| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20080704165354.c39166ab@mailhost.serverengines.com>
Date: Fri, 04 Jul 2008 09:53:54 -0700
From: "Subbu Seetharaman" <subbus@...verengines.com>
To: jeff@...zik.org, netdev@...r.kernel.org
Subject: [PATCH 4/8] benet: hwlib functions
Signed-off-by: Subbu Seetharaman <subbus@...verengines.com>
---
drivers/net/benet/hwlib/cq.c | 211 +++++++
drivers/net/benet/hwlib/eq.c | 299 ++++++++++
drivers/net/benet/hwlib/eth.c | 1273 +++++++++++++++++++++++++++++++++++++++++
3 files changed, 1783 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/benet/hwlib/cq.c
create mode 100644 drivers/net/benet/hwlib/eq.c
create mode 100644 drivers/net/benet/hwlib/eth.c
diff --git a/drivers/net/benet/hwlib/cq.c b/drivers/net/benet/hwlib/cq.c
new file mode 100644
index 0000000..6504586
--- /dev/null
+++ b/drivers/net/benet/hwlib/cq.c
@@ -0,0 +1,211 @@
+/*
+ * Copyright (C) 2005 - 2008 ServerEngines
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation. The full GNU General
+ * Public License is included in this distribution in the file called COPYING.
+ *
+ * Contact Information:
+ * linux-drivers@...verengines.com
+ *
+ * ServerEngines
+ * 209 N. Fair Oaks Ave
+ * Sunnyvale, CA 94085
+ */
+#include "hwlib.h"
+#include "bestatus.h"
+
+/*
+ * Completion Queue Objects
+ */
+/*
+ *============================================================================
+ * P U B L I C R O U T I N E S
+ *============================================================================
+ */
+
+/*
+ This routine creates a completion queue based on the client completion
+ queue configuration information.
+
+
+ FunctionObject - Handle to a function object
+ CqBaseVa - Base VA for a the CQ ring
+ NumEntries - CEV_CQ_CNT_* values
+ solEventEnable - 0 = All CQEs can generate Events if CQ is eventable
+ 1 = only CQEs with solicited bit set are eventable
+ eventable - Eventable CQ, generates interrupts.
+ nodelay - 1 = Force interrupt, relevent if CQ eventable.
+ Interrupt is asserted immediately after EQE
+ write is confirmed, regardless of EQ Timer
+ or watermark settings.
+ wme - Enable watermark based coalescing
+ wmThresh - High watermark(CQ fullness at which event
+ or interrupt should be asserted). These are the
+ CEV_WATERMARK encoded values.
+ EqObject - EQ Handle to assign to this CQ
+ ppCqObject - Internal CQ Handle returned.
+
+ Returns BE_SUCCESS if successfull, otherwise a useful error code is
+ returned.
+
+ IRQL < DISPATCH_LEVEL
+
+*/
+int be_cq_create(struct be_function_object *pfob,
+ struct ring_desc *rd, u32 length, bool solicited_eventable,
+ bool no_delay, u32 wm_thresh,
+ struct be_eq_object *eq_object, struct be_cq_object *cq_object)
+{
+ int status = BE_SUCCESS;
+ u32 num_entries_encoding;
+ u32 num_entries = length / sizeof(struct MCC_CQ_ENTRY_AMAP);
+ struct FWCMD_COMMON_CQ_CREATE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ u32 n;
+ unsigned long irql;
+
+ ASSERT(rd);
+ ASSERT(cq_object);
+ ASSERT(length % sizeof(struct MCC_CQ_ENTRY_AMAP) == 0);
+
+ switch (num_entries) {
+ case 256:
+ num_entries_encoding = CEV_CQ_CNT_256;
+ break;
+ case 512:
+ num_entries_encoding = CEV_CQ_CNT_512;
+ break;
+ case 1024:
+ num_entries_encoding = CEV_CQ_CNT_1024;
+ break;
+ default:
+ ASSERT(0);
+ return BE_STATUS_INVALID_PARAMETER;
+ }
+
+ /*
+ * All cq entries all the same size. Use iSCSI version
+ * as a test for the proper rd length.
+ */
+ memset(cq_object, 0, sizeof(*cq_object));
+
+ atomic_set(&cq_object->ref_count, 0);
+ cq_object->parent_function = pfob;
+ cq_object->eq_object = eq_object;
+ cq_object->num_entries = num_entries;
+ /* save for MCC cq processing */
+ cq_object->va = rd->va;
+
+ /* map into UT. */
+ length = num_entries * sizeof(struct MCC_CQ_ENTRY_AMAP);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ ASSERT(wrb);
+ TRACE(DL_ERR, "No free MCC WRBs in create EQ.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_CQ_CREATE);
+
+ fwcmd->params.request.num_pages = PAGES_SPANNED(OFFSET_IN_PAGE(rd->va),
+ length);
+
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, valid, &fwcmd->params.request.context, 1);
+ n = pfob->pci_function_number;
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, Func, &fwcmd->params.request.context, n);
+
+ n = (eq_object != NULL);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, Eventable,
+ &fwcmd->params.request.context, n);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, Armed, &fwcmd->params.request.context, 1);
+
+ n = eq_object ? eq_object->eq_id : 0;
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, EQID, &fwcmd->params.request.context, n);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, Count,
+ &fwcmd->params.request.context, num_entries_encoding);
+
+ n = 0; /* Protection Domain is always 0 in Linux driver */
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, PD, &fwcmd->params.request.context, n);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, NoDelay,
+ &fwcmd->params.request.context, no_delay);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, SolEvent,
+ &fwcmd->params.request.context, solicited_eventable);
+
+ n = (wm_thresh != 0xFFFFFFFF);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, WME, &fwcmd->params.request.context, n);
+
+ n = (n ? wm_thresh : 0);
+ AMAP_SET_BITS_PTR(CQ_CONTEXT, Watermark,
+ &fwcmd->params.request.context, n);
+ /* Create a page list for the FWCMD. */
+ be_rd_to_pa_list(rd, fwcmd->params.request.pages,
+ ARRAY_SIZE(fwcmd->params.request.pages));
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+ if (status != BE_SUCCESS) {
+ TRACE(DL_ERR, "MCC to create CQ failed.");
+ goto Error;
+ }
+ /* Remember the CQ id. */
+ cq_object->cq_id = fwcmd->params.response.cq_id;
+
+ /* insert this cq into eq_object reference */
+ if (eq_object) {
+ atomic_inc(&eq_object->ref_count);
+ list_add_tail(&cq_object->cqlist_for_eq,
+ &eq_object->cq_list_head);
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+
+ Deferences the given object. Once the object's reference count drops to
+ zero, the object is destroyed and all resources that are held by this object
+ are released. The on-chip context is also destroyed along with the queue
+ ID, and any mappings made into the UT.
+
+ cq_object - CQ handle returned from cq_object_create.
+
+ returns the current reference count on the object
+
+ IRQL: IRQL < DISPATCH_LEVEL
+*/
+int be_cq_destroy(struct be_cq_object *cq_object)
+{
+ int status = 0;
+
+ /* Nothing should reference this CQ at this point. */
+ ASSERT(atomic_read(&cq_object->ref_count) == 0);
+
+ /* Send fwcmd to destroy the CQ. */
+ status = be_function_ring_destroy(cq_object->parent_function,
+ cq_object->cq_id, FWCMD_RING_TYPE_CQ,
+ NULL, NULL, NULL, NULL);
+ ASSERT(status == 0);
+
+ /* Remove reference if this is an eventable CQ. */
+ if (cq_object->eq_object) {
+ atomic_dec(&cq_object->eq_object->ref_count);
+ list_del(&cq_object->cqlist_for_eq);
+ }
+ return BE_SUCCESS;
+}
+
diff --git a/drivers/net/benet/hwlib/eq.c b/drivers/net/benet/hwlib/eq.c
new file mode 100644
index 0000000..db92ccd
--- /dev/null
+++ b/drivers/net/benet/hwlib/eq.c
@@ -0,0 +1,299 @@
+/*
+ * Copyright (C) 2005 - 2008 ServerEngines
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation. The full GNU General
+ * Public License is included in this distribution in the file called COPYING.
+ *
+ * Contact Information:
+ * linux-drivers@...verengines.com
+ *
+ * ServerEngines
+ * 209 N. Fair Oaks Ave
+ * Sunnyvale, CA 94085
+ */
+#include "hwlib.h"
+#include "bestatus.h"
+/*
+ This routine creates an event queue based on the client completion
+ queue configuration information.
+
+ FunctionObject - Handle to a function object
+ EqBaseVa - Base VA for a the EQ ring
+ SizeEncoding - The encoded size for the EQ entries. This value is
+ either CEV_EQ_SIZE_4 or CEV_EQ_SIZE_16
+ NumEntries - CEV_CQ_CNT_* values.
+ Watermark - Enables watermark based coalescing. This parameter
+ must be of the type CEV_WMARK_* if watermarks
+ are enabled. If watermarks to to be disabled
+ this value should be-1.
+ TimerDelay - If a timer delay is enabled this value should be the
+ time of the delay in 8 microsecond units. If
+ delays are not used this parameter should be
+ set to -1.
+ ppEqObject - Internal EQ Handle returned.
+
+ Returns BE_SUCCESS if successfull,, otherwise a useful error code
+ is returned.
+
+ IRQL < DISPATCH_LEVEL
+*/
+int
+be_eq_create(struct be_function_object *pfob,
+ struct ring_desc *rd, u32 eqe_size, u32 num_entries,
+ u32 watermark, /* CEV_WMARK_* or -1 */
+ u32 timer_delay, /* in 8us units, or -1 */
+ struct be_eq_object *eq_object)
+{
+ int status = BE_SUCCESS;
+ u32 num_entries_encoding, eqe_size_encoding, length;
+ struct FWCMD_COMMON_EQ_CREATE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ u32 n;
+ unsigned long irql;
+
+ ASSERT(rd);
+ ASSERT(eq_object);
+
+ switch (num_entries) {
+ case 256:
+ num_entries_encoding = CEV_EQ_CNT_256;
+ break;
+ case 512:
+ num_entries_encoding = CEV_EQ_CNT_512;
+ break;
+ case 1024:
+ num_entries_encoding = CEV_EQ_CNT_1024;
+ break;
+ case 2048:
+ num_entries_encoding = CEV_EQ_CNT_2048;
+ break;
+ case 4096:
+ num_entries_encoding = CEV_EQ_CNT_4096;
+ break;
+ default:
+ ASSERT(0);
+ return BE_STATUS_INVALID_PARAMETER;
+ }
+
+ switch (eqe_size) {
+ case 4:
+ eqe_size_encoding = CEV_EQ_SIZE_4;
+ break;
+ case 16:
+ eqe_size_encoding = CEV_EQ_SIZE_16;
+ break;
+ default:
+ ASSERT(0);
+ return BE_STATUS_INVALID_PARAMETER;
+ }
+
+ if ((eqe_size == 4 && num_entries < 1024) ||
+ (eqe_size == 16 && num_entries == 4096)) {
+ TRACE(DL_ERR, "Bad EQ size. eqe_size:%d num_entries:%d",
+ eqe_size, num_entries);
+ ASSERT(0);
+ return BE_STATUS_INVALID_PARAMETER;
+ }
+
+ memset(eq_object, 0, sizeof(*eq_object));
+
+ atomic_set(&eq_object->ref_count, 0);
+ eq_object->parent_function = pfob;
+ eq_object->eq_id = 0xFFFFFFFF;
+
+ INIT_LIST_HEAD(&eq_object->cq_list_head);
+
+ length = num_entries * eqe_size;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ ASSERT(wrb);
+ TRACE(DL_ERR, "No free MCC WRBs in create EQ.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_EQ_CREATE);
+
+ fwcmd->params.request.num_pages = PAGES_SPANNED(OFFSET_IN_PAGE(rd->va),
+ length);
+ n = pfob->pci_function_number;
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Func, &fwcmd->params.request.context, n);
+
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, valid, &fwcmd->params.request.context, 1);
+
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Size,
+ &fwcmd->params.request.context, eqe_size_encoding);
+
+ n = 0; /* Protection Domain is always 0 in Linux driver */
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, PD, &fwcmd->params.request.context, n);
+
+ /* Let the caller ARM the EQ with the doorbell. */
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Armed, &fwcmd->params.request.context, 0);
+
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Count, &fwcmd->params.request.context,
+ num_entries_encoding);
+
+ n = pfob->pci_function_number * 32;
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, EventVect,
+ &fwcmd->params.request.context, n);
+ if (watermark != -1) {
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, WME,
+ &fwcmd->params.request.context, 1);
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Watermark,
+ &fwcmd->params.request.context, watermark);
+ ASSERT(watermark <= CEV_WMARK_240);
+ } else
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, WME,
+ &fwcmd->params.request.context, 0);
+ if (timer_delay != -1) {
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, TMR,
+ &fwcmd->params.request.context, 1);
+
+ ASSERT(timer_delay <= 250); /* max value according to EAS */
+ timer_delay = min(timer_delay, (u32)250);
+
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, Delay,
+ &fwcmd->params.request.context, timer_delay);
+ } else {
+ AMAP_SET_BITS_PTR(EQ_CONTEXT, TMR,
+ &fwcmd->params.request.context, 0);
+ }
+ /* Create a page list for the FWCMD. */
+ be_rd_to_pa_list(rd, fwcmd->params.request.pages,
+ ARRAY_SIZE(fwcmd->params.request.pages));
+
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+ if (status != BE_SUCCESS) {
+ TRACE(DL_ERR, "MCC to create EQ failed.");
+ goto Error;
+ }
+ /* Get the EQ id. The MPU allocates the IDs. */
+ eq_object->eq_id = fwcmd->params.response.eq_id;
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ Deferences the given object. Once the object's reference count drops to
+ zero, the object is destroyed and all resources that are held by this
+ object are released. The on-chip context is also destroyed along with
+ the queue ID, and any mappings made into the UT.
+
+ eq_object - EQ handle returned from eq_object_create.
+
+ Returns BE_SUCCESS if successfull, otherwise a useful error code
+ is returned.
+
+ IRQL: IRQL < DISPATCH_LEVEL
+*/
+int be_eq_destroy(struct be_eq_object *eq_object)
+{
+ int status = 0;
+
+ ASSERT(atomic_read(&eq_object->ref_count) == 0);
+ /* no CQs should reference this EQ now */
+ ASSERT(list_empty(&eq_object->cq_list_head));
+
+ /* Send fwcmd to destroy the EQ. */
+ status = be_function_ring_destroy(eq_object->parent_function,
+ eq_object->eq_id, FWCMD_RING_TYPE_EQ,
+ NULL, NULL, NULL, NULL);
+ ASSERT(status == 0);
+
+ return BE_SUCCESS;
+}
+/*
+ *---------------------------------------------------------------------------
+ * Function: be_eq_modify_delay
+ * Changes the EQ delay for a group of EQs.
+ * num_eq - The number of EQs in the eq_array to adjust.
+ * This also is the number of delay values in
+ * the eq_delay_array.
+ * eq_array - Array of struct be_eq_object pointers to adjust.
+ * eq_delay_array - Array of "num_eq" timer delays in units
+ * of microseconds. The be_eq_query_delay_range
+ * fwcmd returns the resolution and range of
+ * legal EQ delays.
+ * cb -
+ * cb_context -
+ * q_ctxt - Optional. Pointer to a previously allocated
+ * struct. If the MCC WRB ring is full, this
+ * structure is used to queue the operation. It
+ * will be posted to the MCC ring when space
+ * becomes available. All queued commands will
+ * be posted to the ring in the order they are
+ * received. It is always valid to pass a pointer to
+ * a generic be_generic_q_cntxt. However,
+ * the specific context structs
+ * are generally smaller than the generic struct.
+ * return pend_status - BE_SUCCESS (0) on success.
+ * BE_PENDING (postive value) if the FWCMD
+ * completion is pending. Negative error code on failure.
+ *-------------------------------------------------------------------------
+ */
+int
+be_eq_modify_delay(struct be_function_object *pfob,
+ u32 num_eq, struct be_eq_object **eq_array,
+ u32 *eq_delay_array, mcc_wrb_cqe_callback cb,
+ void *cb_context, struct be_eq_modify_delay_q_ctxt *q_ctxt)
+{
+ struct FWCMD_COMMON_MODIFY_EQ_DELAY *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ struct be_generic_q_ctxt *gen_ctxt = NULL;
+ u32 i;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ gen_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ gen_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_MODIFY_EQ_DELAY);
+
+ ASSERT(num_eq > 0);
+ ASSERT(num_eq <= ARRAY_SIZE(fwcmd->params.request.delay));
+ fwcmd->params.request.num_eq = num_eq;
+ for (i = 0; i < num_eq; i++) {
+ fwcmd->params.request.delay[i].eq_id = eq_array[i]->eq_id;
+ fwcmd->params.request.delay[i].delay_in_microseconds =
+ eq_delay_array[i];
+ }
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, gen_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, NULL);
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
diff --git a/drivers/net/benet/hwlib/eth.c b/drivers/net/benet/hwlib/eth.c
new file mode 100644
index 0000000..fced8a8
--- /dev/null
+++ b/drivers/net/benet/hwlib/eth.c
@@ -0,0 +1,1273 @@
+/*
+ * Copyright (C) 2005 - 2008 ServerEngines
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation. The full GNU General
+ * Public License is included in this distribution in the file called COPYING.
+ *
+ * Contact Information:
+ * linux-drivers@...verengines.com
+ *
+ * ServerEngines
+ * 209 N. Fair Oaks Ave
+ * Sunnyvale, CA 94085
+ */
+#include <linux/if_ether.h>
+#include "hwlib.h"
+#include "bestatus.h"
+
+/*
+ *---------------------------------------------------------
+ * Function: be_eth_sq_create_ex
+ * Creates an ethernet send ring - extended version with
+ * additional parameters.
+ * pfob -
+ * rd - ring address
+ * length_in_bytes -
+ * type - The type of ring to create.
+ * ulp - The requested ULP number for the ring.
+ * This should be zero based, i.e. 0,1,2. This must
+ * be valid NIC ULP based on the firmware config.
+ * All doorbells for this ring must be sent to
+ * this ULP. The first network ring allocated for
+ * each ULP are higher performance than subsequent rings.
+ * cq_object - cq object for completions
+ * ex_parameters - Additional parameters (that may increase in
+ * future revisions). These parameters are only used
+ * for certain ring types -- see
+ * struct be_eth_sq_parameters for details.
+ * eth_sq -
+ * return status - BE_SUCCESS (0) on success. Negative error code on failure.
+ *---------------------------------------------------------
+ */
+int
+be_eth_sq_create_ex(struct be_function_object *pfob, struct ring_desc *rd,
+ u32 length, u32 type, u32 ulp, struct be_cq_object *cq_object,
+ struct be_eth_sq_parameters *ex_parameters,
+ struct be_ethsq_object *eth_sq)
+{
+ struct FWCMD_COMMON_ETH_TX_CREATE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ u32 n;
+ unsigned long irql;
+
+ ASSERT(rd);
+ ASSERT(eth_sq);
+ ASSERT(ex_parameters);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ memset(eth_sq, 0, sizeof(*eth_sq));
+
+ eth_sq->parent_function = pfob;
+ eth_sq->bid = 0xFFFFFFFF;
+ eth_sq->cq_object = cq_object;
+
+ /* Translate hwlib interface to arm interface. */
+ switch (type) {
+ case BE_ETH_TX_RING_TYPE_FORWARDING:
+ type = ETH_TX_RING_TYPE_FORWARDING;
+ break;
+ case BE_ETH_TX_RING_TYPE_STANDARD:
+ type = ETH_TX_RING_TYPE_STANDARD;
+ break;
+ case BE_ETH_TX_RING_TYPE_BOUND:
+ ASSERT(ex_parameters->port < 2);
+ type = ETH_TX_RING_TYPE_BOUND;
+ break;
+ default:
+ TRACE(DL_ERR, "Invalid eth tx ring type:%d", type);
+ return BE_NOT_OK;
+ break;
+ }
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ ASSERT(wrb);
+ TRACE(DL_ERR, "No free MCC WRBs in create EQ.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ /* NIC must be supported by the current config. */
+ ASSERT(pfob->fw_config.nic_ulp_mask);
+
+ /*
+ * The ulp parameter must select a valid NIC ULP
+ * for the current config.
+ */
+ ASSERT((1 << ulp) & pfob->fw_config.nic_ulp_mask);
+
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_ETH_TX_CREATE);
+ fwcmd->header.request.port_number = ex_parameters->port;
+
+ AMAP_SET_BITS_PTR(ETX_CONTEXT, pd_id,
+ &fwcmd->params.request.context, 0);
+
+ n = be_ring_length_to_encoding(length, sizeof(struct ETH_WRB_AMAP));
+ AMAP_SET_BITS_PTR(ETX_CONTEXT, tx_ring_size,
+ &fwcmd->params.request.context, n);
+
+ AMAP_SET_BITS_PTR(ETX_CONTEXT, cq_id_send,
+ &fwcmd->params.request.context, cq_object->cq_id);
+
+ n = pfob->pci_function_number;
+ AMAP_SET_BITS_PTR(ETX_CONTEXT, func, &fwcmd->params.request.context, n);
+
+ fwcmd->params.request.type = type;
+ fwcmd->params.request.ulp_num = (1 << ulp);
+ fwcmd->params.request.num_pages = DIV_ROUND_UP(length, PAGE_SIZE);
+ ASSERT(PAGES_SPANNED(rd->va, rd->length) >=
+ fwcmd->params.request.num_pages);
+
+ /* Create a page list for the FWCMD. */
+ be_rd_to_pa_list(rd, fwcmd->params.request.pages,
+ ARRAY_SIZE(fwcmd->params.request.pages));
+
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+ if (status != BE_SUCCESS) {
+ TRACE(DL_ERR, "MCC to create etx queue failed.");
+ goto Error;
+ }
+ /* save the butler ID */
+ eth_sq->bid = fwcmd->params.response.cid;
+
+ /* add a reference to the corresponding CQ */
+ atomic_inc(&cq_object->ref_count);
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+
+/*
+ This routine destroys an ethernet send queue
+
+ EthSq - EthSq Handle returned from EthSqCreate
+
+ This function always return BE_SUCCESS.
+
+ This function frees memory allocated by EthSqCreate for the EthSq Object.
+
+*/
+int be_eth_sq_destroy(struct be_ethsq_object *eth_sq)
+{
+ int status = 0;
+
+ /* Send fwcmd to destroy the queue. */
+ status = be_function_ring_destroy(eth_sq->parent_function, eth_sq->bid,
+ FWCMD_RING_TYPE_ETH_TX, NULL, NULL, NULL, NULL);
+ ASSERT(status == 0);
+
+ /* Derefence any associated CQs. */
+ atomic_dec(ð_sq->cq_object->ref_count);
+ return status;
+}
+/*
+ This routine attempts to set the transmit flow control parameters.
+
+ FunctionObject - Handle to a function object
+
+ txfc_enable - transmit flow control enable - true for
+ enable, false for disable
+
+ rxfc_enable - receive flow control enable - true for
+ enable, false for disable
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int error
+ code is returned.
+
+ IRQL: < DISPATCH_LEVEL
+
+ This function always fails in non-privileged machine context.
+*/
+int
+be_eth_set_flow_control(struct be_function_object *pfob,
+ bool txfc_enable, bool rxfc_enable)
+{
+ struct FWCMD_COMMON_SET_FLOW_CONTROL *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_SET_FLOW_CONTROL);
+
+ fwcmd->params.request.rx_flow_control = rxfc_enable;
+ fwcmd->params.request.tx_flow_control = txfc_enable;
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "set flow control fwcmd failed.");
+ goto error;
+ }
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine attempts to get the transmit flow control parameters.
+
+ pfob - Handle to a function object
+
+ txfc_enable - transmit flow control enable - true for
+ enable, false for disable
+
+ rxfc_enable - receive flow control enable - true for enable,
+ false for disable
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int error code
+ is returned.
+
+ IRQL: < DISPATCH_LEVEL
+
+ This function always fails in non-privileged machine context.
+*/
+int
+be_eth_get_flow_control(struct be_function_object *pfob,
+ bool *txfc_enable, bool *rxfc_enable)
+{
+ struct FWCMD_COMMON_GET_FLOW_CONTROL *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_GET_FLOW_CONTROL);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "get flow control fwcmd failed.");
+ goto error;
+ }
+
+ *txfc_enable = fwcmd->params.response.tx_flow_control;
+ *rxfc_enable = fwcmd->params.response.rx_flow_control;
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ *---------------------------------------------------------
+ * Function: be_eth_set_qos
+ * This function sets the ethernet transmit Quality of Service (QoS)
+ * characteristics of BladeEngine for the domain. All ethernet
+ * transmit rings of the domain will evenly share the bandwidth.
+ * The exeception to sharing is the host primary (super) ethernet
+ * transmit ring as well as the host ethernet forwarding ring
+ * for missed offload data.
+ * pfob -
+ * max_bps - the maximum bits per second in units of
+ * 10 Mbps (valid 0-100)
+ * max_pps - the maximum packets per second in units
+ * of 1 Kpps (0 indicates no limit)
+ * return status - BE_SUCCESS (0) on success. Negative error code on failure.
+ *---------------------------------------------------------
+ */
+int
+be_eth_set_qos(struct be_function_object *pfob, u32 max_bps, u32 max_pps)
+{
+ struct FWCMD_COMMON_SET_QOS *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_SET_QOS);
+
+ /* Set fields in fwcmd */
+ fwcmd->params.request.max_bits_per_second_NIC = max_bps;
+ fwcmd->params.request.max_packets_per_second_NIC = max_pps;
+ fwcmd->params.request.valid_flags = QOS_BITS_NIC | QOS_PKTS_NIC;
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0)
+ TRACE(DL_ERR, "network set qos fwcmd failed.");
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ *---------------------------------------------------------
+ * Function: be_eth_get_qos
+ * This function retrieves the ethernet transmit Quality of Service (QoS)
+ * characteristics for the domain.
+ * max_bps - the maximum bits per second in units of
+ * 10 Mbps (valid 0-100)
+ * max_pps - the maximum packets per second in units of
+ * 1 Kpps (0 indicates no limit)
+ * return status - BE_SUCCESS (0) on success. Negative error code on failure.
+ *---------------------------------------------------------
+ */
+int
+be_eth_get_qos(struct be_function_object *pfob, u32 *max_bps, u32 *max_pps)
+{
+ struct FWCMD_COMMON_GET_QOS *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_GET_QOS);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "network get qos fwcmd failed.");
+ goto error;
+ }
+
+ *max_bps = fwcmd->params.response.max_bits_per_second_NIC;
+ *max_pps = fwcmd->params.response.max_packets_per_second_NIC;
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ *---------------------------------------------------------
+ * Function: be_eth_set_frame_size
+ * This function sets the ethernet maximum frame size. The previous
+ * values are returned.
+ * pfob -
+ * tx_frame_size - maximum transmit frame size in bytes
+ * rx_frame_size - maximum receive frame size in bytes
+ * return status - BE_SUCCESS (0) on success. Negative error code on failure.
+ *---------------------------------------------------------
+ */
+int
+be_eth_set_frame_size(struct be_function_object *pfob,
+ u32 *tx_frame_size, u32 *rx_frame_size)
+{
+ struct FWCMD_COMMON_SET_FRAME_SIZE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_SET_FRAME_SIZE);
+ fwcmd->params.request.max_tx_frame_size = *tx_frame_size;
+ fwcmd->params.request.max_rx_frame_size = *rx_frame_size;
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "network set frame size fwcmd failed.");
+ goto error;
+ }
+
+ *tx_frame_size = fwcmd->params.response.chip_max_tx_frame_size;
+ *rx_frame_size = fwcmd->params.response.chip_max_rx_frame_size;
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+
+/*
+ This routine creates a Ethernet receive ring.
+
+ pfob - handle to a function object
+ rq_base_va - base VA for the default receive ring. this must be
+ exactly 8K in length and continguous physical memory.
+ cq_object - handle to a previously created CQ to be associated
+ with the RQ.
+ pp_eth_rq - pointer to an opqaue handle where an eth
+ receive object is returned.
+ Returns BE_SUCCESS if successfull, , otherwise a useful
+ int error code is returned.
+
+ IRQL: < DISPATCH_LEVEL
+ this function allocates a struct be_ethrq_object *object.
+ there must be no more than 1 of these per function object, unless the
+ function object supports RSS (is networking and on the host).
+ the rq_base_va must point to a buffer of exactly 8K.
+ the erx::host_cqid (or host_stor_cqid) register and erx::ring_page registers
+ will be updated as appropriate on return
+*/
+int
+be_eth_rq_create(struct be_function_object *pfob,
+ struct ring_desc *rd, struct be_cq_object *cq_object,
+ struct be_cq_object *bcmc_cq_object,
+ struct be_ethrq_object *eth_rq)
+{
+ int status = 0;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ struct FWCMD_COMMON_ETH_RX_CREATE *fwcmd = NULL;
+ unsigned long irql;
+
+ /* MPU will set the */
+ ASSERT(rd);
+ ASSERT(eth_rq);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ eth_rq->parent_function = pfob;
+ eth_rq->cq_object = cq_object;
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_ETH_RX_CREATE);
+
+ fwcmd->params.request.num_pages = 2; /* required length */
+ fwcmd->params.request.cq_id = cq_object->cq_id;
+
+ if (bcmc_cq_object)
+ fwcmd->params.request.bcmc_cq_id = bcmc_cq_object->cq_id;
+ else
+ fwcmd->params.request.bcmc_cq_id = 0xFFFF;
+
+ /* Create a page list for the FWCMD. */
+ be_rd_to_pa_list(rd, fwcmd->params.request.pages,
+ ARRAY_SIZE(fwcmd->params.request.pages));
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+ if (status != BE_SUCCESS) {
+ TRACE(DL_ERR, "fwcmd to map eth rxq frags failed.");
+ goto Error;
+ }
+ /* Save the ring ID for cleanup. */
+ eth_rq->rid = fwcmd->params.response.id;
+
+ atomic_inc(&cq_object->ref_count);
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine destroys an Ethernet receive queue
+
+ eth_rq - ethernet receive queue handle returned from eth_rq_create
+
+ Returns BE_SUCCESS on success and an appropriate int on failure.
+
+ This function frees resourcs allocated by EthRqCreate.
+ The erx::host_cqid (or host_stor_cqid) register and erx::ring_page
+ registers will be updated as appropriate on return
+ IRQL: < DISPATCH_LEVEL
+*/
+
+static void be_eth_rq_destroy_internal_cb(void *context, int status,
+ struct MCC_WRB_AMAP *wrb)
+{
+ struct be_ethrq_object *eth_rq = (struct be_ethrq_object *) context;
+
+ if (status != BE_SUCCESS) {
+ TRACE(DL_ERR, "Destroy eth rq failed in internal callback.\n");
+ } else {
+ /* Dereference any CQs associated with this queue. */
+ atomic_dec(ð_rq->cq_object->ref_count);
+ }
+
+ return;
+}
+
+int be_eth_rq_destroy(struct be_ethrq_object *eth_rq)
+{
+ int status = BE_SUCCESS;
+
+ /* Send fwcmd to destroy the RQ. */
+ status = be_function_ring_destroy(eth_rq->parent_function,
+ eth_rq->rid, FWCMD_RING_TYPE_ETH_RX, NULL, NULL,
+ be_eth_rq_destroy_internal_cb, eth_rq);
+
+ return status;
+}
+
+/*
+ *---------------------------------------------------------------------------
+ * Function: be_eth_rq_destroy_options
+ * Destroys an ethernet receive ring with finer granularity options
+ * than the standard be_eth_rq_destroy() API function.
+ * eth_rq -
+ * flush - Set to 1 to flush the ring, set to 0 to bypass the flush
+ * cb - Callback function on completion
+ * cb_context - Callback context
+ * return status - BE_SUCCESS (0) on success. Negative error code on failure.
+ *----------------------------------------------------------------------------
+ */
+int
+be_eth_rq_destroy_options(struct be_ethrq_object *eth_rq, bool flush,
+ mcc_wrb_cqe_callback cb, void *cb_context)
+{
+ struct FWCMD_COMMON_RING_DESTROY *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = BE_SUCCESS;
+ struct be_function_object *pfob = NULL;
+ unsigned long irql;
+
+ pfob = eth_rq->parent_function;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ TRACE(DL_INFO, "Destroy eth_rq ring id:%d, flush:%d", eth_rq->rid,
+ flush);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ ASSERT(wrb);
+ TRACE(DL_ERR, "No free MCC WRBs in destroy eth_rq ring.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_RING_DESTROY);
+
+ fwcmd->params.request.id = eth_rq->rid;
+ fwcmd->params.request.ring_type = FWCMD_RING_TYPE_ETH_RX;
+ fwcmd->params.request.bypass_flush = ((0 == flush) ? 1 : 0);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, cb, cb_context,
+ be_eth_rq_destroy_internal_cb, eth_rq, fwcmd, NULL);
+
+ if (status != BE_SUCCESS && status != BE_PENDING) {
+ TRACE(DL_ERR, "eth_rq ring destroy failed. id:%d, flush:%d",
+ eth_rq->rid, flush);
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine queries the frag size for erx.
+
+ pfob - handle to a function object
+
+ frag_size_bytes - erx frag size in bytes that is/was set.
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int error
+ code is returned.
+
+ IRQL: < DISPATCH_LEVEL
+
+*/
+int
+be_eth_rq_get_frag_size(struct be_function_object *pfob, u32 *frag_size_bytes)
+{
+ struct FWCMD_ETH_GET_RX_FRAG_SIZE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ ASSERT(frag_size_bytes);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ return BE_STATUS_NO_MCC_WRB;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, ETH_GET_RX_FRAG_SIZE);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "get frag size fwcmd failed.");
+ goto error;
+ }
+
+ *frag_size_bytes = 1 << fwcmd->params.response.actual_fragsize_log2;
+
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine attempts to set the frag size for erx. If the frag size is
+ already set, the attempt fails and the current frag size is returned.
+
+ pfob - Handle to a function object
+
+ frag_size - Erx frag size in bytes that is/was set.
+
+ current_frag_size_bytes - Pointer to location where currrent frag
+ is to be rturned
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int error
+ code is returned.
+
+ IRQL: < DISPATCH_LEVEL
+
+ This function always fails in non-privileged machine context.
+*/
+int
+be_eth_rq_set_frag_size(struct be_function_object *pfob,
+ u32 frag_size, u32 *frag_size_bytes)
+{
+ struct FWCMD_ETH_SET_RX_FRAG_SIZE *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ unsigned long irql;
+
+ ASSERT(frag_size_bytes);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto error;
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, ETH_SET_RX_FRAG_SIZE);
+
+ ASSERT(frag_size >= 128 && frag_size <= 16 * 1024);
+
+ /* This is the log2 of the fragsize. This is not the exact
+ * ERX encoding. */
+ fwcmd->params.request.new_fragsize_log2 = __ilog2_u32(frag_size);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, NULL, NULL,
+ NULL, NULL, fwcmd, NULL);
+
+ if (status != 0) {
+ TRACE(DL_ERR, "set frag size fwcmd failed.");
+ goto error;
+ }
+
+ *frag_size_bytes = 1 << fwcmd->params.response.actual_fragsize_log2;
+error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+
+/*
+ This routine gets or sets a mac address for a domain
+ given the port and mac.
+
+ FunctionObject - Function object handle.
+ port1 - Set to TRUE if this function will set/get the Port 1
+ address. Only the host may set this to TRUE.
+ mac1 - Set to TRUE if this function will set/get the
+ MAC 1 address. Only the host may set this to TRUE.
+ write - Set to TRUE if this function should write the mac address.
+ mac_address - Buffer of the mac address to read or write.
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int is returned.
+
+ IRQL: < DISPATCH_LEVEL
+*/
+int be_rxf_mac_address_read_write(struct be_function_object *pfob,
+ bool port1, /* VM must always set to false */
+ bool mac1, /* VM must always set to false */
+ bool mgmt, bool write,
+ bool permanent, u8 *mac_address,
+ mcc_wrb_cqe_callback cb, /* optional */
+ void *cb_context) /* optional */
+{
+ int status = BE_SUCCESS;
+ union {
+ struct FWCMD_COMMON_NTWK_MAC_QUERY *query;
+ struct FWCMD_COMMON_NTWK_MAC_SET *set;
+ } fwcmd = {NULL};
+ struct MCC_WRB_AMAP *wrb = NULL;
+ u32 type = 0;
+ unsigned long irql;
+ struct be_mcc_wrb_response_copy rc;
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ ASSERT(mac_address);
+
+ ASSERT(port1 == false);
+ ASSERT(mac1 == false);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ TRACE(DL_ERR, "MCC wrb peek failed.");
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+
+ if (mgmt) {
+ type = MAC_ADDRESS_TYPE_MANAGEMENT;
+ } else {
+ if (pfob->type == BE_FUNCTION_TYPE_NETWORK)
+ type = MAC_ADDRESS_TYPE_NETWORK;
+ else
+ type = MAC_ADDRESS_TYPE_STORAGE;
+ }
+
+ if (write) {
+ /* Prepares an embedded fwcmd, including
+ * request/response sizes.
+ */
+ fwcmd.set = BE_PREPARE_EMBEDDED_FWCMD(pfob,
+ wrb, COMMON_NTWK_MAC_SET);
+
+ fwcmd.set->params.request.invalidate = 0;
+ fwcmd.set->params.request.mac1 = (mac1 ? 1 : 0);
+ fwcmd.set->params.request.port = (port1 ? 1 : 0);
+ fwcmd.set->params.request.type = type;
+
+ /* Copy the mac address to set. */
+ fwcmd.set->params.request.mac.SizeOfStructure =
+ sizeof(fwcmd.set->params.request.mac);
+ memcpy(fwcmd.set->params.request.mac.MACAddress,
+ mac_address, ETH_ALEN);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL,
+ cb, cb_context, NULL, NULL, fwcmd.set, NULL);
+
+ } else {
+
+ /*
+ * Prepares an embedded fwcmd, including
+ * request/response sizes.
+ */
+ fwcmd.query = BE_PREPARE_EMBEDDED_FWCMD(pfob,
+ wrb, COMMON_NTWK_MAC_QUERY);
+
+ fwcmd.query->params.request.mac1 = (mac1 ? 1 : 0);
+ fwcmd.query->params.request.port = (port1 ? 1 : 0);
+ fwcmd.query->params.request.type = type;
+ fwcmd.query->params.request.permanent = permanent;
+
+ rc.length = FIELD_SIZEOF(struct FWCMD_COMMON_NTWK_MAC_QUERY,
+ params.response.mac.MACAddress);
+ rc.fwcmd_offset = offsetof(struct FWCMD_COMMON_NTWK_MAC_QUERY,
+ params.response.mac.MACAddress);
+ rc.va = mac_address;
+ /* Post the f/w command (with a copy for the response) */
+ status = be_function_post_mcc_wrb(pfob, wrb, NULL, cb,
+ cb_context, NULL, NULL, fwcmd.query, &rc);
+ }
+
+ if (status < 0) {
+ TRACE(DL_ERR, "mac set/query failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine writes data to context memory.
+
+ pfob - Function object handle.
+ mac_table - Set to the 128-bit multicast address hash table.
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int is returned.
+
+ IRQL: < DISPATCH_LEVEL
+*/
+
+int be_rxf_multicast_config(struct be_function_object *pfob,
+ bool promiscuous, u32 num, u8 *mac_table,
+ mcc_wrb_cqe_callback cb, /* optional */
+ void *cb_context,
+ struct be_multicast_q_ctxt *q_ctxt)
+{
+ int status = BE_SUCCESS;
+ struct FWCMD_COMMON_NTWK_MULTICAST_SET *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+
+ ASSERT(num <= ARRAY_SIZE(fwcmd->params.request.mac));
+
+ if (num > ARRAY_SIZE(fwcmd->params.request.mac)) {
+ TRACE(DL_ERR, "Too many multicast addresses. BE supports %d.",
+ ARRAY_SIZE(fwcmd->params.request.mac));
+ return BE_NOT_OK;
+ }
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_NTWK_MULTICAST_SET);
+
+ fwcmd->params.request.promiscuous = promiscuous;
+ if (!promiscuous) {
+ fwcmd->params.request.num_mac = num;
+ if (num > 0) {
+ ASSERT(mac_table);
+ memcpy(fwcmd->params.request.mac,
+ mac_table, ETH_ALEN * num);
+ }
+ }
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, NULL);
+ if (status < 0) {
+ TRACE(DL_ERR, "multicast fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+/*
+ This routine adds or removes a vlan tag from the rxf table.
+
+ FunctionObject - Function object handle.
+ VLanTag - VLan tag to add or remove.
+ Add - Set to TRUE if this will add a vlan tag
+
+ Returns BE_SUCCESS if successfull, otherwise a useful int is returned.
+
+ IRQL: < DISPATCH_LEVEL
+*/
+int be_rxf_vlan_config(struct be_function_object *pfob,
+ bool promiscuous, u32 num, u16 *vlan_tag_array,
+ mcc_wrb_cqe_callback cb, /* optional */
+ void *cb_context,
+ struct be_vlan_q_ctxt *q_ctxt) /* optional */
+{
+ int status = BE_SUCCESS;
+ struct FWCMD_COMMON_NTWK_VLAN_CONFIG *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+
+ if (num > ARRAY_SIZE(fwcmd->params.request.vlan_tag)) {
+ TRACE(DL_ERR, "Too many VLAN tags.");
+ return BE_NOT_OK;
+ }
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_NTWK_VLAN_CONFIG);
+
+ fwcmd->params.request.promiscuous = promiscuous;
+ if (!promiscuous) {
+ fwcmd->params.request.num_vlan = num;
+
+ if (num > 0) {
+ ASSERT(vlan_tag_array);
+ memcpy(fwcmd->params.request.vlan_tag, vlan_tag_array,
+ num * sizeof(vlan_tag_array[0]));
+ }
+ }
+
+ /* Post the commadn */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, NULL);
+ if (status < 0) {
+ TRACE(DL_ERR, "vlan fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+
+int be_rxf_link_status(struct be_function_object *pfob,
+ struct BE_LINK_STATUS *link_status,
+ mcc_wrb_cqe_callback cb,
+ void *cb_context,
+ struct be_link_status_q_ctxt *q_ctxt)
+{
+ struct FWCMD_COMMON_NTWK_LINK_STATUS_QUERY *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+ struct be_mcc_wrb_response_copy rc;
+
+ ASSERT(link_status);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb,
+ COMMON_NTWK_LINK_STATUS_QUERY);
+
+ rc.length = FIELD_SIZEOF(struct FWCMD_COMMON_NTWK_LINK_STATUS_QUERY,
+ params.response);
+ rc.fwcmd_offset = offsetof(struct FWCMD_COMMON_NTWK_LINK_STATUS_QUERY,
+ params.response);
+ rc.va = link_status;
+ /* Post or queue the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, &rc);
+
+ if (status < 0) {
+ TRACE(DL_ERR, "link status fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+int
+be_rxf_query_eth_statistics(struct be_function_object *pfob,
+ struct FWCMD_ETH_GET_STATISTICS *va_for_fwcmd,
+ u64 pa_for_fwcmd, mcc_wrb_cqe_callback cb,
+ void *cb_context,
+ struct be_nonembedded_q_ctxt *q_ctxt)
+{
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+
+ ASSERT(va_for_fwcmd);
+ ASSERT(pa_for_fwcmd);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+
+ TRACE(DL_INFO, "Query eth stats. fwcmd va:%p pa:0x%08x_%08x",
+ va_for_fwcmd, upper_32_bits(pa_for_fwcmd), (u32)pa_for_fwcmd);
+
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ va_for_fwcmd = BE_PREPARE_NONEMBEDDED_FWCMD(pfob, wrb,
+ va_for_fwcmd, pa_for_fwcmd, ETH_GET_STATISTICS);
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, va_for_fwcmd, NULL);
+ if (status < 0) {
+ TRACE(DL_ERR, "eth stats fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+int
+be_rxf_promiscuous(struct be_function_object *pfob,
+ bool enable_port0, bool enable_port1,
+ mcc_wrb_cqe_callback cb, void *cb_context,
+ struct be_promiscuous_q_ctxt *q_ctxt)
+{
+ struct FWCMD_ETH_PROMISCUOUS *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, ETH_PROMISCUOUS);
+
+ fwcmd->params.request.port0_promiscuous = enable_port0;
+ fwcmd->params.request.port1_promiscuous = enable_port1;
+
+ /* Post the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, NULL);
+
+ if (status < 0) {
+ TRACE(DL_ERR, "promiscuous fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
+
+
+/*
+ *-------------------------------------------------------------------------
+ * Function: be_rxf_filter_config
+ * Configures BladeEngine ethernet receive filter settings.
+ * pfob -
+ * settings - Pointer to the requested filter settings.
+ * The response from BladeEngine will be placed back
+ * in this structure.
+ * cb - optional
+ * cb_context - optional
+ * q_ctxt - Optional. Pointer to a previously allocated struct.
+ * If the MCC WRB ring is full, this structure is
+ * used to queue the operation. It will be posted
+ * to the MCC ring when space becomes available. All
+ * queued commands will be posted to the ring in
+ * the order they are received. It is always valid
+ * to pass a pointer to a generic
+ * be_generic_q_ctxt. However, the specific
+ * context structs are generally smaller than
+ * the generic struct.
+ * return pend_status - BE_SUCCESS (0) on success.
+ * BE_PENDING (postive value) if the FWCMD
+ * completion is pending. Negative error code on failure.
+ *---------------------------------------------------------------------------
+ */
+int
+be_rxf_filter_config(struct be_function_object *pfob,
+ struct NTWK_RX_FILTER_SETTINGS *settings,
+ mcc_wrb_cqe_callback cb, void *cb_context,
+ struct be_rxf_filter_q_ctxt *q_ctxt)
+{
+ struct FWCMD_COMMON_NTWK_RX_FILTER *fwcmd = NULL;
+ struct MCC_WRB_AMAP *wrb = NULL;
+ int status = 0;
+ struct be_generic_q_ctxt *generic_ctxt = NULL;
+ unsigned long irql;
+ struct be_mcc_wrb_response_copy rc;
+
+ ASSERT(settings);
+
+ spin_lock_irqsave(&pfob->post_lock, irql);
+
+ wrb = be_function_peek_mcc_wrb(pfob);
+
+ if (!wrb) {
+ if (q_ctxt && cb) {
+ wrb = (struct MCC_WRB_AMAP *) &q_ctxt->wrb_header;
+ generic_ctxt = (struct be_generic_q_ctxt *) q_ctxt;
+ generic_ctxt->context.bytes = sizeof(*q_ctxt);
+ } else {
+ status = BE_STATUS_NO_MCC_WRB;
+ goto Error;
+ }
+ }
+ /* Prepares an embedded fwcmd, including request/response sizes. */
+ fwcmd = BE_PREPARE_EMBEDDED_FWCMD(pfob, wrb, COMMON_NTWK_RX_FILTER);
+ memcpy(&fwcmd->params.request, settings, sizeof(*settings));
+
+ rc.length = FIELD_SIZEOF(struct FWCMD_COMMON_NTWK_RX_FILTER,
+ params.response);
+ rc.fwcmd_offset = offsetof(struct FWCMD_COMMON_NTWK_RX_FILTER,
+ params.response);
+ rc.va = settings;
+ /* Post or queue the f/w command */
+ status = be_function_post_mcc_wrb(pfob, wrb, generic_ctxt,
+ cb, cb_context, NULL, NULL, fwcmd, &rc);
+
+ if (status < 0) {
+ TRACE(DL_ERR, "RXF/ERX filter config fwcmd failed.");
+ goto Error;
+ }
+
+Error:
+ spin_unlock_irqrestore(&pfob->post_lock, irql);
+ if (pfob->pend_queue_driving && pfob->mcc) {
+ pfob->pend_queue_driving = 0;
+ be_drive_mcc_wrb_queue(pfob->mcc);
+ }
+ return status;
+}
--
1.5.5
___________________________________________________________________________________
This message, together with any attachment(s), contains confidential and proprietary information of
ServerEngines Corporation and is intended only for the designated recipient(s) named above. Any unauthorized
review, printing, retention, copying, disclosure or distribution is strictly prohibited. If you are not the
intended recipient of this message, please immediately advise the sender by reply email message and
delete all copies of this message and any attachment(s). Thank you.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists