,14 @@ struct igb_adapter {
char fw_version[32];
};
-#define IGB_FLAG_HAS_MSI (1 << 0)
-#define IGB_FLAG_DCA_ENABLED (1 << 1)
-#define IGB_FLAG_QUAD_PORT_A (1 << 2)
-#define IGB_FLAG_QUEUE_PAIRS (1 << 3)
-#define IGB_FLAG_DMAC (1 << 4)
-#define IGB_FLAG_PTP (1 << 5)
+#define IGB_FLAG_HAS_MSI (1 << 0)
+#define IGB_FLAG_DCA_ENABLED (1 << 1)
+#define IGB_FLAG_QUAD_PORT_A (1 << 2)
+#define IGB_FLAG_QUEUE_PAIRS (1 << 3)
+#define IGB_FLAG_DMAC (1 << 4)
+#define IGB_FLAG_PTP (1 << 5)
+#define IGB_FLAG_RSS_FIELD_IPV4_UDP (1 << 6)
+#define IGB_FLAG_RSS_FIELD_IPV6_UDP (1 << 7)
/* DMA Coalescing defines */
#define IGB_MIN_TXPBSIZE 20408
diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index d8b1bee..2b82a53 100644
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -2350,6 +2350,185 @@ static int igb_get_ts_info(struct net_device *dev,
}
}
+static int igb_get_rss_hash_opts(struct igb_adapter *adapter,
+ struct ethtool_rxnfc *cmd)
+{
+ cmd->data = 0;
+
+ /* Report default options for RSS on igb */
+ switch (cmd->flow_type) {
+ case TCP_V4_FLOW:
+ cmd->data |= RXH_L4_B_0_1 | RXH_L4_B_2_3;
+ case UDP_V4_FLOW:
+ if (adapter->flags & IGB_FLAG_RSS_FIELD_IPV4_UDP)
+ cmd->data |= RXH_L4_B_0_1 | RXH_L4_B_2_3;
+ case SCTP_V4_FLOW:
+ case AH_ESP_V4_FLOW:
+ case AH_V4_FLOW:
+ case ESP_V4_FLOW:
+ case IPV4_FLOW:
+ cmd->data |= RXH_IP_SRC | RXH_IP_DST;
+ break;
+ case TCP_V6_FLOW:
+ cmd->data |= RXH_L4_B_0_1 | RXH_L4_B_2_3;
+ case UDP_V6_FLOW:
+ if (adapter->flags & IGB_FLAG_RSS_FIELD_IPV6_UDP)
+ cmd->data |= RXH_L4_B_0_1 | RXH_L4_B_2_3;
+ case SCTP_V6_FLOW:
+ case AH_ESP_V6_FLOW:
+ case AH_V6_FLOW:
+ case ESP_V6_FLOW:
+ case IPV6_FLOW:
+ cmd->data |= RXH_IP_SRC | RXH_IP_DST;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int igb_get_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd,
+ u32 *rule_locs)
+{
+ struct igb_adapter *adapter = netdev_priv(dev);
+ int ret = -EOPNOTSUPP;
+
+ switch (cmd->cmd) {
+ case ETHTOOL_GRXRINGS:
+ cmd->data = adapter->num_rx_queues;
+ ret = 0;
+ break;
+ case ETHTOOL_GRXFH:
+ ret = igb_get_rss_hash_opts(adapter, cmd);
+ break;
+ default:
+ break;
+ }
+
+ return ret;
+}
+
+#define UDP_RSS_FLAGS (IGB_FLAG_RSS_FIELD_IPV4_UDP | \
+ IGB_FLAG_RSS_FIELD_IPV6_UDP)
+static int igb_set_rss_hash_opt(struct igb_adapter *adapter,
+ struct ethtool_rxnfc *nfc)
+{
+ u32 flags = adapter->flags;
+
+ /* RSS does not support anything other than hashing
+ * to queues on src and dst IPs and ports
+ */
+ if (nfc->data & ~(RXH_IP_SRC | RXH_IP_DST |
+ RXH_L4_B_0_1 | RXH_L4_B_2_3))
+ return -EINVAL;
+
+ switch (nfc->flow_type) {
+ case TCP_V4_FLOW:
+ case TCP_V6_FLOW:
+ if (!(nfc->data & RXH_IP_SRC) ||
+ !(nfc->data & RXH_IP_DST) ||
+ !(nfc->data & RXH_L4_B_0_1) ||
+ !(nfc->data & RXH_L4_B_2_3))
+ return -EINVAL;
+ break;
+ case UDP_V4_FLOW:
+ if (!(nfc->data & RXH_IP_SRC) ||
+ !(nfc->data & RXH_IP_DST))
+ return -EINVAL;
+ switch (nfc->data & (RXH_L4_B_0_1 | RXH_L4_B_2_3)) {
+ case 0:
+ flags &= ~IGB_FLAG_RSS_FIELD_IPV4_UDP;
+ break;
+ case (RXH_L4_B_0_1 | RXH_L4_B_2_3):
+ flags |= IGB_FLAG_RSS_FIELD_IPV4_UDP;
+ break;
+ default:
+ return -EINVAL;
+ }
+ break;
+ case UDP_V6_FLOW:
+ if (!(nfc->data & RXH_IP_SRC) ||
+ !(nfc->data & RXH_IP_DST))
+ return -EINVAL;
+ switch (nfc->data & (RXH_L4_B_0_1 | RXH_L4_B_2_3)) {
+ case 0:
+ flags &= ~IGB_FLAG_RSS_FIELD_IPV6_UDP;
+ break;
+ case (RXH_L4_B_0_1 | RXH_L4_B_2_3):
