* [net-next v2 12/13] i40e: Persistent LLDP support
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem
Cc: Sylwia Wnuczko, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: Sylwia Wnuczko <sylwia.wnuczko@intel.com>
This patch adds a function to read NVM module data and uses it to
read current LLDP agent configuration from NVM API version 1.8.
Signed-off-by: Sylwia Wnuczko <sylwia.wnuczko@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_dcb.c | 18 +++-
drivers/net/ethernet/intel/i40e/i40e_dcb.h | 2 +
drivers/net/ethernet/intel/i40e/i40e_nvm.c | 101 ++++++++++++++++++
.../net/ethernet/intel/i40e/i40e_prototype.h | 6 ++
drivers/net/ethernet/intel/i40e/i40e_type.h | 1 +
5 files changed, 127 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_dcb.c b/drivers/net/ethernet/intel/i40e/i40e_dcb.c
index 292eeb3def10..200a1cb3b536 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_dcb.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_dcb.c
@@ -877,7 +877,23 @@ i40e_status i40e_init_dcb(struct i40e_hw *hw, bool enable_mib_change)
return I40E_NOT_SUPPORTED;
/* Read LLDP NVM area */
- ret = i40e_read_lldp_cfg(hw, &lldp_cfg);
+ if (hw->flags & I40E_HW_FLAG_FW_LLDP_PERSISTENT) {
+ u8 offset = 0;
+
+ if (hw->mac.type == I40E_MAC_XL710)
+ offset = I40E_LLDP_CURRENT_STATUS_XL710_OFFSET;
+ else if (hw->mac.type == I40E_MAC_X722)
+ offset = I40E_LLDP_CURRENT_STATUS_X722_OFFSET;
+ else
+ return I40E_NOT_SUPPORTED;
+
+ ret = i40e_read_nvm_module_data(hw,
+ I40E_SR_EMP_SR_SETTINGS_PTR,
+ offset, 1,
+ &lldp_cfg.adminstatus);
+ } else {
+ ret = i40e_read_lldp_cfg(hw, &lldp_cfg);
+ }
if (ret)
return I40E_ERR_NOT_READY;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_dcb.h b/drivers/net/ethernet/intel/i40e/i40e_dcb.h
index ddb48ae7cce4..2a80c5daa376 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_dcb.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_dcb.h
@@ -30,6 +30,8 @@
#define I40E_CEE_SUBTYPE_APP_PRI 4
#define I40E_CEE_MAX_FEAT_TYPE 3
+#define I40E_LLDP_CURRENT_STATUS_XL710_OFFSET 0x2B
+#define I40E_LLDP_CURRENT_STATUS_X722_OFFSET 0x31
/* Defines for LLDP TLV header */
#define I40E_LLDP_TLV_LEN_SHIFT 0
#define I40E_LLDP_TLV_LEN_MASK (0x01FF << I40E_LLDP_TLV_LEN_SHIFT)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_nvm.c b/drivers/net/ethernet/intel/i40e/i40e_nvm.c
index c508b75c3c09..e4d8d20baf3b 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_nvm.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_nvm.c
@@ -321,6 +321,77 @@ i40e_status i40e_read_nvm_word(struct i40e_hw *hw, u16 offset,
return ret_code;
}
+/**
+ * i40e_read_nvm_module_data - Reads NVM Buffer to specified memory location
+ * @hw: pointer to the HW structure
+ * @module_ptr: Pointer to module in words with respect to NVM beginning
+ * @offset: offset in words from module start
+ * @words_data_size: Words to read from NVM
+ * @data_ptr: Pointer to memory location where resulting buffer will be stored
+ **/
+i40e_status i40e_read_nvm_module_data(struct i40e_hw *hw,
+ u8 module_ptr, u16 offset,
+ u16 words_data_size,
+ u16 *data_ptr)
+{
+ i40e_status status;
+ u16 ptr_value = 0;
+ u32 flat_offset;
+
+ if (module_ptr != 0) {
+ status = i40e_read_nvm_word(hw, module_ptr, &ptr_value);
+ if (status) {
+ i40e_debug(hw, I40E_DEBUG_ALL,
+ "Reading nvm word failed.Error code: %d.\n",
+ status);
+ return I40E_ERR_NVM;
+ }
+ }
+#define I40E_NVM_INVALID_PTR_VAL 0x7FFF
+#define I40E_NVM_INVALID_VAL 0xFFFF
+
+ /* Pointer not initialized */
+ if (ptr_value == I40E_NVM_INVALID_PTR_VAL ||
+ ptr_value == I40E_NVM_INVALID_VAL)
+ return I40E_ERR_BAD_PTR;
+
+ /* Check whether the module is in SR mapped area or outside */
+ if (ptr_value & I40E_PTR_TYPE) {
+ /* Pointer points outside of the Shared RAM mapped area */
+ ptr_value &= ~I40E_PTR_TYPE;
+
+ /* PtrValue in 4kB units, need to convert to words */
+ ptr_value /= 2;
+ flat_offset = ((u32)ptr_value * 0x1000) + (u32)offset;
+ status = i40e_acquire_nvm(hw, I40E_RESOURCE_READ);
+ if (!status) {
+ status = i40e_aq_read_nvm(hw, 0, 2 * flat_offset,
+ 2 * words_data_size,
+ data_ptr, true, NULL);
+ i40e_release_nvm(hw);
+ if (status) {
+ i40e_debug(hw, I40E_DEBUG_ALL,
+ "Reading nvm aq failed.Error code: %d.\n",
+ status);
+ return I40E_ERR_NVM;
+ }
+ } else {
+ return I40E_ERR_NVM;
+ }
+ } else {
+ /* Read from the Shadow RAM */
+ status = i40e_read_nvm_buffer(hw, ptr_value + offset,
+ &words_data_size, data_ptr);
+ if (status) {
+ i40e_debug(hw, I40E_DEBUG_ALL,
+ "Reading nvm buffer failed.Error code: %d.\n",
+ status);
+ }
+ }
+
+ return status;
+}
+
/**
* i40e_read_nvm_buffer_srctl - Reads Shadow RAM buffer via SRCTL register
* @hw: pointer to the HW structure
@@ -429,6 +500,36 @@ static i40e_status __i40e_read_nvm_buffer(struct i40e_hw *hw,
return i40e_read_nvm_buffer_srctl(hw, offset, words, data);
}
+/**
+ * i40e_read_nvm_buffer - Reads Shadow RAM buffer and acquire lock if necessary
+ * @hw: pointer to the HW structure
+ * @offset: offset of the Shadow RAM word to read (0x000000 - 0x001FFF).
+ * @words: (in) number of words to read; (out) number of words actually read
+ * @data: words read from the Shadow RAM
+ *
+ * Reads 16 bit words (data buffer) from the SR using the i40e_read_nvm_srrd()
+ * method. The buffer read is preceded by the NVM ownership take
+ * and followed by the release.
+ **/
+i40e_status i40e_read_nvm_buffer(struct i40e_hw *hw, u16 offset,
+ u16 *words, u16 *data)
+{
+ i40e_status ret_code = 0;
+
+ if (hw->flags & I40E_HW_FLAG_AQ_SRCTL_ACCESS_ENABLE) {
+ ret_code = i40e_acquire_nvm(hw, I40E_RESOURCE_READ);
+ if (!ret_code) {
+ ret_code = i40e_read_nvm_buffer_aq(hw, offset, words,
+ data);
+ i40e_release_nvm(hw);
+ }
+ } else {
+ ret_code = i40e_read_nvm_buffer_srctl(hw, offset, words, data);
+ }
+
+ return ret_code;
+}
+
/**
* i40e_write_nvm_aq - Writes Shadow RAM.
* @hw: pointer to the HW structure.
diff --git a/drivers/net/ethernet/intel/i40e/i40e_prototype.h b/drivers/net/ethernet/intel/i40e/i40e_prototype.h
index 9c810d54df1c..5250441bf75b 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_prototype.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_prototype.h
@@ -315,6 +315,12 @@ i40e_status i40e_acquire_nvm(struct i40e_hw *hw,
void i40e_release_nvm(struct i40e_hw *hw);
i40e_status i40e_read_nvm_word(struct i40e_hw *hw, u16 offset,
u16 *data);
+i40e_status i40e_read_nvm_module_data(struct i40e_hw *hw,
+ u8 module_ptr, u16 offset,
+ u16 words_data_size,
+ u16 *data_ptr);
+i40e_status i40e_read_nvm_buffer(struct i40e_hw *hw, u16 offset,
+ u16 *words, u16 *data);
i40e_status i40e_update_nvm_checksum(struct i40e_hw *hw);
i40e_status i40e_validate_nvm_checksum(struct i40e_hw *hw,
u16 *checksum);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_type.h b/drivers/net/ethernet/intel/i40e/i40e_type.h
index 7c1d57683b53..b43ec94a0f29 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_type.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_type.h
@@ -1318,6 +1318,7 @@ struct i40e_hw_port_stats {
#define I40E_SR_VPD_PTR 0x2F
#define I40E_SR_PCIE_ALT_AUTO_LOAD_PTR 0x3E
#define I40E_SR_SW_CHECKSUM_WORD 0x3F
+#define I40E_SR_EMP_SR_SETTINGS_PTR 0x48
/* Auxiliary field, mask and shift definition for Shadow RAM and NVM Flash */
#define I40E_SR_VPD_MODULE_MAX_SIZE 1024
--
2.21.0
^ permalink raw reply related
* [net-next v2 02/13] i40e: Check if transceiver implements DDM before access
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem
Cc: Mauro S. M. Rodrigues, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: "Mauro S. M. Rodrigues" <maurosr@linux.vnet.ibm.com>
Similar to the ixgbe issue fixed in:
655c91414579 ("ixgbe: Check DDM existence in transceiver before access)
i40e has the same issue when reading eeprom from SFP's module that comply
with SFF-8472 but not implement the Digital Diagnostic Monitoring (DDM)
interface described in it. The existence of such area is specified by bit
6 of byte 92, set to 1 if implemented.
Without this patch, due to not checking this bit i40e fails to read SFP
module's eeprom with the follow message:
ethtool -m enP51p1s0f0
Cannot get Module EEPROM data: Input/output error
Because it fails to read the additional 256 bytes in which it was assumed
to exist the DDM data.
Signed-off-by: "Mauro S. M. Rodrigues" <maurosr@linux.vnet.ibm.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_ethtool.c | 6 ++++++
drivers/net/ethernet/intel/i40e/i40e_type.h | 1 +
2 files changed, 7 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 01e4615b1b4b..41e1240acaea 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -5137,6 +5137,12 @@ static int i40e_get_module_info(struct net_device *netdev,
/* Module is not SFF-8472 compliant */
modinfo->type = ETH_MODULE_SFF_8079;
modinfo->eeprom_len = ETH_MODULE_SFF_8079_LEN;
+ } else if (!(sff8472_swap & I40E_MODULE_SFF_DDM_IMPLEMENTED)) {
+ /* Module is SFF-8472 compliant but doesn't implement
+ * Digital Diagnostic Monitoring (DDM).
+ */
+ modinfo->type = ETH_MODULE_SFF_8079;
+ modinfo->eeprom_len = ETH_MODULE_SFF_8079_LEN;
} else {
modinfo->type = ETH_MODULE_SFF_8472;
modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_type.h b/drivers/net/ethernet/intel/i40e/i40e_type.h
index 8f43aa47c263..2a6219d66771 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_type.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_type.h
@@ -443,6 +443,7 @@ struct i40e_nvm_access {
#define I40E_MODULE_SFF_8472_COMP 0x5E
#define I40E_MODULE_SFF_8472_SWAP 0x5C
#define I40E_MODULE_SFF_ADDR_MODE 0x04
+#define I40E_MODULE_SFF_DDM_IMPLEMENTED 0x40
#define I40E_MODULE_TYPE_QSFP_PLUS 0x0D
#define I40E_MODULE_TYPE_QSFP28 0x11
#define I40E_MODULE_QSFP_MAX_LEN 640
--
2.21.0
^ permalink raw reply related
* [net-next v2 06/13] i40e: check_recovery_mode had wrong if statement
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem
Cc: Adrian Podlawski, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: Adrian Podlawski <adrian.podlawski@intel.com>
Function check_recovery_mode had wrong if statement.
Now we check proper FWS1B register values, which are responsible for
the recovery mode. Recovery mode has 4 values for x710 and 2 for x722.
That's why we need 6 different flags which are defined in the code.
Now in the if statement, we recognize type of mac address
and register value.
Without those changes driver could show wrong state.
Signed-off-by: Adrian Podlawski <adrian.podlawski@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 17 ++++++++++++++---
drivers/net/ethernet/intel/i40e/i40e_register.h | 6 ++++++
2 files changed, 20 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index b807dd6b1417..4551d97771c9 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -14578,9 +14578,20 @@ void i40e_set_fec_in_flags(u8 fec_cfg, u32 *flags)
**/
static bool i40e_check_recovery_mode(struct i40e_pf *pf)
{
- u32 val = rd32(&pf->hw, I40E_GL_FWSTS);
-
- if (val & I40E_GL_FWSTS_FWS1B_MASK) {
+ u32 val = rd32(&pf->hw, I40E_GL_FWSTS) & I40E_GL_FWSTS_FWS1B_MASK;
+ bool is_recovery_mode = false;
+
+ if (pf->hw.mac.type == I40E_MAC_XL710)
+ is_recovery_mode =
+ val == I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_CORER_MASK ||
+ val == I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_GLOBR_MASK ||
+ val == I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_TRANSITION_MASK ||
+ val == I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_NVM_MASK;
+ if (pf->hw.mac.type == I40E_MAC_X722)
+ is_recovery_mode =
+ val == I40E_X722_GL_FWSTS_FWS1B_REC_MOD_CORER_MASK ||
+ val == I40E_X722_GL_FWSTS_FWS1B_REC_MOD_GLOBR_MASK;
+ if (is_recovery_mode) {
dev_notice(&pf->pdev->dev, "Firmware recovery mode detected. Limiting functionality.\n");
dev_notice(&pf->pdev->dev, "Refer to the Intel(R) Ethernet Adapters and Devices User Guide for details on firmware recovery mode.\n");
set_bit(__I40E_RECOVERY_MODE, pf->state);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_register.h b/drivers/net/ethernet/intel/i40e/i40e_register.h
index 330ac19a5dae..d35d690ca10f 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_register.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_register.h
@@ -363,6 +363,12 @@
#define I40E_GL_FWSTS_FWRI_MASK I40E_MASK(0x1, I40E_GL_FWSTS_FWRI_SHIFT)
#define I40E_GL_FWSTS_FWS1B_SHIFT 16
#define I40E_GL_FWSTS_FWS1B_MASK I40E_MASK(0xFF, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_CORER_MASK I40E_MASK(0x30, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_GLOBR_MASK I40E_MASK(0x31, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_TRANSITION_MASK I40E_MASK(0x32, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_XL710_GL_FWSTS_FWS1B_REC_MOD_NVM_MASK I40E_MASK(0x33, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_X722_GL_FWSTS_FWS1B_REC_MOD_CORER_MASK I40E_MASK(0xB, I40E_GL_FWSTS_FWS1B_SHIFT)
+#define I40E_X722_GL_FWSTS_FWS1B_REC_MOD_GLOBR_MASK I40E_MASK(0xC, I40E_GL_FWSTS_FWS1B_SHIFT)
#define I40E_GLGEN_CLKSTAT 0x000B8184 /* Reset: POR */
#define I40E_GLGEN_CLKSTAT_CLKMODE_SHIFT 0
#define I40E_GLGEN_CLKSTAT_CLKMODE_MASK I40E_MASK(0x1, I40E_GLGEN_CLKSTAT_CLKMODE_SHIFT)
--
2.21.0
^ permalink raw reply related
* [net-next v2 05/13] i40e: Add drop mode parameter to set mac config
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem
Cc: Sylwia Wnuczko, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: Sylwia Wnuczko <sylwia.wnuczko@intel.com>
This patch adds "drop mode" parameter to set mac config AQ command.
This bit controls the behavior when a no-drop packet is blocking a TC
queue.
0 – The PF driver is notified.
1 – The blocking packet is dropped and then the PF driver is notified.
Signed-off-by: Sylwia Wnuczko <sylwia.wnuczko@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_adminq.c | 4 ++-
.../net/ethernet/intel/i40e/i40e_adminq_cmd.h | 29 ++++++++++---------
drivers/net/ethernet/intel/i40e/i40e_type.h | 1 +
3 files changed, 19 insertions(+), 15 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq.c b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
index 814acbe79ffd..72c04881d290 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_adminq.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
@@ -610,8 +610,10 @@ i40e_status i40e_init_adminq(struct i40e_hw *hw)
if (hw->aq.api_maj_ver > 1 ||
(hw->aq.api_maj_ver == 1 &&
- hw->aq.api_min_ver >= 8))
+ hw->aq.api_min_ver >= 8)) {
hw->flags |= I40E_HW_FLAG_FW_LLDP_PERSISTENT;
+ hw->flags |= I40E_HW_FLAG_DROP_MODE;
+ }
if (hw->aq.api_maj_ver > I40E_FW_API_VERSION_MAJOR) {
ret_code = I40E_ERR_FIRMWARE_API_VERSION;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h b/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
index 6536023fa074..4d966d80305f 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
@@ -2051,20 +2051,21 @@ I40E_CHECK_CMD_LENGTH(i40e_aq_set_phy_config);
struct i40e_aq_set_mac_config {
__le16 max_frame_size;
u8 params;
-#define I40E_AQ_SET_MAC_CONFIG_CRC_EN 0x04
-#define I40E_AQ_SET_MAC_CONFIG_PACING_MASK 0x78
-#define I40E_AQ_SET_MAC_CONFIG_PACING_SHIFT 3
-#define I40E_AQ_SET_MAC_CONFIG_PACING_NONE 0x0
-#define I40E_AQ_SET_MAC_CONFIG_PACING_1B_13TX 0xF
-#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_9TX 0x9
-#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_4TX 0x8
-#define I40E_AQ_SET_MAC_CONFIG_PACING_3DW_7TX 0x7
-#define I40E_AQ_SET_MAC_CONFIG_PACING_2DW_3TX 0x6
-#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_1TX 0x5
-#define I40E_AQ_SET_MAC_CONFIG_PACING_3DW_2TX 0x4
-#define I40E_AQ_SET_MAC_CONFIG_PACING_7DW_3TX 0x3
-#define I40E_AQ_SET_MAC_CONFIG_PACING_4DW_1TX 0x2
-#define I40E_AQ_SET_MAC_CONFIG_PACING_9DW_1TX 0x1
+#define I40E_AQ_SET_MAC_CONFIG_CRC_EN 0x04
+#define I40E_AQ_SET_MAC_CONFIG_PACING_MASK 0x78
+#define I40E_AQ_SET_MAC_CONFIG_PACING_SHIFT 3
+#define I40E_AQ_SET_MAC_CONFIG_PACING_NONE 0x0
+#define I40E_AQ_SET_MAC_CONFIG_PACING_1B_13TX 0xF
+#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_9TX 0x9
+#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_4TX 0x8
+#define I40E_AQ_SET_MAC_CONFIG_PACING_3DW_7TX 0x7
+#define I40E_AQ_SET_MAC_CONFIG_PACING_2DW_3TX 0x6
+#define I40E_AQ_SET_MAC_CONFIG_PACING_1DW_1TX 0x5
+#define I40E_AQ_SET_MAC_CONFIG_PACING_3DW_2TX 0x4
+#define I40E_AQ_SET_MAC_CONFIG_PACING_7DW_3TX 0x3
+#define I40E_AQ_SET_MAC_CONFIG_PACING_4DW_1TX 0x2
+#define I40E_AQ_SET_MAC_CONFIG_PACING_9DW_1TX 0x1
+#define I40E_AQ_SET_MAC_CONFIG_DROP_BLOCKING_PACKET_EN 0x80
u8 tx_timer_priority; /* bitmap */
__le16 tx_timer_value;
__le16 fc_refresh_threshold;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_type.h b/drivers/net/ethernet/intel/i40e/i40e_type.h
index 2a6219d66771..7c1d57683b53 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_type.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_type.h
@@ -624,6 +624,7 @@ struct i40e_hw {
#define I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK BIT_ULL(3)
#define I40E_HW_FLAG_FW_LLDP_STOPPABLE BIT_ULL(4)
#define I40E_HW_FLAG_FW_LLDP_PERSISTENT BIT_ULL(5)
+#define I40E_HW_FLAG_DROP_MODE BIT_ULL(7)
u64 flags;
/* Used in set switch config AQ command */
--
2.21.0
^ permalink raw reply related
* [net-next v2 07/13] i40e: Update FW API version to 1.9
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem
Cc: Piotr Azarewicz, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: Piotr Azarewicz <piotr.azarewicz@intel.com>
Upcoming FW increment API version to 1.9 due to Extend PHY access AQ
command support. SW is ready for that support as well.
Signed-off-by: Piotr Azarewicz <piotr.azarewicz@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h b/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
index 4d966d80305f..21cccec328e3 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_adminq_cmd.h
@@ -11,8 +11,8 @@
*/
#define I40E_FW_API_VERSION_MAJOR 0x0001
-#define I40E_FW_API_VERSION_MINOR_X722 0x0008
-#define I40E_FW_API_VERSION_MINOR_X710 0x0008
+#define I40E_FW_API_VERSION_MINOR_X722 0x0009
+#define I40E_FW_API_VERSION_MINOR_X710 0x0009
#define I40E_FW_MINOR_VERSION(_h) ((_h)->mac.type == I40E_MAC_XL710 ? \
I40E_FW_API_VERSION_MINOR_X710 : \
--
2.21.0
^ permalink raw reply related
* [net-next v2 03/13] i40e: add check on i40e_configure_tx_ring() return value
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem; +Cc: huhai, netdev, nhorman, sassmann, Andrew Bowers, Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: huhai <huhai@kylinos.cn>
When i40e_configure_tx_ring(vsi->tx_rings[i]) returns an error, we should
exit from i40e_vsi_configure_tx and return the error, instead of continuing
to check whether xdp is enable, and configure the xdp transmit ring.
Signed-off-by: huhai <huhai@kylinos.cn>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 6d456e579314..b807dd6b1417 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -3364,7 +3364,7 @@ static int i40e_vsi_configure_tx(struct i40e_vsi *vsi)
for (i = 0; (i < vsi->num_queue_pairs) && !err; i++)
err = i40e_configure_tx_ring(vsi->tx_rings[i]);
- if (!i40e_enabled_xdp_vsi(vsi))
+ if (err || !i40e_enabled_xdp_vsi(vsi))
return err;
for (i = 0; (i < vsi->num_queue_pairs) && !err; i++)
--
2.21.0
^ permalink raw reply related
* [net-next v2 01/13] i40e: reduce stack usage in i40e_set_fc
From: Jeff Kirsher @ 2019-08-22 20:30 UTC (permalink / raw)
To: davem; +Cc: Arnd Bergmann, netdev, nhorman, sassmann, Andrew Bowers,
Jeff Kirsher
In-Reply-To: <20190822203039.15668-1-jeffrey.t.kirsher@intel.com>
From: Arnd Bergmann <arnd@arndb.de>
The functions i40e_aq_get_phy_abilities_resp() and i40e_set_fc() both
have giant structure on the stack, which makes each one use stack frames
larger than 500 bytes.
As clang decides one function into the other, we get a warning for
exceeding the frame size limit on 32-bit architectures:
drivers/net/ethernet/intel/i40e/i40e_common.c:1654:23: error: stack frame size of 1116 bytes in function 'i40e_set_fc' [-Werror,-Wframe-larger-than=]
When building with gcc, the inlining does not happen, but i40e_set_fc()
calls i40e_aq_get_phy_abilities_resp() anyway, so they add up on the
kernel stack just as much.
The parts that actually use large stacks don't overlap, so make sure
each one is a separate function, and mark them as noinline_for_stack to
prevent the compilers from combining them again.
Fixes: 0a862b43acc6 ("i40e/i40evf: Add module_types and update_link_info")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_common.c | 91 +++++++++++--------
1 file changed, 51 insertions(+), 40 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_common.c b/drivers/net/ethernet/intel/i40e/i40e_common.c
index 906cf68d3453..7af1b7477140 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_common.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_common.c
@@ -1643,25 +1643,15 @@ enum i40e_status_code i40e_aq_set_phy_config(struct i40e_hw *hw,
return status;
}
-/**
- * i40e_set_fc
- * @hw: pointer to the hw struct
- * @aq_failures: buffer to return AdminQ failure information
- * @atomic_restart: whether to enable atomic link restart
- *
- * Set the requested flow control mode using set_phy_config.
- **/
-enum i40e_status_code i40e_set_fc(struct i40e_hw *hw, u8 *aq_failures,
- bool atomic_restart)
+static noinline_for_stack enum i40e_status_code
+i40e_set_fc_status(struct i40e_hw *hw,
+ struct i40e_aq_get_phy_abilities_resp *abilities,
+ bool atomic_restart)
{
- enum i40e_fc_mode fc_mode = hw->fc.requested_mode;
- struct i40e_aq_get_phy_abilities_resp abilities;
struct i40e_aq_set_phy_config config;
- enum i40e_status_code status;
+ enum i40e_fc_mode fc_mode = hw->fc.requested_mode;
u8 pause_mask = 0x0;
- *aq_failures = 0x0;
-
switch (fc_mode) {
case I40E_FC_FULL:
pause_mask |= I40E_AQ_PHY_FLAG_PAUSE_TX;
@@ -1677,6 +1667,48 @@ enum i40e_status_code i40e_set_fc(struct i40e_hw *hw, u8 *aq_failures,
break;
}
+ memset(&config, 0, sizeof(struct i40e_aq_set_phy_config));
+ /* clear the old pause settings */
+ config.abilities = abilities->abilities & ~(I40E_AQ_PHY_FLAG_PAUSE_TX) &
+ ~(I40E_AQ_PHY_FLAG_PAUSE_RX);
+ /* set the new abilities */
+ config.abilities |= pause_mask;
+ /* If the abilities have changed, then set the new config */
+ if (config.abilities == abilities->abilities)
+ return 0;
+
+ /* Auto restart link so settings take effect */
+ if (atomic_restart)
+ config.abilities |= I40E_AQ_PHY_ENABLE_ATOMIC_LINK;
+ /* Copy over all the old settings */
+ config.phy_type = abilities->phy_type;
+ config.phy_type_ext = abilities->phy_type_ext;
+ config.link_speed = abilities->link_speed;
+ config.eee_capability = abilities->eee_capability;
+ config.eeer = abilities->eeer_val;
+ config.low_power_ctrl = abilities->d3_lpan;
+ config.fec_config = abilities->fec_cfg_curr_mod_ext_info &
+ I40E_AQ_PHY_FEC_CONFIG_MASK;
+
+ return i40e_aq_set_phy_config(hw, &config, NULL);
+}
+
+/**
+ * i40e_set_fc
+ * @hw: pointer to the hw struct
+ * @aq_failures: buffer to return AdminQ failure information
+ * @atomic_restart: whether to enable atomic link restart
+ *
+ * Set the requested flow control mode using set_phy_config.
+ **/
+enum i40e_status_code i40e_set_fc(struct i40e_hw *hw, u8 *aq_failures,
+ bool atomic_restart)
+{
+ struct i40e_aq_get_phy_abilities_resp abilities;
+ enum i40e_status_code status;
+
+ *aq_failures = 0x0;
+
/* Get the current phy config */
status = i40e_aq_get_phy_capabilities(hw, false, false, &abilities,
NULL);
@@ -1685,31 +1717,10 @@ enum i40e_status_code i40e_set_fc(struct i40e_hw *hw, u8 *aq_failures,
return status;
}
- memset(&config, 0, sizeof(struct i40e_aq_set_phy_config));
- /* clear the old pause settings */
- config.abilities = abilities.abilities & ~(I40E_AQ_PHY_FLAG_PAUSE_TX) &
- ~(I40E_AQ_PHY_FLAG_PAUSE_RX);
- /* set the new abilities */
- config.abilities |= pause_mask;
- /* If the abilities have changed, then set the new config */
- if (config.abilities != abilities.abilities) {
- /* Auto restart link so settings take effect */
- if (atomic_restart)
- config.abilities |= I40E_AQ_PHY_ENABLE_ATOMIC_LINK;
- /* Copy over all the old settings */
- config.phy_type = abilities.phy_type;
- config.phy_type_ext = abilities.phy_type_ext;
- config.link_speed = abilities.link_speed;
- config.eee_capability = abilities.eee_capability;
- config.eeer = abilities.eeer_val;
- config.low_power_ctrl = abilities.d3_lpan;
- config.fec_config = abilities.fec_cfg_curr_mod_ext_info &
- I40E_AQ_PHY_FEC_CONFIG_MASK;
- status = i40e_aq_set_phy_config(hw, &config, NULL);
+ status = i40e_set_fc_status(hw, &abilities, atomic_restart);
+ if (status)
+ *aq_failures |= I40E_SET_FC_AQ_FAIL_SET;
- if (status)
- *aq_failures |= I40E_SET_FC_AQ_FAIL_SET;
- }
/* Update the link info */
status = i40e_update_link_info(hw);
if (status) {
@@ -2537,7 +2548,7 @@ i40e_status i40e_get_link_status(struct i40e_hw *hw, bool *link_up)
* i40e_updatelink_status - update status of the HW network link
* @hw: pointer to the hw struct
**/
-i40e_status i40e_update_link_info(struct i40e_hw *hw)
+noinline_for_stack i40e_status i40e_update_link_info(struct i40e_hw *hw)
{
struct i40e_aq_get_phy_abilities_resp abilities;
i40e_status status = 0;
--
2.21.0
^ permalink raw reply related
* Re: [PATCH net-next v5] sched: Add dualpi2 qdisc
From: Dave Taht @ 2019-08-22 20:33 UTC (permalink / raw)
To: Tilmans, Olivier (Nokia - BE/Antwerp)
Cc: Eric Dumazet, Stephen Hemminger, Olga Albisser,
De Schepper, Koen (Nokia - BE/Antwerp), Bob Briscoe, Henrik Steen,
Jamal Hadi Salim, Cong Wang, Jiri Pirko, David S. Miller,
linux-kernel@vger.kernel.org, netdev@vger.kernel.org
In-Reply-To: <20190822080045.27609-1-olivier.tilmans@nokia-bell-labs.com>
This is vastly improved code, thank you!
1) Since we're still duking it out over the meaning of the bits - not
just the SCE thing, but as best as I can
tell (but could be wrong) the NQB idea wants to put something into the
l4s fast queue? Or is NQB supposed to
be a third queue?
In those cases, the ecn_mask should just be mask.
2) Is the intent to make the drop probability 0 by default? (10 in the
pie rfc, not mentioned in the l4s rfc as yet)
3) has this been tested on a hw mq system as yet? (10gigE is typically
64 queues)
^ permalink raw reply
* Re: [PATCH] net/ncsi: Fix the payload copying for the request coming from Netlink
From: Vijay Khemka @ 2019-08-22 20:34 UTC (permalink / raw)
To: Justin.Lee1@Dell.com, netdev@vger.kernel.org,
openbmc@lists.ozlabs.org, linux-kernel@vger.kernel.org,
sam@mendozajonas.com, davem@davemloft.net
In-Reply-To: <a94e5fa397a64ae3a676ec11ea09aaba@AUSX13MPS302.AMER.DELL.COM>
LGTM.
On 8/21/19, 2:26 PM, "openbmc on behalf of Justin.Lee1@Dell.com" <openbmc-bounces+vijaykhemka=fb.com@lists.ozlabs.org on behalf of Justin.Lee1@Dell.com> wrote:
The request coming from Netlink should use the OEM generic handler.
The standard command handler expects payload in bytes/words/dwords
but the actual payload is stored in data if the request is coming from Netlink.
Signed-off-by: Justin Lee <justin.lee1@dell.com>
Reviewed-by: Vijay Khemka <vijaykhemka@fb.com>
---
net/ncsi/ncsi-cmd.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/net/ncsi/ncsi-cmd.c b/net/ncsi/ncsi-cmd.c
index eab4346..0187e65 100644
--- a/net/ncsi/ncsi-cmd.c
+++ b/net/ncsi/ncsi-cmd.c
@@ -309,14 +309,21 @@ static struct ncsi_request *ncsi_alloc_command(struct ncsi_cmd_arg *nca)
int ncsi_xmit_cmd(struct ncsi_cmd_arg *nca)
{
+ struct ncsi_cmd_handler *nch = NULL;
struct ncsi_request *nr;
+ unsigned char type;
struct ethhdr *eh;
- struct ncsi_cmd_handler *nch = NULL;
int i, ret;
+ /* Use OEM generic handler for Netlink request */
+ if (nca->req_flags == NCSI_REQ_FLAG_NETLINK_DRIVEN)
+ type = NCSI_PKT_CMD_OEM;
+ else
+ type = nca->type;
+
/* Search for the handler */
for (i = 0; i < ARRAY_SIZE(ncsi_cmd_handlers); i++) {
- if (ncsi_cmd_handlers[i].type == nca->type) {
+ if (ncsi_cmd_handlers[i].type == type) {
if (ncsi_cmd_handlers[i].handler)
nch = &ncsi_cmd_handlers[i];
else
--
2.9.3
^ permalink raw reply
* Re: New skb extension for use by LSMs (skb "security blob")?
From: Casey Schaufler @ 2019-08-22 20:35 UTC (permalink / raw)
To: Florian Westphal
Cc: Paul Moore, netdev, linux-security-module, selinux, casey
In-Reply-To: <20190822201520.GJ20113@breakpoint.cc>
On 8/22/2019 1:15 PM, Florian Westphal wrote:
> Casey Schaufler <casey@schaufler-ca.com> wrote:
>> Given that the original objection to using a skb extension for a
>> security blob was that an extension is dynamic, and that the ubiquitous
>> nature of LSM use makes that unreasonable, it would seem that supporting
>> the security blob as a basic part if the skb would be the obvious and
>> correct solution. If the normal case is that there is an LSM that would
>> befit from the native (unextended) support of a blob, it would seem
>> that that is the case that should be optimized.
> What is this "blob"? i.e., what would you like to add to sk_buff to make
> whatever use cases you have in mind work?
In LSM terminology a blob is a set of data managed and used by
the LSM (either in the infrastructure or the security module).
Blob pointers are included in the system data structures to which
they relate. The inode has an i_security field, which is a void *.
If the secmark where replaced by a security blob, the u32 secmark field
in an sk_buff would be replaced by a void * security field.
^ permalink raw reply
* [pull request][net 0/4] Mellanox, mlx5 fixes 2019-08-22
From: Saeed Mahameed @ 2019-08-22 20:41 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev@vger.kernel.org, Saeed Mahameed
Hi Dave,
This series introduces some fixes to mlx5 driver.
1) Form Moshe, two fixes for firmware health reporter
2) From Eran, two ktls fixes.
Please pull and let me know if there is any problem.
No -stable this time :) ..
Thanks,
Saeed.
---
The following changes since commit cc07db5a5b100bc8eaab5097a23d72f858979750:
gve: Copy and paste bug in gve_get_stats() (2019-08-21 20:50:26 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2019-08-22
for you to fetch changes up to a195784c105b2907b45fd62307d9ce821da9dc20:
net/mlx5e: Remove ethernet segment from dump WQE (2019-08-22 13:38:48 -0700)
----------------------------------------------------------------
mlx5-fixes-2019-08-22
----------------------------------------------------------------
Eran Ben Elisha (2):
net/mlx5e: Add num bytes metadata to WQE info
net/mlx5e: Remove ethernet segment from dump WQE
Moshe Shemesh (2):
net/mlx5: Fix crdump chunks print
net/mlx5: Fix delay in fw fatal report handling due to fw report
.../ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c | 38 ++++++++++------------
drivers/net/ethernet/mellanox/mlx5/core/health.c | 22 +++++++------
2 files changed, 29 insertions(+), 31 deletions(-)
^ permalink raw reply
* [net 1/4] net/mlx5: Fix crdump chunks print
From: Saeed Mahameed @ 2019-08-22 20:41 UTC (permalink / raw)
To: David S. Miller
Cc: netdev@vger.kernel.org, Moshe Shemesh, Eran Ben Elisha,
Saeed Mahameed
In-Reply-To: <20190822204121.16954-1-saeedm@mellanox.com>
From: Moshe Shemesh <moshe@mellanox.com>
Crdump repeats itself every chunk of 256bytes.
That is due to bug of missing progressing offset while copying the data
from buffer to devlink_fmsg.
Fixes: 9b1f29823605 ("net/mlx5: Add support for FW fatal reporter dump")
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/health.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 9314777d99e3..cc5887f52679 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -590,7 +590,8 @@ mlx5_fw_fatal_reporter_dump(struct devlink_health_reporter *reporter,
data_size = crdump_size - offset;
else
data_size = MLX5_CR_DUMP_CHUNK_SIZE;
- err = devlink_fmsg_binary_put(fmsg, cr_data, data_size);
+ err = devlink_fmsg_binary_put(fmsg, (char *)cr_data + offset,
+ data_size);
if (err)
goto free_data;
}
--
2.21.0
^ permalink raw reply related
* [net 2/4] net/mlx5: Fix delay in fw fatal report handling due to fw report
From: Saeed Mahameed @ 2019-08-22 20:41 UTC (permalink / raw)
To: David S. Miller
Cc: netdev@vger.kernel.org, Moshe Shemesh, Eran Ben Elisha,
Saeed Mahameed
In-Reply-To: <20190822204121.16954-1-saeedm@mellanox.com>
From: Moshe Shemesh <moshe@mellanox.com>
When fw fatal error occurs, poll health() first detects and reports on a
fw error. Afterwards, it detects and reports on the fw fatal error
itself.
That can cause a long delay in fw fatal error handling which waits in a
queue for the fw error handling to be finished. The fw error handle will
try asking for fw core dump command while fw in fatal state may not
respond and driver will wait for command timeout.
Changing the flow to detect and handle first fw fatal errors and only if
no fatal error detected look for a fw error to handle.
Fixes: d1bf0e2cc4a6 ("net/mlx5: Report devlink health on FW issues")
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/health.c | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index cc5887f52679..d685122d9ff7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -701,6 +701,16 @@ static void poll_health(struct timer_list *t)
if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
goto out;
+ fatal_error = check_fatal_sensors(dev);
+
+ if (fatal_error && !health->fatal_error) {
+ mlx5_core_err(dev, "Fatal error %u detected\n", fatal_error);
+ dev->priv.health.fatal_error = fatal_error;
+ print_health_info(dev);
+ mlx5_trigger_health_work(dev);
+ goto out;
+ }
+
count = ioread32be(health->health_counter);
if (count == health->prev)
++health->miss_counter;
@@ -719,15 +729,6 @@ static void poll_health(struct timer_list *t)
if (health->synd && health->synd != prev_synd)
queue_work(health->wq, &health->report_work);
- fatal_error = check_fatal_sensors(dev);
-
- if (fatal_error && !health->fatal_error) {
- mlx5_core_err(dev, "Fatal error %u detected\n", fatal_error);
- dev->priv.health.fatal_error = fatal_error;
- print_health_info(dev);
- mlx5_trigger_health_work(dev);
- }
-
out:
mod_timer(&health->timer, get_next_poll_jiffies());
}
--
2.21.0
^ permalink raw reply related
* [net 3/4] net/mlx5e: Add num bytes metadata to WQE info
From: Saeed Mahameed @ 2019-08-22 20:41 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev@vger.kernel.org, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <20190822204121.16954-1-saeedm@mellanox.com>
From: Eran Ben Elisha <eranbe@mellanox.com>
For TLS WQEs, metadata info did not include num_bytes. Due to this issue,
tx_tls_dump_bytes counter did not increment.
Modify tx_fill_wi() to fill num bytes. When it is called for non-traffic
WQE, zero is expected.
Fixes: d2ead1f360e8 ("net/mlx5e: Add kTLS TX HW offload support")
Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
index 8b93101e1a09..0681735ea398 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
@@ -109,13 +109,15 @@ build_progress_params(struct mlx5e_tx_wqe *wqe, u16 pc, u32 sqn,
static void tx_fill_wi(struct mlx5e_txqsq *sq,
u16 pi, u8 num_wqebbs,
- skb_frag_t *resync_dump_frag)
+ skb_frag_t *resync_dump_frag,
+ u32 num_bytes)
{
struct mlx5e_tx_wqe_info *wi = &sq->db.wqe_info[pi];
wi->skb = NULL;
wi->num_wqebbs = num_wqebbs;
wi->resync_dump_frag = resync_dump_frag;
+ wi->num_bytes = num_bytes;
}
void mlx5e_ktls_tx_offload_set_pending(struct mlx5e_ktls_offload_context_tx *priv_tx)
@@ -143,7 +145,7 @@ post_static_params(struct mlx5e_txqsq *sq,
umr_wqe = mlx5e_sq_fetch_wqe(sq, MLX5E_KTLS_STATIC_UMR_WQE_SZ, &pi);
build_static_params(umr_wqe, sq->pc, sq->sqn, priv_tx, fence);
- tx_fill_wi(sq, pi, MLX5E_KTLS_STATIC_WQEBBS, NULL);
+ tx_fill_wi(sq, pi, MLX5E_KTLS_STATIC_WQEBBS, NULL, 0);
sq->pc += MLX5E_KTLS_STATIC_WQEBBS;
}
@@ -157,7 +159,7 @@ post_progress_params(struct mlx5e_txqsq *sq,
wqe = mlx5e_sq_fetch_wqe(sq, MLX5E_KTLS_PROGRESS_WQE_SZ, &pi);
build_progress_params(wqe, sq->pc, sq->sqn, priv_tx, fence);
- tx_fill_wi(sq, pi, MLX5E_KTLS_PROGRESS_WQEBBS, NULL);
+ tx_fill_wi(sq, pi, MLX5E_KTLS_PROGRESS_WQEBBS, NULL, 0);
sq->pc += MLX5E_KTLS_PROGRESS_WQEBBS;
}
@@ -296,7 +298,7 @@ tx_post_resync_dump(struct mlx5e_txqsq *sq, struct sk_buff *skb,
dseg->byte_count = cpu_to_be32(fsz);
mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
- tx_fill_wi(sq, pi, num_wqebbs, frag);
+ tx_fill_wi(sq, pi, num_wqebbs, frag, fsz);
sq->pc += num_wqebbs;
WARN(num_wqebbs > MLX5E_KTLS_MAX_DUMP_WQEBBS,
@@ -323,7 +325,7 @@ static void tx_post_fence_nop(struct mlx5e_txqsq *sq)
struct mlx5_wq_cyc *wq = &sq->wq;
u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
- tx_fill_wi(sq, pi, 1, NULL);
+ tx_fill_wi(sq, pi, 1, NULL, 0);
mlx5e_post_nop_fence(wq, sq->sqn, &sq->pc);
}
--
2.21.0
^ permalink raw reply related
* [net 4/4] net/mlx5e: Remove ethernet segment from dump WQE
From: Saeed Mahameed @ 2019-08-22 20:41 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev@vger.kernel.org, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <20190822204121.16954-1-saeedm@mellanox.com>
From: Eran Ben Elisha <eranbe@mellanox.com>
Dump WQE shall not include Ethernet segment. Define mlx5e_dump_wqe to be
used for "Dump WQEs" instead of sharing it with the general mlx5e_tx_wqe
layout.
Fixes: d2ead1f360e8 ("net/mlx5e: Add kTLS TX HW offload support")
Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/en_accel/ktls_tx.c | 26 +++++++------------
1 file changed, 10 insertions(+), 16 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
index 0681735ea398..7833ddef0427 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
@@ -250,43 +250,37 @@ tx_post_resync_params(struct mlx5e_txqsq *sq,
mlx5e_ktls_tx_post_param_wqes(sq, priv_tx, skip_static_post, true);
}
+struct mlx5e_dump_wqe {
+ struct mlx5_wqe_ctrl_seg ctrl;
+ struct mlx5_wqe_data_seg data;
+};
+
static int
tx_post_resync_dump(struct mlx5e_txqsq *sq, struct sk_buff *skb,
skb_frag_t *frag, u32 tisn, bool first)
{
struct mlx5_wqe_ctrl_seg *cseg;
- struct mlx5_wqe_eth_seg *eseg;
struct mlx5_wqe_data_seg *dseg;
- struct mlx5e_tx_wqe *wqe;
+ struct mlx5e_dump_wqe *wqe;
dma_addr_t dma_addr = 0;
- u16 ds_cnt, ds_cnt_inl;
u8 num_wqebbs;
- u16 pi, ihs;
+ u16 ds_cnt;
int fsz;
-
- ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
- ihs = eth_get_headlen(skb->dev, skb->data, skb_headlen(skb));
- ds_cnt_inl = DIV_ROUND_UP(ihs - INL_HDR_START_SZ, MLX5_SEND_WQE_DS);
- ds_cnt += ds_cnt_inl;
- ds_cnt += 1; /* one frag */
+ u16 pi;
wqe = mlx5e_sq_fetch_wqe(sq, sizeof(*wqe), &pi);
+ ds_cnt = sizeof(*wqe) / MLX5_SEND_WQE_DS;
num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
cseg = &wqe->ctrl;
- eseg = &wqe->eth;
- dseg = wqe->data;
+ dseg = &wqe->data;
cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_DUMP);
cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
cseg->tisn = cpu_to_be32(tisn << 8);
cseg->fm_ce_se = first ? MLX5_FENCE_MODE_INITIATOR_SMALL : 0;
- eseg->inline_hdr.sz = cpu_to_be16(ihs);
- memcpy(eseg->inline_hdr.start, skb->data, ihs);
- dseg += ds_cnt_inl;
-
fsz = skb_frag_size(frag);
dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
DMA_TO_DEVICE);
--
2.21.0
^ permalink raw reply related
* [PATCH v2] riscv: add support for SECCOMP and SECCOMP_FILTER
From: David Abdurachmanov @ 2019-08-22 20:55 UTC (permalink / raw)
To: Paul Walmsley, Palmer Dabbelt, Albert Ou, Oleg Nesterov,
Kees Cook, Andy Lutomirski, Will Drewry, Shuah Khan,
Alexei Starovoitov, Daniel Borkmann, Martin KaFai Lau, Song Liu,
Yonghong Song, David Abdurachmanov, Thomas Gleixner,
Allison Randal, Alexios Zavras, Anup Patel, Vincent Chen,
Alan Kao, linux-riscv, linux-kernel, linux-kselftest, netdev, bpf
Cc: david.abdurachmanov, me
This patch was extensively tested on Fedora/RISCV (applied by default on
top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc
on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased:
https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org
Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov <david.abdurachmanov@sifive.com>
---
arch/riscv/Kconfig | 14 ++++++++++
arch/riscv/include/asm/seccomp.h | 10 +++++++
arch/riscv/include/asm/thread_info.h | 5 +++-
arch/riscv/kernel/entry.S | 27 +++++++++++++++++--
arch/riscv/kernel/ptrace.c | 10 +++++++
tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++-
6 files changed, 70 insertions(+), 4 deletions(-)
create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index 59a4727ecd6c..441e63ff5adc 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -31,6 +31,7 @@ config RISCV
select GENERIC_SMP_IDLE_THREAD
select GENERIC_ATOMIC64 if !64BIT
select HAVE_ARCH_AUDITSYSCALL
+ select HAVE_ARCH_SECCOMP_FILTER
select HAVE_MEMBLOCK_NODE_MAP
select HAVE_DMA_CONTIGUOUS
select HAVE_FUTEX_CMPXCHG if FUTEX
@@ -235,6 +236,19 @@ menu "Kernel features"
source "kernel/Kconfig.hz"
+config SECCOMP
+ bool "Enable seccomp to safely compute untrusted bytecode"
+ help
+ This kernel feature is useful for number crunching applications
+ that may need to compute untrusted bytecode during their
+ execution. By using pipes or other transports made available to
+ the process as file descriptors supporting the read/write
+ syscalls, it's possible to isolate those applications in
+ their own address space using seccomp. Once seccomp is
+ enabled via prctl(PR_SET_SECCOMP), it cannot be disabled
+ and the task is only allowed to execute a few safe syscalls
+ defined by each seccomp mode.
+
endmenu
menu "Boot options"
diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h
new file mode 100644
index 000000000000..bf7744ee3b3d
--- /dev/null
+++ b/arch/riscv/include/asm/seccomp.h
@@ -0,0 +1,10 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef _ASM_SECCOMP_H
+#define _ASM_SECCOMP_H
+
+#include <asm/unistd.h>
+
+#include <asm-generic/seccomp.h>
+
+#endif /* _ASM_SECCOMP_H */
diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h
index 905372d7eeb8..a0b2a29a0da1 100644
--- a/arch/riscv/include/asm/thread_info.h
+++ b/arch/riscv/include/asm/thread_info.h
@@ -75,6 +75,7 @@ struct thread_info {
#define TIF_MEMDIE 5 /* is terminating due to OOM killer */
#define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */
#define TIF_SYSCALL_AUDIT 7 /* syscall auditing */
+#define TIF_SECCOMP 8 /* syscall secure computing */
#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE)
#define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME)
@@ -82,11 +83,13 @@ struct thread_info {
#define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED)
#define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT)
#define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT)
+#define _TIF_SECCOMP (1 << TIF_SECCOMP)
#define _TIF_WORK_MASK \
(_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED)
#define _TIF_SYSCALL_WORK \
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
+ (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \
+ _TIF_SECCOMP )
#endif /* _ASM_RISCV_THREAD_INFO_H */
diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S
index bc7a56e1ca6f..0bbedfa3e47d 100644
--- a/arch/riscv/kernel/entry.S
+++ b/arch/riscv/kernel/entry.S
@@ -203,8 +203,25 @@ check_syscall_nr:
/* Check to make sure we don't jump to a bogus syscall number. */
li t0, __NR_syscalls
la s0, sys_ni_syscall
- /* Syscall number held in a7 */
- bgeu a7, t0, 1f
+ /*
+ * The tracer can change syscall number to valid/invalid value.
+ * We use syscall_set_nr helper in syscall_trace_enter thus we
+ * cannot trust the current value in a7 and have to reload from
+ * the current task pt_regs.
+ */
+ REG_L a7, PT_A7(sp)
+ /*
+ * Syscall number held in a7.
+ * If syscall number is above allowed value, redirect to ni_syscall.
+ */
+ bge a7, t0, 1f
+ /*
+ * Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.
+ * If yes, we pretend it was executed.
+ */
+ li t1, -1
+ beq a7, t1, ret_from_syscall_rejected
+ /* Call syscall */
la s0, sys_call_table
slli t0, a7, RISCV_LGPTR
add s0, s0, t0
@@ -215,6 +232,12 @@ check_syscall_nr:
ret_from_syscall:
/* Set user a0 to kernel a0 */
REG_S a0, PT_A0(sp)
+ /*
+ * We didn't execute the actual syscall.
+ * Seccomp already set return value for the current task pt_regs.
+ * (If it was configured with SECCOMP_RET_ERRNO/TRACE)
+ */
+ret_from_syscall_rejected:
/* Trace syscalls, but only if requested by the user. */
REG_L t0, TASK_TI_FLAGS(tp)
andi t0, t0, _TIF_SYSCALL_WORK
diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c
index 368751438366..63e47c9f85f0 100644
--- a/arch/riscv/kernel/ptrace.c
+++ b/arch/riscv/kernel/ptrace.c
@@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs)
if (tracehook_report_syscall_entry(regs))
syscall_set_nr(current, regs, -1);
+ /*
+ * Do the secure computing after ptrace; failures should be fast.
+ * If this fails we might have return value in a0 from seccomp
+ * (via SECCOMP_RET_ERRNO/TRACE).
+ */
+ if (secure_computing(NULL) == -1) {
+ syscall_set_nr(current, regs, -1);
+ return;
+ }
+
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
if (test_thread_flag(TIF_SYSCALL_TRACEPOINT))
trace_sys_enter(regs, syscall_get_nr(current, regs));
diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c
index 6ef7f16c4cf5..492e0adad9d3 100644
--- a/tools/testing/selftests/seccomp/seccomp_bpf.c
+++ b/tools/testing/selftests/seccomp/seccomp_bpf.c
@@ -112,6 +112,8 @@ struct seccomp_data {
# define __NR_seccomp 383
# elif defined(__aarch64__)
# define __NR_seccomp 277
+# elif defined(__riscv)
+# define __NR_seccomp 277
# elif defined(__hppa__)
# define __NR_seccomp 338
# elif defined(__powerpc__)
@@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally)
# define ARCH_REGS struct user_pt_regs
# define SYSCALL_NUM regs[8]
# define SYSCALL_RET regs[0]
+#elif defined(__riscv) && __riscv_xlen == 64
+# define ARCH_REGS struct user_regs_struct
+# define SYSCALL_NUM a7
+# define SYSCALL_RET a0
#elif defined(__hppa__)
# define ARCH_REGS struct user_regs_struct
# define SYSCALL_NUM gr[20]
@@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata,
EXPECT_EQ(0, ret) {}
#if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
+ defined(__s390__) || defined(__hppa__) || defined(__riscv)
{
regs.SYSCALL_NUM = syscall;
}
--
2.21.0
^ permalink raw reply related
* Re: [PATCH net-next v2 2/3] net: ethernet: mediatek: Re-add support SGMII
From: Russell King - ARM Linux admin @ 2019-08-22 21:11 UTC (permalink / raw)
To: René van Dorst
Cc: Nelson Chang, Frank Wunderlich, netdev, Sean Wang, linux-mips,
linux-mediatek, John Crispin, Matthias Brugger, Stefan Roese,
David S . Miller, linux-arm-kernel
In-Reply-To: <20190822195033.Horde.hEW8FBGNfFrugQOCv0gaDfx@www.vdorst.com>
Hi Rene,
On Thu, Aug 22, 2019 at 07:50:33PM +0000, René van Dorst wrote:
> Quoting Russell King - ARM Linux admin <linux@armlinux.org.uk>:
> > Isn't this set for Cisco SGMII as well as for 802.3 1000BASE-X and
> > the up-clocked 2500BASE-X modes?
> >
> > If so, is there a reason why 10Mbps and 100Mbps speeds aren't
> > supported on Cisco SGMII links?
>
> I can only tell a bit about the mt7622 SOC, datasheet tells me that:
>
> The SGMII is the interface between 10/100/1000/2500 Mbps PHY and Ethernet MAC,
> the spec is raised by Cisco in 1999, which is aims for pin reduction compare
> with the GMII. It uses 2 differential data pair for TX and RX with clock
> embedded bit stream to convey frame data and port ability information.
> The core leverages the 1000Base-X PCS and Auto-Negotiation from IEEE 802.3
> specification (clause 36/37). This IP can support up to 3.125G baud for
> 2.5Gbps
> (proprietary 2500Base-X) data rate of MAC by overclocking.
>
> Also features tells me: Support 10/100/1000/2500 Mbps in full duplex mode and
> 10/100 Mbps in half duplex mode.
Yep, that is what I'd expect. Thanks.
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
^ permalink raw reply
* Re: [PATCH] net/ipv4/tcp_bpf: Delete an unnecessary check before the function call “consume_skb”
From: Song Liu @ 2019-08-22 21:12 UTC (permalink / raw)
To: Markus Elfring
Cc: Networking, bpf, Alexei Starovoitov, Alexey Kuznetsov,
Daniel Borkmann, David S. Miller, Eric Dumazet, Hideaki Yoshifuji,
John Fastabend, Martin KaFai Lau, Song Liu, Yonghong Song, LKML,
kernel-janitors
In-Reply-To: <25ca9c48-1ac9-daa1-8472-0a53e4beed6a@web.de>
On Thu, Aug 22, 2019 at 10:19 AM Markus Elfring <Markus.Elfring@web.de> wrote:
>
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Thu, 22 Aug 2019 18:20:42 +0200
>
> The consume_skb() function performs also input parameter validation.
> Thus the test around the call is not needed.
>
> This issue was detected by using the Coccinelle software.
>
> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Song Liu <songliubraving@fb.com>
^ permalink raw reply
* Re: [PATCH] net/core/skmsg: Delete an unnecessary check before the function call “consume_skb”
From: Song Liu @ 2019-08-22 21:13 UTC (permalink / raw)
To: Markus Elfring
Cc: Networking, bpf, Daniel Borkmann, David S. Miller, John Fastabend,
LKML, kernel-janitors
In-Reply-To: <2755f638-7846-91f2-74f4-61031f3e34c8@web.de>
On Thu, Aug 22, 2019 at 10:18 AM Markus Elfring <Markus.Elfring@web.de> wrote:
>
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Thu, 22 Aug 2019 18:00:40 +0200
>
> The consume_skb() function performs also input parameter validation.
> Thus the test around the call is not needed.
>
> This issue was detected by using the Coccinelle software.
>
> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Song Liu <songliubraving@fb.com>
^ permalink raw reply
* [PATCH v2 4/5] spi: spi-fsl-dspi: Use poll mode in case the platform IRQ is missing
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie; +Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean
In-Reply-To: <20190822211514.19288-1-olteanv@gmail.com>
On platforms like LS1021A which use TCFQ mode, an interrupt needs to be
processed after each byte is TXed/RXed. I tried to make the DSPI
implementation on this SoC operate in other, more efficient modes (EOQ,
DMA) but it looks like it simply isn't possible.
Therefore allow the driver to operate in poll mode, to ease a bit of
this absurd amount of IRQ load generated in TCFQ mode. Doing so reduces
both the net time it takes to transmit a SPI message, as well as the
inter-frame jitter that occurs while doing so.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
drivers/spi/spi-fsl-dspi.c | 87 ++++++++++++++++++++++++++++----------
1 file changed, 64 insertions(+), 23 deletions(-)
diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index 6d2c7984ab0e..77db43f1290f 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -647,19 +647,12 @@ static void dspi_eoq_read(struct fsl_dspi *dspi)
dspi_push_rx(dspi, fifo_read(dspi));
}
-static irqreturn_t dspi_interrupt(int irq, void *dev_id)
+static int dspi_rxtx(struct fsl_dspi *dspi)
{
- struct fsl_dspi *dspi = (struct fsl_dspi *)dev_id;
struct spi_message *msg = dspi->cur_msg;
enum dspi_trans_mode trans_mode;
- u32 spi_sr, spi_tcr;
u16 spi_tcnt;
-
- regmap_read(dspi->regmap, SPI_SR, &spi_sr);
- regmap_write(dspi->regmap, SPI_SR, spi_sr);
-
- if (!(spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)))
- return IRQ_NONE;
+ u32 spi_tcr;
/* Get transfer counter (in number of SPI transfers). It was
* reset to 0 when transfer(s) were started.
@@ -675,17 +668,55 @@ static irqreturn_t dspi_interrupt(int irq, void *dev_id)
else if (trans_mode == DSPI_TCFQ_MODE)
dspi_tcfq_read(dspi);
- if (!dspi->len) {
- dspi->waitflags = 1;
- wake_up_interruptible(&dspi->waitq);
- return IRQ_HANDLED;
- }
+ if (!dspi->len)
+ /* Success! */
+ return 0;
if (trans_mode == DSPI_EOQ_MODE)
dspi_eoq_write(dspi);
else if (trans_mode == DSPI_TCFQ_MODE)
dspi_tcfq_write(dspi);
+ return -EINPROGRESS;
+}
+
+static int dspi_poll(struct fsl_dspi *dspi)
+{
+ int tries = 1000;
+ u32 spi_sr;
+
+ do {
+ regmap_read(dspi->regmap, SPI_SR, &spi_sr);
+ regmap_write(dspi->regmap, SPI_SR, spi_sr);
+
+ if (spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF))
+ break;
+ } while (--tries);
+
+ if (!tries)
+ return -ETIMEDOUT;
+
+ return dspi_rxtx(dspi);
+}
+
+static irqreturn_t dspi_interrupt(int irq, void *dev_id)
+{
+ struct fsl_dspi *dspi = (struct fsl_dspi *)dev_id;
+ u32 spi_sr;
+
+ regmap_read(dspi->regmap, SPI_SR, &spi_sr);
+ regmap_write(dspi->regmap, SPI_SR, spi_sr);
+
+ if (!(spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)))
+ return IRQ_NONE;
+
+ dspi_rxtx(dspi);
+
+ if (!dspi->len) {
+ dspi->waitflags = 1;
+ wake_up_interruptible(&dspi->waitq);
+ }
+
return IRQ_HANDLED;
}
@@ -773,13 +804,18 @@ static int dspi_transfer_one_message(struct spi_controller *ctlr,
goto out;
}
- if (trans_mode != DSPI_DMA_MODE) {
- if (wait_event_interruptible(dspi->waitq,
- dspi->waitflags))
- dev_err(&dspi->pdev->dev,
- "wait transfer complete fail!\n");
+ if (!dspi->irq) {
+ do {
+ status = dspi_poll(dspi);
+ } while (status == -EINPROGRESS);
+ } else if (trans_mode != DSPI_DMA_MODE) {
+ status = wait_event_interruptible(dspi->waitq,
+ dspi->waitflags);
dspi->waitflags = 0;
}
+ if (status)
+ dev_err(&dspi->pdev->dev,
+ "Waiting for transfer to complete failed!\n");
if (transfer->delay_usecs)
udelay(transfer->delay_usecs);
@@ -1079,10 +1115,13 @@ static int dspi_probe(struct platform_device *pdev)
goto out_ctlr_put;
dspi_init(dspi);
+
dspi->irq = platform_get_irq(pdev, 0);
- if (dspi->irq < 0) {
- ret = dspi->irq;
- goto out_clk_put;
+ if (dspi->irq <= 0) {
+ dev_info(&pdev->dev,
+ "can't get platform irq, using poll mode\n");
+ dspi->irq = 0;
+ goto poll_mode;
}
ret = devm_request_irq(&pdev->dev, dspi->irq, dspi_interrupt,
@@ -1092,6 +1131,9 @@ static int dspi_probe(struct platform_device *pdev)
goto out_clk_put;
}
+ init_waitqueue_head(&dspi->waitq);
+
+poll_mode:
if (dspi->devtype_data->trans_mode == DSPI_DMA_MODE) {
ret = dspi_request_dma(dspi, res->start);
if (ret < 0) {
@@ -1103,7 +1145,6 @@ static int dspi_probe(struct platform_device *pdev)
ctlr->max_speed_hz =
clk_get_rate(dspi->clk) / dspi->devtype_data->max_clock_factor;
- init_waitqueue_head(&dspi->waitq);
platform_set_drvdata(pdev, ctlr);
ret = spi_register_controller(ctlr);
--
2.17.1
^ permalink raw reply related
* [PATCH v2 5/5] ARM: dts: ls1021a-tsn: Use the DSPI controller in poll mode
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie
Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean,
Rob Herring, Shawn Guo
In-Reply-To: <20190822211514.19288-1-olteanv@gmail.com>
Connected to the LS1021A DSPI is the SJA1105 DSA switch. This
constitutes 4 of the 6 Ethernet ports on this board.
As the SJA1105 is a PTP switch, constant disciplining of its PTP clock
is necessary, and that translates into a lot of SPI I/O even when
otherwise idle.
Switching to using the DSPI in poll mode has several distinct
benefits:
- With interrupts, the DSPI driver in TCFQ mode raises an IRQ after each
transmitted byte. There is more time wasted for the "waitq" event than
for actual I/O. And the DSPI IRQ count is by far the largest in
/proc/interrupts on this board (larger than Ethernet). I should
mention that due to various LS1021A errata, other operating modes than
TCFQ are not available.
- The SPI I/O time is both lower, and more consistently so. For a TSN
switch it is important that all SPI transfers take a deterministic
time to complete.
Reading the PTP clock is an important example.
Egressing through the switch requires some setup in advance (an SPI
write command). Without this patch, that operation required a
--tx_timestamp_timeout 50 (ms), now it can be done with
--tx_timestamp_timeout 10.
Yet another example is reconstructing timestamps, which has a hard
deadline because the PTP timestamping counter wraps around in 0.135
seconds. Combined with other I/O needed for that to happen, there is
a real risk that the deadline is not always met.
See drivers/net/dsa/sja1105/ for more info about the above.
Cc: Rob Herring <robh@kernel.org>
Cc: Shawn Guo <shawnguo@kernel.org>
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
arch/arm/boot/dts/ls1021a-tsn.dts | 1 +
1 file changed, 1 insertion(+)
diff --git a/arch/arm/boot/dts/ls1021a-tsn.dts b/arch/arm/boot/dts/ls1021a-tsn.dts
index 5b7689094b70..1c09cfc766af 100644
--- a/arch/arm/boot/dts/ls1021a-tsn.dts
+++ b/arch/arm/boot/dts/ls1021a-tsn.dts
@@ -33,6 +33,7 @@
};
&dspi0 {
+ /delete-property/ interrupts;
bus-num = <0>;
status = "okay";
--
2.17.1
^ permalink raw reply related
* [PATCH v2 2/5] spi: spi-fsl-dspi: Exit the ISR with IRQ_NONE when it's not ours
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie; +Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean
In-Reply-To: <20190822211514.19288-1-olteanv@gmail.com>
The DSPI interrupt can be shared between two controllers at least on the
LX2160A. In that case, the driver for one controller might misbehave and
consume the other's interrupt. Fix this by actually checking if any of
the bits in the status register have been asserted.
Fixes: 13aed2392741 ("spi: spi-fsl-dspi: use IRQF_SHARED mode to request IRQ")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
drivers/spi/spi-fsl-dspi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index c90db7db4121..6ef2279a3699 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -659,7 +659,7 @@ static irqreturn_t dspi_interrupt(int irq, void *dev_id)
regmap_write(dspi->regmap, SPI_SR, spi_sr);
if (!(spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)))
- return IRQ_HANDLED;
+ return IRQ_NONE;
/* Get transfer counter (in number of SPI transfers). It was
* reset to 0 when transfer(s) were started.
--
2.17.1
^ permalink raw reply related
* [PATCH v2 1/5] spi: spi-fsl-dspi: Reduce indentation level in dspi_interrupt
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie; +Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean
In-Reply-To: <20190822211514.19288-1-olteanv@gmail.com>
If the entire function depends on the SPI status register having the
interrupt bits asserted, then just check it and exit early if those bits
aren't set (such as in the case of the shared IRQ being triggered for
the other peripheral). Cosmetic patch.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
drivers/spi/spi-fsl-dspi.c | 79 +++++++++++++++++++-------------------
1 file changed, 40 insertions(+), 39 deletions(-)
diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index 790cb02fc181..c90db7db4121 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -658,47 +658,48 @@ static irqreturn_t dspi_interrupt(int irq, void *dev_id)
regmap_read(dspi->regmap, SPI_SR, &spi_sr);
regmap_write(dspi->regmap, SPI_SR, spi_sr);
+ if (!(spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)))
+ return IRQ_HANDLED;
+
+ /* Get transfer counter (in number of SPI transfers). It was
+ * reset to 0 when transfer(s) were started.
+ */
+ regmap_read(dspi->regmap, SPI_TCR, &spi_tcr);
+ spi_tcnt = SPI_TCR_GET_TCNT(spi_tcr);
+ /* Update total number of bytes that were transferred */
+ msg->actual_length += spi_tcnt * dspi->bytes_per_word;
+
+ trans_mode = dspi->devtype_data->trans_mode;
+ switch (trans_mode) {
+ case DSPI_EOQ_MODE:
+ dspi_eoq_read(dspi);
+ break;
+ case DSPI_TCFQ_MODE:
+ dspi_tcfq_read(dspi);
+ break;
+ default:
+ dev_err(&dspi->pdev->dev, "unsupported trans_mode %u\n",
+ trans_mode);
+ return IRQ_HANDLED;
+ }
- if (spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)) {
- /* Get transfer counter (in number of SPI transfers). It was
- * reset to 0 when transfer(s) were started.
- */
- regmap_read(dspi->regmap, SPI_TCR, &spi_tcr);
- spi_tcnt = SPI_TCR_GET_TCNT(spi_tcr);
- /* Update total number of bytes that were transferred */
- msg->actual_length += spi_tcnt * dspi->bytes_per_word;
-
- trans_mode = dspi->devtype_data->trans_mode;
- switch (trans_mode) {
- case DSPI_EOQ_MODE:
- dspi_eoq_read(dspi);
- break;
- case DSPI_TCFQ_MODE:
- dspi_tcfq_read(dspi);
- break;
- default:
- dev_err(&dspi->pdev->dev, "unsupported trans_mode %u\n",
- trans_mode);
- return IRQ_HANDLED;
- }
+ if (!dspi->len) {
+ dspi->waitflags = 1;
+ wake_up_interruptible(&dspi->waitq);
+ return IRQ_HANDLED;
+ }
- if (!dspi->len) {
- dspi->waitflags = 1;
- wake_up_interruptible(&dspi->waitq);
- } else {
- switch (trans_mode) {
- case DSPI_EOQ_MODE:
- dspi_eoq_write(dspi);
- break;
- case DSPI_TCFQ_MODE:
- dspi_tcfq_write(dspi);
- break;
- default:
- dev_err(&dspi->pdev->dev,
- "unsupported trans_mode %u\n",
- trans_mode);
- }
- }
+ switch (trans_mode) {
+ case DSPI_EOQ_MODE:
+ dspi_eoq_write(dspi);
+ break;
+ case DSPI_TCFQ_MODE:
+ dspi_tcfq_write(dspi);
+ break;
+ default:
+ dev_err(&dspi->pdev->dev,
+ "unsupported trans_mode %u\n",
+ trans_mode);
}
return IRQ_HANDLED;
--
2.17.1
^ permalink raw reply related
* [PATCH v2 3/5] spi: spi-fsl-dspi: Remove impossible to reach error check
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie; +Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean
In-Reply-To: <20190822211514.19288-1-olteanv@gmail.com>
dspi->devtype_data is under the total control of the driver. Therefore,
a bad value is a driver bug and checking it at runtime (and during an
ISR, at that!) is pointless.
The second "else if" check is only for clarity (instead of a broader
"else") in case other transfer modes are added in the future. But the
printing is dead code and can be removed.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
---
drivers/spi/spi-fsl-dspi.c | 24 ++++--------------------
1 file changed, 4 insertions(+), 20 deletions(-)
diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index 6ef2279a3699..6d2c7984ab0e 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -670,18 +670,10 @@ static irqreturn_t dspi_interrupt(int irq, void *dev_id)
msg->actual_length += spi_tcnt * dspi->bytes_per_word;
trans_mode = dspi->devtype_data->trans_mode;
- switch (trans_mode) {
- case DSPI_EOQ_MODE:
+ if (trans_mode == DSPI_EOQ_MODE)
dspi_eoq_read(dspi);
- break;
- case DSPI_TCFQ_MODE:
+ else if (trans_mode == DSPI_TCFQ_MODE)
dspi_tcfq_read(dspi);
- break;
- default:
- dev_err(&dspi->pdev->dev, "unsupported trans_mode %u\n",
- trans_mode);
- return IRQ_HANDLED;
- }
if (!dspi->len) {
dspi->waitflags = 1;
@@ -689,18 +681,10 @@ static irqreturn_t dspi_interrupt(int irq, void *dev_id)
return IRQ_HANDLED;
}
- switch (trans_mode) {
- case DSPI_EOQ_MODE:
+ if (trans_mode == DSPI_EOQ_MODE)
dspi_eoq_write(dspi);
- break;
- case DSPI_TCFQ_MODE:
+ else if (trans_mode == DSPI_TCFQ_MODE)
dspi_tcfq_write(dspi);
- break;
- default:
- dev_err(&dspi->pdev->dev,
- "unsupported trans_mode %u\n",
- trans_mode);
- }
return IRQ_HANDLED;
}
--
2.17.1
^ permalink raw reply related
* [PATCH v2 0/5] Poll mode for NXP DSPI driver
From: Vladimir Oltean @ 2019-08-22 21:15 UTC (permalink / raw)
To: broonie; +Cc: linux-spi, linux-kernel, devicetree, netdev, Vladimir Oltean
This series makes the EOQ and TCFQ operating modes of spi-fsl-dspi work
when no platform IRQ is defined in the DT.
The NXP LS1021A-TSN board needs this setting for performance reasons,
and I am adding the corresponding DTS patch to this series to avoid
runtime errors that may occur if the DTS and the driver patches part
ways and get included through separate trees.
The series also contains a bug fix for the shared IRQ of the DSPI
driver. I am going to respin a version of it as a separate patch for
inclusion in stable trees, independent of this patchset.
Vladimir Oltean (5):
spi: spi-fsl-dspi: Reduce indentation level in dspi_interrupt
spi: spi-fsl-dspi: Exit the ISR with IRQ_NONE when it's not ours
spi: spi-fsl-dspi: Remove impossible to reach error check
spi: spi-fsl-dspi: Use poll mode in case the platform IRQ is missing
ARM: dts: ls1021a-tsn: Use the DSPI controller in poll mode
arch/arm/boot/dts/ls1021a-tsn.dts | 1 +
drivers/spi/spi-fsl-dspi.c | 128 ++++++++++++++++++------------
2 files changed, 78 insertions(+), 51 deletions(-)
--
2.17.1
^ 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