+ flags |= IGB_FLAG_RSS_FIELD_IPV6_UDP;
+ break;
+ default:
+ return -EINVAL;
+ }
+ break;
+ case AH_ESP_V4_FLOW:
+ case AH_V4_FLOW:
+ case ESP_V4_FLOW:
+ case SCTP_V4_FLOW:
+ case AH_ESP_V6_FLOW:
+ case AH_V6_FLOW:
+ case ESP_V6_FLOW:
+ case SCTP_V6_FLOW:
+ if (!(nfc->data & RXH_IP_SRC) ||
+ !(nfc->data & RXH_IP_DST) ||
+ (nfc->data & RXH_L4_B_0_1) ||
+ (nfc->data & RXH_L4_B_2_3))
+ return -EINVAL;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ /* if we changed something we need to update flags */
+ if (flags != adapter->flags) {
+ struct e1000_hw *hw = &adapter->hw;
+ u32 mrqc = rd32(E1000_MRQC);
+
+ if ((flags & UDP_RSS_FLAGS) &&
+ !(adapter->flags & UDP_RSS_FLAGS))
+ dev_err(&adapter->pdev->dev,
+ "enabling UDP RSS: fragmented packets may arrive out of order to the stack above\n");
+
+ adapter->flags = flags;
+
+ /* Perform hash on these packet types */
+ mrqc |= E1000_MRQC_RSS_FIELD_IPV4 |
+ E1000_MRQC_RSS_FIELD_IPV4_TCP |
+ E1000_MRQC_RSS_FIELD_IPV6 |
+ E1000_MRQC_RSS_FIELD_IPV6_TCP;
+
+ mrqc &= ~(E1000_MRQC_RSS_FIELD_IPV4_UDP |
+ E1000_MRQC_RSS_FIELD_IPV6_UDP);
+
+ if (flags & IGB_FLAG_RSS_FIELD_IPV4_UDP)
+ mrqc |= E1000_MRQC_RSS_FIELD_IPV4_UDP;
+
+ if (flags & IGB_FLAG_RSS_FIELD_IPV6_UDP)
+ mrqc |= E1000_MRQC_RSS_FIELD_IPV6_UDP;
+
+ wr32(E1000_MRQC, mrqc);
+ }
+
+ return 0;
+}
+
+static int igb_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
+{
+ struct igb_adapter *adapter = netdev_priv(dev);
+ int ret = -EOPNOTSUPP;
+
+ switch (cmd->cmd) {
+ case ETHTOOL_SRXFH:
+ ret = igb_set_rss_hash_opt(adapter, cmd);
+ break;
+ default:
+ break;
+ }
+
+ return ret;
+}
+
static int igb_ethtool_begin(struct net_device *netdev)
{
struct igb_adapter *adapter = netdev_priv(netdev);
@@ -2390,6 +2569,8 @@ static const struct ethtool_ops igb_ethtool_ops = {
.get_coalesce = igb_get_coalesce,
.set_coalesce = igb_set_coalesce,
.get_ts_info = igb_get_ts_info,
+ .get_rxnfc = igb_get_rxnfc,
+ .set_rxnfc = igb_set_rxnfc,
.begin = igb_ethtool_begin,
.complete = igb_ethtool_complete,
};
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 082ce73..da9fd7b 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -2874,18 +2874,21 @@ static void igb_setup_mrqc(struct igb_adapter *adapter)
/* Don't need to set TUOFL or IPOFL, they default to 1 */
wr32(E1000_RXCSUM, rxcsum);
- /*
- * Generate RSS hash based on TCP port numbers and/or
- * IPv4/v6 src and dst addresses since UDP cannot be
- * hashed reliably due to IP fragmentation
- */
+ /* Generate RSS hash based on packet types, TCP/UDP
+ * port numbers and/or IPv4/v6 src and dst addresses
+ */
mrqc = E1000_MRQC_RSS_FIELD_IPV4 |
E1000_MRQC_RSS_FIELD_IPV4_TCP |
E1000_MRQC_RSS_FIELD_IPV6 |
E1000_MRQC_RSS_FIELD_IPV6_TCP |
E1000_MRQC_RSS_FIELD_IPV6_TCP_EX;
+ if (adapter->flags & IGB_FLAG_RSS_FIELD_IPV4_UDP)
+ mrqc |= E1000_MRQC_RSS_FIELD_IPV4_UDP;
+ if (adapter->flags & IGB_FLAG_RSS_FIELD_IPV6_UDP)
+ mrqc |= E1000_MRQC_RSS_FIELD_IPV6_UDP;
+
/* If VMDq is enabled then we set the appropriate mode for that, else
* we default to RSS so that an RSS hash is calculated per packet even
* if we are only using one queue */
--
1.7.11.7
^ permalink raw reply related
* [net-next 08/11] igb: Make TSO check for CHECKSUM_PARTIAL to avoid skb_is_gso check
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
This change is meant to reduce the overhead for workloads that are not
using either TSO or checksum offloads. Most of the time the compiler
should jump ahead after failing this check to the VLAN check since in the
igb_tx_csum call we start with that check as well.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/igb_main.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index da9fd7b..b0730e9 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -3998,6 +3998,9 @@ static int igb_tso(struct igb_ring *tx_ring,
u32 vlan_macip_lens, type_tucmd;
u32 mss_l4len_idx, l4len;
+ if (skb->ip_summed != CHECKSUM_PARTIAL)
+ return 0;
+
if (!skb_is_gso(skb))
return 0;
--
1.7.11.7
^ permalink raw reply related
* [net-next 09/11] igb: Update igb Tx flags to improve code efficiency
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
This change is meant to improve the efficiency of the Tx flags in igb by
aligning them with the values that will later be written into either the
cmd_type or olinfo. By doing this we are able to reduce most of these
functions to either just a simple shift followed by an or in the case of
cmd_type, or an and followed by an or in the case of olinfo.
In order to avoid type conversion errors I also adjusted the locations
where we were switching between CPU and little endian.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/igb.h | 17 ++++++---
drivers/net/ethernet/intel/igb/igb_main.c | 61 +++++++++++++++++--------------
2 files changed, 45 insertions(+), 33 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index b23a667..d8fd5b6 100644
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -150,11 +150,18 @@ struct vf_data_storage {
#define IGB_MNG_VLAN_NONE -1
-#define IGB_TX_FLAGS_CSUM 0x00000001
-#define IGB_TX_FLAGS_VLAN 0x00000002
-#define IGB_TX_FLAGS_TSO 0x00000004
-#define IGB_TX_FLAGS_IPV4 0x00000008
-#define IGB_TX_FLAGS_TSTAMP 0x00000010
+enum igb_tx_flags {
+ /* cmd_type flags */
+ IGB_TX_FLAGS_VLAN = 0x01,
+ IGB_TX_FLAGS_TSO = 0x02,
+ IGB_TX_FLAGS_TSTAMP = 0x04,
+
+ /* olinfo flags */
+ IGB_TX_FLAGS_IPV4 = 0x10,
+ IGB_TX_FLAGS_CSUM = 0x20,
+};
+
+/* VLAN info */
#define IGB_TX_FLAGS_VLAN_MASK 0xffff0000
#define IGB_TX_FLAGS_VLAN_SHIFT 16
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index b0730e9..0a0bd81 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -4121,24 +4121,32 @@ static void igb_tx_csum(struct igb_ring *tx_ring, struct igb_tx_buffer *first)
igb_tx_ctxtdesc(tx_ring, vlan_macip_lens, type_tucmd, mss_l4len_idx);
}
-static __le32 igb_tx_cmd_type(u32 tx_flags)
+#define IGB_SET_FLAG(_input, _flag, _result) \
+ ((_flag <= _result) ? \
+ ((u32)(_input & _flag) * (_result / _flag)) : \
+ ((u32)(_input & _flag) / (_flag / _result)))
+
+static u32 igb_tx_cmd_type(struct sk_buff *skb, u32 tx_flags)
{
/* set type for advanced descriptor with frame checksum insertion */
- __le32 cmd_type = cpu_to_le32(E1000_ADVTXD_DTYP_DATA |
- E1000_ADVTXD_DCMD_IFCS |
- E1000_ADVTXD_DCMD_DEXT);
+ u32 cmd_type = E1000_ADVTXD_DTYP_DATA |
+ E1000_ADVTXD_DCMD_DEXT |
+ E1000_ADVTXD_DCMD_IFCS;
/* set HW vlan bit if vlan is present */
- if (tx_flags & IGB_TX_FLAGS_VLAN)
- cmd_type |= cpu_to_le32(E1000_ADVTXD_DCMD_VLE);
+ cmd_type |= IGB_SET_FLAG(tx_flags, IGB_TX_FLAGS_VLAN,
+ (E1000_ADVTXD_DCMD_VLE));
+
+ /* set segmentation bits for TSO */
+ cmd_type |= IGB_SET_FLAG(tx_flags, IGB_TX_FLAGS_TSO,
+ (E1000_ADVTXD_DCMD_TSE));
/* set timestamp bit if present */
- if (unlikely(tx_flags & IGB_TX_FLAGS_TSTAMP))
- cmd_type |= cpu_to_le32(E1000_ADVTXD_MAC_TSTAMP);
+ cmd_type |= IGB_SET_FLAG(tx_flags, IGB_TX_FLAGS_TSTAMP,
+ (E1000_ADVTXD_MAC_TSTAMP));
- /* set segmentation bits for TSO */
- if (tx_flags & IGB_TX_FLAGS_TSO)
- cmd_type |= cpu_to_le32(E1000_ADVTXD_DCMD_TSE);
+ /* insert frame checksum */
+ cmd_type ^= IGB_SET_FLAG(skb->no_fcs, 1, E1000_ADVTXD_DCMD_IFCS);
return cmd_type;
}
@@ -4149,19 +4157,19 @@ static void igb_tx_olinfo_status(struct igb_ring *tx_ring,
{
u32 olinfo_status = paylen << E1000_ADVTXD_PAYLEN_SHIFT;
- /* 82575 requires a unique index per ring if any offload is enabled */
- if ((tx_flags & (IGB_TX_FLAGS_CSUM | IGB_TX_FLAGS_VLAN)) &&
- test_bit(IGB_RING_FLAG_TX_CTX_IDX, &tx_ring->flags))
+ /* 82575 requires a unique index per ring */
+ if (test_bit(IGB_RING_FLAG_TX_CTX_IDX, &tx_ring->flags))
olinfo_status |= tx_ring->reg_idx << 4;
/* insert L4 checksum */
- if (tx_flags & IGB_TX_FLAGS_CSUM) {
- olinfo_status |= E1000_TXD_POPTS_TXSM << 8;
+ olinfo_status |= IGB_SET_FLAG(tx_flags,
+ IGB_TX_FLAGS_CSUM,
+ (E1000_TXD_POPTS_TXSM << 8));
- /* insert IPv4 checksum */
- if (tx_flags & IGB_TX_FLAGS_IPV4)
- olinfo_status |= E1000_TXD_POPTS_IXSM << 8;
- }
+ /* insert IPv4 checksum */
+ olinfo_status |= IGB_SET_FLAG(tx_flags,
+ IGB_TX_FLAGS_IPV4,
+ (E1000_TXD_POPTS_IXSM << 8));
tx_desc->read.olinfo_status = cpu_to_le32(olinfo_status);
}
@@ -4185,14 +4193,13 @@ static void igb_tx_map(struct igb_ring *tx_ring,
unsigned int data_len = skb->data_len;
unsigned int size = skb_headlen(skb);
unsigned int paylen = skb->len - hdr_len;
- __le32 cmd_type;
u32 tx_flags = first->tx_flags;
+ u32 cmd_type = igb_tx_cmd_type(skb, tx_flags);
u16 i = tx_ring->next_to_use;
tx_desc = IGB_TX_DESC(tx_ring, i);
igb_tx_olinfo_status(tx_ring, tx_desc, tx_flags, paylen);
- cmd_type = igb_tx_cmd_type(tx_flags);
dma = dma_map_single(tx_ring->dev, skb->data, size, DMA_TO_DEVICE);
if (dma_mapping_error(tx_ring->dev, dma))
@@ -4206,7 +4213,7 @@ static void igb_tx_map(struct igb_ring *tx_ring,
for (;;) {
while (unlikely(size > IGB_MAX_DATA_PER_TXD)) {
tx_desc->read.cmd_type_len =
- cmd_type | cpu_to_le32(IGB_MAX_DATA_PER_TXD);
+ cpu_to_le32(cmd_type ^ IGB_MAX_DATA_PER_TXD);
i++;
tx_desc++;
@@ -4225,7 +4232,7 @@ static void igb_tx_map(struct igb_ring *tx_ring,
if (likely(!data_len))
break;
- tx_desc->read.cmd_type_len = cmd_type | cpu_to_le32(size);
+ tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type ^ size);
i++;
tx_desc++;
@@ -4255,10 +4262,8 @@ static void igb_tx_map(struct igb_ring *tx_ring,
netdev_tx_sent_queue(txring_txq(tx_ring), first->bytecount);
/* write last descriptor with RS and EOP bits */
- cmd_type |= cpu_to_le32(size) | cpu_to_le32(IGB_TXD_DCMD);
- if (unlikely(skb->no_fcs))
- cmd_type &= ~(cpu_to_le32(E1000_ADVTXD_DCMD_IFCS));
- tx_desc->read.cmd_type_len = cmd_type;
+ cmd_type |= size | IGB_TXD_DCMD;
+ tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
/* set the timestamp */
first->time_stamp = jiffies;
--
1.7.11.7
^ permalink raw reply related
* [net-next 11/11] igb: Ethtool support to enable and disable EEE
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Akeem G. Abodunrin, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: "Akeem G. Abodunrin" <akeem.g.abodunrin@intel.com>
This patch allows users to enable and disable EEE using Ethtool.
It also allows users to get EEE settings, as supported by the device.
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/e1000_defines.h | 1 +
drivers/net/ethernet/intel/igb/igb_ethtool.c | 100 +++++++++++++++++++++++++
2 files changed, 101 insertions(+)
diff --git a/drivers/net/ethernet/intel/igb/e1000_defines.h b/drivers/net/ethernet/intel/igb/e1000_defines.h
index e647cff..198d148 100644
--- a/drivers/net/ethernet/intel/igb/e1000_defines.h
+++ b/drivers/net/ethernet/intel/igb/e1000_defines.h
@@ -874,6 +874,7 @@
#define E1000_EEER_FRC_AN 0x10000000 /* Enable EEE in loopback */
#define E1000_EEER_LPI_FC 0x00040000 /* EEE Enable on FC */
#define E1000_EEE_SU_LPI_CLK_STP 0X00800000 /* EEE LPI Clock Stop */
+#define E1000_EEER_EEE_NEG 0x20000000 /* EEE capability nego */
/* SerDes Control */
#define E1000_GEN_CTL_READY 0x80000000
diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index 2b82a53..0acf590 100644
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -2529,6 +2529,104 @@ static int igb_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
return ret;
}
+static int igb_get_eee(struct net_device *netdev, struct ethtool_eee *edata)
+{
+ struct igb_adapter *adapter = netdev_priv(netdev);
+ struct e1000_hw *hw = &adapter->hw;
+ u32 ipcnfg, eeer;
+
+ if ((hw->mac.type < e1000_i350) ||
+ (hw->phy.media_type != e1000_media_type_copper))
+ return -EOPNOTSUPP;
+
+ edata->supported = (SUPPORTED_1000baseT_Full |
+ SUPPORTED_100baseT_Full);
+
+ ipcnfg = rd32(E1000_IPCNFG);
+ eeer = rd32(E1000_EEER);
+
+ /* EEE status on negotiated link */
+ if (ipcnfg & E1000_IPCNFG_EEE_1G_AN)
+ edata->advertised = ADVERTISED_1000baseT_Full;
+
+ if (ipcnfg & E1000_IPCNFG_EEE_100M_AN)
+ edata->advertised |= ADVERTISED_100baseT_Full;
+
+ if (eeer & E1000_EEER_EEE_NEG)
+ edata->eee_active = true;
+
+ edata->eee_enabled = !hw->dev_spec._82575.eee_disable;
+
+ if (eeer & E1000_EEER_TX_LPI_EN)
+ edata->tx_lpi_enabled = true;
+
+ /* Report correct negotiated EEE status for devices that
+ * wrongly report EEE at half-duplex
+ */
+ if (adapter->link_duplex == HALF_DUPLEX) {
+ edata->eee_enabled = false;
+ edata->eee_active = false;
+ edata->tx_lpi_enabled = false;
+ edata->advertised &= ~edata->advertised;
+ }
+
+ return 0;
+}
+
+static int igb_set_eee(struct net_device *netdev,
+ struct ethtool_eee *edata)
+{
+ struct igb_adapter *adapter = netdev_priv(netdev);
+ struct e1000_hw *hw = &adapter->hw;
+ struct ethtool_eee eee_curr;
+ s32 ret_val;
+
+ if ((hw->mac.type < e1000_i350) ||
+ (hw->phy.media_type != e1000_media_type_copper))
+ return -EOPNOTSUPP;
+
+ ret_val = igb_get_eee(netdev, &eee_curr);
+ if (ret_val)
+ return ret_val;
+
+ if (eee_curr.eee_enabled) {
+ if (eee_curr.tx_lpi_enabled != edata->tx_lpi_enabled) {
+ dev_err(&adapter->pdev->dev,
+ "Setting EEE tx-lpi is not supported\n");
+ return -EINVAL;
+ }
+
+ /* Tx LPI timer is not implemented currently */
+ if (edata->tx_lpi_timer) {
+ dev_err(&adapter->pdev->dev,
+ "Setting EEE Tx LPI timer is not supported\n");
+ return -EINVAL;
+ }
+
+ if (eee_curr.advertised != edata->advertised) {
+ dev_err(&adapter->pdev->dev,
+ "Setting EEE Advertisement is not supported\n");
+ return -EINVAL;
+ }
+
+ } else if (!edata->eee_enabled) {
+ dev_err(&adapter->pdev->dev,
+ "Setting EEE options are not supported with EEE disabled\n");
+ return -EINVAL;
+ }
+
+ if (hw->dev_spec._82575.eee_disable != !edata->eee_enabled) {
+ hw->dev_spec._82575.eee_disable = !edata->eee_enabled;
+ igb_set_eee_i350(hw);
+
+ /* reset link */
+ if (!netif_running(netdev))
+ igb_reset(adapter);
+ }
+
+ return 0;
+}
+
static int igb_ethtool_begin(struct net_device *netdev)
{
struct igb_adapter *adapter = netdev_priv(netdev);
@@ -2571,6 +2669,8 @@ static const struct ethtool_ops igb_ethtool_ops = {
.get_ts_info = igb_get_ts_info,
.get_rxnfc = igb_get_rxnfc,
.set_rxnfc = igb_set_rxnfc,
+ .get_eee = igb_get_eee,
+ .set_eee = igb_set_eee,
.begin = igb_ethtool_begin,
.complete = igb_ethtool_complete,
};
--
1.7.11.7
^ permalink raw reply related
* [net-next 10/11] igb: Improve performance and reduce size of igb_tx_map
From: Jeff Kirsher @ 2012-11-13 14:03 UTC (permalink / raw)
To: davem; +Cc: Alexander Duyck, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1352815405-751-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Alexander Duyck <alexander.h.duyck@intel.com>
This change is meant to both improve the performance and reduce the size of
igb_tx_map. To do this I have expanded the work done in the main loop by
pushing first into tx_buffer. This allows us to pull in the dma_mapping_error
check, the tx_buffer value assignment, and the initial DMA value assignment to
the Tx descriptor. The net result is that the function reduces in size by a
little over a 100 bytes and is about 1% or 2% faster.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/igb_main.c | 47 +++++++++++++++----------------
1 file changed, 22 insertions(+), 25 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 0a0bd81..7044aaa 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -4188,29 +4188,34 @@ static void igb_tx_map(struct igb_ring *tx_ring,
struct sk_buff *skb = first->skb;
struct igb_tx_buffer *tx_buffer;
union e1000_adv_tx_desc *tx_desc;
+ struct skb_frag_struct *frag;
dma_addr_t dma;
- struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[0];
- unsigned int data_len = skb->data_len;
- unsigned int size = skb_headlen(skb);
- unsigned int paylen = skb->len - hdr_len;
+ unsigned int data_len, size;
u32 tx_flags = first->tx_flags;
u32 cmd_type = igb_tx_cmd_type(skb, tx_flags);
u16 i = tx_ring->next_to_use;
tx_desc = IGB_TX_DESC(tx_ring, i);
- igb_tx_olinfo_status(tx_ring, tx_desc, tx_flags, paylen);
+ igb_tx_olinfo_status(tx_ring, tx_desc, tx_flags, skb->len - hdr_len);
+
+ size = skb_headlen(skb);
+ data_len = skb->data_len;
dma = dma_map_single(tx_ring->dev, skb->data, size, DMA_TO_DEVICE);
- if (dma_mapping_error(tx_ring->dev, dma))
- goto dma_error;
- /* record length, and DMA address */
- dma_unmap_len_set(first, len, size);
- dma_unmap_addr_set(first, dma, dma);
- tx_desc->read.buffer_addr = cpu_to_le64(dma);
+ tx_buffer = first;
+
+ for (frag = &skb_shinfo(skb)->frags[0];; frag++) {
+ if (dma_mapping_error(tx_ring->dev, dma))
+ goto dma_error;
+
+ /* record length, and DMA address */
+ dma_unmap_len_set(tx_buffer, len, size);
+ dma_unmap_addr_set(tx_buffer, dma, dma);
+
+ tx_desc->read.buffer_addr = cpu_to_le64(dma);
- for (;;) {
while (unlikely(size > IGB_MAX_DATA_PER_TXD)) {
tx_desc->read.cmd_type_len =
cpu_to_le32(cmd_type ^ IGB_MAX_DATA_PER_TXD);
@@ -4221,11 +4226,11 @@ static void igb_tx_map(struct igb_ring *tx_ring,
tx_desc = IGB_TX_DESC(tx_ring, 0);
i = 0;
}
+ tx_desc->read.olinfo_status = 0;
dma += IGB_MAX_DATA_PER_TXD;
size -= IGB_MAX_DATA_PER_TXD;
- tx_desc->read.olinfo_status = 0;
tx_desc->read.buffer_addr = cpu_to_le64(dma);
}
@@ -4240,31 +4245,23 @@ static void igb_tx_map(struct igb_ring *tx_ring,
tx_desc = IGB_TX_DESC(tx_ring, 0);
i = 0;
}
+ tx_desc->read.olinfo_status = 0;
size = skb_frag_size(frag);
data_len -= size;
dma = skb_frag_dma_map(tx_ring->dev, frag, 0,
- size, DMA_TO_DEVICE);
- if (dma_mapping_error(tx_ring->dev, dma))
- goto dma_error;
+ size, DMA_TO_DEVICE);
tx_buffer = &tx_ring->tx_buffer_info[i];
- dma_unmap_len_set(tx_buffer, len, size);
- dma_unmap_addr_set(tx_buffer, dma, dma);
-
- tx_desc->read.olinfo_status = 0;
- tx_desc->read.buffer_addr = cpu_to_le64(dma);
-
- frag++;
}
- netdev_tx_sent_queue(txring_txq(tx_ring), first->bytecount);
-
/* write last descriptor with RS and EOP bits */
cmd_type |= size | IGB_TXD_DCMD;
tx_desc->read.cmd_type_len = cpu_to_le32(cmd_type);
+ netdev_tx_sent_queue(txring_txq(tx_ring), first->bytecount);
+
/* set the timestamp */
first->time_stamp = jiffies;
--
1.7.11.7
^ permalink raw reply related
* Re: [net-next 01/11] ixgbe: Do not use DCA to prefetch the entire packet into the cache
From: Eric Dumazet @ 2012-11-13 14:13 UTC (permalink / raw)
To: Jeff Kirsher; +Cc: davem, Alexander Duyck, netdev, gospo, sassmann
In-Reply-To: <1352815405-751-2-git-send-email-jeffrey.t.kirsher@intel.com>
On Tue, 2012-11-13 at 06:03 -0800, Jeff Kirsher wrote:
> From: Alexander Duyck <alexander.h.duyck@intel.com>
>
> The way the code was previously written it was causing DCA to prefetch the
> entire packet into the cache when it was enabled. That is excessive as we
> only really need the headers.
>
> We are now prefetching the headers via software so doing this from DCA would
> be redundant anyway. So clear the bit that was causing us to prefetch the
> packet data and instead only use DCA for the descriptor rings.
>
> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
> Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> ---
Excellent !
My own ixgbe cards are moving so I cant test this, do you guys have some
numbers to share ?
Thanks
^ permalink raw reply
* [PATCH 3.7.0-rc4] of/net/mdio-gpio: Fix pdev->id issue when using devicetrees.
From: Srinivas KANDAGATLA @ 2012-11-13 14:26 UTC (permalink / raw)
To: netdev; +Cc: davem, devicetree-discuss, srinivas.kandagatla
From: Srinivas Kandagatla <srinivas.kandagatla@st.com>
When the mdio-gpio driver is probed via device trees, the platform
device id is set as -1, However the id is re-used in the code while
creating an mdio bus.
So, setting up the id via aliases from device tree is a sensible
solution to fix this issue.
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@st.com>
---
.../devicetree/bindings/net/mdio-gpio.txt | 9 ++++++++-
drivers/net/phy/mdio-gpio.c | 1 +
2 files changed, 9 insertions(+), 1 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/mdio-gpio.txt b/Documentation/devicetree/bindings/net/mdio-gpio.txt
index bc95495..c79bab0 100644
--- a/Documentation/devicetree/bindings/net/mdio-gpio.txt
+++ b/Documentation/devicetree/bindings/net/mdio-gpio.txt
@@ -8,9 +8,16 @@ gpios property as described in section VIII.1 in the following order:
MDC, MDIO.
+Note: Each gpio-mdio bus should have an alias correctly numbered in "aliases"
+node.
+
Example:
-mdio {
+aliases {
+ mdio-gpio0 = <&mdio0>;
+};
+
+mdio0: mdio {
compatible = "virtual,mdio-gpio";
#address-cells = <1>;
#size-cells = <0>;
diff --git a/drivers/net/phy/mdio-gpio.c b/drivers/net/phy/mdio-gpio.c
index 899274f..e3f3115 100644
--- a/drivers/net/phy/mdio-gpio.c
+++ b/drivers/net/phy/mdio-gpio.c
@@ -56,6 +56,7 @@ static void *mdio_gpio_of_get_data(struct platform_device *pdev)
if (ret < 0)
return NULL;
pdata->mdio = ret;
+ pdev->id = of_alias_get_id(np, "mdio-gpio");
return pdata;
}
--
1.7.0.4
^ permalink raw reply related
* Relax, You've Made The Right Choice!
From: kid_from_hell2507 @ 2012-11-13 14:23 UTC (permalink / raw)
To: lotto68
http://bit.ly/PQxeyZ Local Sex Search
^ permalink raw reply
* [PATCH v6] Network driver for the Armada 370 and Armada XP ARM Marvell SoCs
From: Thomas Petazzoni @ 2012-11-13 15:00 UTC (permalink / raw)
To: David S. Miller
Cc: Francois Romieu, Lennert Buytenhek, netdev, linux-arm-kernel,
Jason Cooper, Andrew Lunn, Gregory Clement, Lior Amsalem,
Dmitri Epshtein
David,
This patch set adds a new network driver for the network unit
available in the newest Marvell ARM SoCs Armada 370 and Armada XP, as
well as the necessary Device Tree information to use this driver in
the two evaluation platforms of those SoCs.
The previous versions of this patch set have been sent on September
4th (v1), October 11th (v2), October 23rd (v3), October 26th (v4),
November 12th (v5) and now comes the v6 of the driver. The number of
comments over the last versions have been really small, and I would
really appreciate if this driver could land into the 3.8 kernel
release.
People interested in testing this driver can find it at:
git@github.com:MISL-EBU-System-SW/mainline-public.git marvell-neta-v6
In details:
* Patch 1 contains a small driver for the MDIO interface of this
Ethernet controller. Having a separate driver is useful to more
easily handle concurrent accesses on this MDIO interface that is
shared between all Ethernet ports.
* Patch 3 contains the driver itself. The commit log contains a
detailed explanation about why a new driver is needed for this new
Marvell SoC, compared to older Marvell SoCs (Orion, Kirkwood, Dove)
that use the mv643xx_eth driver.
* Patch 4 adds the necessary entry to the MAINTAINERS file.
* Patch 5 adds the SoC-level Device Tree information for Armada 370
and Armada XP.
* Patch 6 adds the board-level Device Tree information for the
Marvell evaluation boards of Armada 370 and Armada XP.
* Patch 6 adds the board-level Device Tree information for the
PlatHome OpenBlocks AX3-4 platform (based on the Armada XP SoC).
* Patch 7 adds the board-level Device Tree information for the
GlobalScale Mirabox platform (based on the Armada 370 SoC).
Changes since v5:
* Take into account comments from François Romieu (mainly coding
style fixes + addition of a mvneta_rxq_desc_is_first_last() helper
function)
* Fix a wrong argument passed to dma_alloc_coherent():
DMA_BIDIRECTIONAL should have been GFP_KERNEL. Thanks to François
Romieu for having pointed the issue in his review.
* Removed calls to smp_call_function_many() that were useless.
* Ordered alphabetically entries in the Kconfig and Makefile.
Changes since v4:
* Added a separate MDIO driver, which allow to easily handle
concurrent accesses to the MDIO interface.
* The Device Tree now has separate nodes for the PHY devices, which
belong to the MDIO bus handled by the separate MDIO driver.
* Fix tabulation issues in some Device Tree files.
* Rebased on top of 3.7-rc5
* Added the Device Tree code necessary for the GlobalScale Mirabox
platform and the PlatHome OpenBlocks AX3-4 platform.
Changes since v3:
* Use phy_find_first() to get the correct PHY. Suggested by Florian
Fainelli.
* Make pp->cause_rx_tx a simple variable instead of a per-CPU array
since it is not useful. Fixes a comment raised by David Miller.
Changes since v2:
* Change compatible string from 'marvell,neta' to
'marvell,armada-370-neta'. Requested by Rob Herring.
* Rename Ethernet DT nodes from eth@... to ethernet@... Requested by
Rob Herring.
* Remove device_type DT property. Requested by Rob Herring.
* Change the PHY interface for eth0/eth1 to be rgmii-id, which allows
to enable TX/RX delay mechanisms at the PHY level. This fixes CRC
errors on received packets during iperf tests (it was a bug in v2).
* Remove the mvneta_ prefix from module parameters. Requested by
Baruch Siach.
* Many code style improvements suggested by François Romieu.
* Properly stop/restart the TX queue when the number of TX
descriptors available becomes low, instead of returning
NETDEV_TX_BUSY. Requested by François Romieu.
* Properly drop packets on the TX path when DMA mapping functions
return an error, instead of returning NETDEV_TX_BUSY. Requested by
François Romieu.
* Rebased on top of Linux 3.7-rc2.
Changes since v1:
* Reduced the Cc: list in order to make the patch set acceptable for
the netdev@ mailing list.
* Merge the mvneta.h contents into mvneta.c, since the header was
only used by the driver. Requested by Arnd Bergmann.
* Completely reorganize the organization of the register list and
register values, in order to make it more consistent, and hopefully
easier to read (especially easier to match register values with the
corresponding register).
* Integrate with the phylib, as suggested by Florian Fainelli, and
remove the link management code that has become useless as the
result of this integration
* Fix many small details suggested by Florian Fainelli in his review
of the first driver
* Simplify various parts of the driver (descriptors array allocation,
data structures, etc.)
Thanks,
Thomas Petazzoni
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox