Haim Dreyfuss 47fe2f8ed5 iwlwifi: mvm Support new MCC update response
Change MCC update response API to be compatible with new FW API.
While at it change v2 which is not in use anymore to v3 and cleanup
mcc_update v1 command and response which is obsolete.

Signed-off-by: Haim Dreyfuss <haim.dreyfuss@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
2018-10-06 10:25:51 +03:00

641 lines
18 KiB
C

/******************************************************************************
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
*
* GPL LICENSE SUMMARY
*
* Copyright(c) 2012 - 2014 Intel Corporation. All rights reserved.
* Copyright(c) 2013 - 2015 Intel Mobile Communications GmbH
* Copyright(c) 2016 - 2017 Intel Deutschland GmbH
* Copyright(c) 2018 Intel Corporation
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of version 2 of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* The full GNU General Public License is included in this distribution
* in the file called COPYING.
*
* Contact Information:
* Intel Linux Wireless <linuxwifi@intel.com>
* Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
*
* BSD LICENSE
*
* Copyright(c) 2012 - 2014 Intel Corporation. All rights reserved.
* Copyright(c) 2013 - 2015 Intel Mobile Communications GmbH
* Copyright(c) 2016 - 2017 Intel Deutschland GmbH
* Copyright(c) 2018 Intel Corporation
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* * Neither the name Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
*****************************************************************************/
#include <linux/firmware.h>
#include <linux/rtnetlink.h>
#include "iwl-trans.h"
#include "iwl-csr.h"
#include "mvm.h"
#include "iwl-eeprom-parse.h"
#include "iwl-eeprom-read.h"
#include "iwl-nvm-parse.h"
#include "iwl-prph.h"
#include "fw/acpi.h"
/* Default NVM size to read */
#define IWL_NVM_DEFAULT_CHUNK_SIZE (2 * 1024)
#define NVM_WRITE_OPCODE 1
#define NVM_READ_OPCODE 0
/* load nvm chunk response */
enum {
READ_NVM_CHUNK_SUCCEED = 0,
READ_NVM_CHUNK_NOT_VALID_ADDRESS = 1
};
/*
* prepare the NVM host command w/ the pointers to the nvm buffer
* and send it to fw
*/
static int iwl_nvm_write_chunk(struct iwl_mvm *mvm, u16 section,
u16 offset, u16 length, const u8 *data)
{
struct iwl_nvm_access_cmd nvm_access_cmd = {
.offset = cpu_to_le16(offset),
.length = cpu_to_le16(length),
.type = cpu_to_le16(section),
.op_code = NVM_WRITE_OPCODE,
};
struct iwl_host_cmd cmd = {
.id = NVM_ACCESS_CMD,
.len = { sizeof(struct iwl_nvm_access_cmd), length },
.flags = CMD_WANT_SKB | CMD_SEND_IN_RFKILL,
.data = { &nvm_access_cmd, data },
/* data may come from vmalloc, so use _DUP */
.dataflags = { 0, IWL_HCMD_DFL_DUP },
};
struct iwl_rx_packet *pkt;
struct iwl_nvm_access_resp *nvm_resp;
int ret;
ret = iwl_mvm_send_cmd(mvm, &cmd);
if (ret)
return ret;
pkt = cmd.resp_pkt;
/* Extract & check NVM write response */
nvm_resp = (void *)pkt->data;
if (le16_to_cpu(nvm_resp->status) != READ_NVM_CHUNK_SUCCEED) {
IWL_ERR(mvm,
"NVM access write command failed for section %u (status = 0x%x)\n",
section, le16_to_cpu(nvm_resp->status));
ret = -EIO;
}
iwl_free_resp(&cmd);
return ret;
}
static int iwl_nvm_read_chunk(struct iwl_mvm *mvm, u16 section,
u16 offset, u16 length, u8 *data)
{
struct iwl_nvm_access_cmd nvm_access_cmd = {
.offset = cpu_to_le16(offset),
.length = cpu_to_le16(length),
.type = cpu_to_le16(section),
.op_code = NVM_READ_OPCODE,
};
struct iwl_nvm_access_resp *nvm_resp;
struct iwl_rx_packet *pkt;
struct iwl_host_cmd cmd = {
.id = NVM_ACCESS_CMD,
.flags = CMD_WANT_SKB | CMD_SEND_IN_RFKILL,
.data = { &nvm_access_cmd, },
};
int ret, bytes_read, offset_read;
u8 *resp_data;
cmd.len[0] = sizeof(struct iwl_nvm_access_cmd);
ret = iwl_mvm_send_cmd(mvm, &cmd);
if (ret)
return ret;
pkt = cmd.resp_pkt;
/* Extract NVM response */
nvm_resp = (void *)pkt->data;
ret = le16_to_cpu(nvm_resp->status);
bytes_read = le16_to_cpu(nvm_resp->length);
offset_read = le16_to_cpu(nvm_resp->offset);
resp_data = nvm_resp->data;
if (ret) {
if ((offset != 0) &&
(ret == READ_NVM_CHUNK_NOT_VALID_ADDRESS)) {
/*
* meaning of NOT_VALID_ADDRESS:
* driver try to read chunk from address that is
* multiple of 2K and got an error since addr is empty.
* meaning of (offset != 0): driver already
* read valid data from another chunk so this case
* is not an error.
*/
IWL_DEBUG_EEPROM(mvm->trans->dev,
"NVM access command failed on offset 0x%x since that section size is multiple 2K\n",
offset);
ret = 0;
} else {
IWL_DEBUG_EEPROM(mvm->trans->dev,
"NVM access command failed with status %d (device: %s)\n",
ret, mvm->cfg->name);
ret = -EIO;
}
goto exit;
}
if (offset_read != offset) {
IWL_ERR(mvm, "NVM ACCESS response with invalid offset %d\n",
offset_read);
ret = -EINVAL;
goto exit;
}
/* Write data to NVM */
memcpy(data + offset, resp_data, bytes_read);
ret = bytes_read;
exit:
iwl_free_resp(&cmd);
return ret;
}
static int iwl_nvm_write_section(struct iwl_mvm *mvm, u16 section,
const u8 *data, u16 length)
{
int offset = 0;
/* copy data in chunks of 2k (and remainder if any) */
while (offset < length) {
int chunk_size, ret;
chunk_size = min(IWL_NVM_DEFAULT_CHUNK_SIZE,
length - offset);
ret = iwl_nvm_write_chunk(mvm, section, offset,
chunk_size, data + offset);
if (ret < 0)
return ret;
offset += chunk_size;
}
return 0;
}
/*
* Reads an NVM section completely.
* NICs prior to 7000 family doesn't have a real NVM, but just read
* section 0 which is the EEPROM. Because the EEPROM reading is unlimited
* by uCode, we need to manually check in this case that we don't
* overflow and try to read more than the EEPROM size.
* For 7000 family NICs, we supply the maximal size we can read, and
* the uCode fills the response with as much data as we can,
* without overflowing, so no check is needed.
*/
static int iwl_nvm_read_section(struct iwl_mvm *mvm, u16 section,
u8 *data, u32 size_read)
{
u16 length, offset = 0;
int ret;
/* Set nvm section read length */
length = IWL_NVM_DEFAULT_CHUNK_SIZE;
ret = length;
/* Read the NVM until exhausted (reading less than requested) */
while (ret == length) {
/* Check no memory assumptions fail and cause an overflow */
if ((size_read + offset + length) >
mvm->cfg->base_params->eeprom_size) {
IWL_ERR(mvm, "EEPROM size is too small for NVM\n");
return -ENOBUFS;
}
ret = iwl_nvm_read_chunk(mvm, section, offset, length, data);
if (ret < 0) {
IWL_DEBUG_EEPROM(mvm->trans->dev,
"Cannot read NVM from section %d offset %d, length %d\n",
section, offset, length);
return ret;
}
offset += ret;
}
iwl_nvm_fixups(mvm->trans->hw_id, section, data, offset);
IWL_DEBUG_EEPROM(mvm->trans->dev,
"NVM section %d read completed\n", section);
return offset;
}
static struct iwl_nvm_data *
iwl_parse_nvm_sections(struct iwl_mvm *mvm)
{
struct iwl_nvm_section *sections = mvm->nvm_sections;
const __be16 *hw;
const __le16 *sw, *calib, *regulatory, *mac_override, *phy_sku;
bool lar_enabled;
int regulatory_type;
/* Checking for required sections */
if (mvm->trans->cfg->nvm_type != IWL_NVM_EXT) {
if (!mvm->nvm_sections[NVM_SECTION_TYPE_SW].data ||
!mvm->nvm_sections[mvm->cfg->nvm_hw_section_num].data) {
IWL_ERR(mvm, "Can't parse empty OTP/NVM sections\n");
return NULL;
}
} else {
if (mvm->trans->cfg->nvm_type == IWL_NVM_SDP)
regulatory_type = NVM_SECTION_TYPE_REGULATORY_SDP;
else
regulatory_type = NVM_SECTION_TYPE_REGULATORY;
/* SW and REGULATORY sections are mandatory */
if (!mvm->nvm_sections[NVM_SECTION_TYPE_SW].data ||
!mvm->nvm_sections[regulatory_type].data) {
IWL_ERR(mvm,
"Can't parse empty family 8000 OTP/NVM sections\n");
return NULL;
}
/* MAC_OVERRIDE or at least HW section must exist */
if (!mvm->nvm_sections[mvm->cfg->nvm_hw_section_num].data &&
!mvm->nvm_sections[NVM_SECTION_TYPE_MAC_OVERRIDE].data) {
IWL_ERR(mvm,
"Can't parse mac_address, empty sections\n");
return NULL;
}
/* PHY_SKU section is mandatory in B0 */
if (!mvm->nvm_sections[NVM_SECTION_TYPE_PHY_SKU].data) {
IWL_ERR(mvm,
"Can't parse phy_sku in B0, empty sections\n");
return NULL;
}
}
hw = (const __be16 *)sections[mvm->cfg->nvm_hw_section_num].data;
sw = (const __le16 *)sections[NVM_SECTION_TYPE_SW].data;
calib = (const __le16 *)sections[NVM_SECTION_TYPE_CALIBRATION].data;
mac_override =
(const __le16 *)sections[NVM_SECTION_TYPE_MAC_OVERRIDE].data;
phy_sku = (const __le16 *)sections[NVM_SECTION_TYPE_PHY_SKU].data;
regulatory = mvm->trans->cfg->nvm_type == IWL_NVM_SDP ?
(const __le16 *)sections[NVM_SECTION_TYPE_REGULATORY_SDP].data :
(const __le16 *)sections[NVM_SECTION_TYPE_REGULATORY].data;
lar_enabled = !iwlwifi_mod_params.lar_disable &&
fw_has_capa(&mvm->fw->ucode_capa,
IWL_UCODE_TLV_CAPA_LAR_SUPPORT);
return iwl_parse_nvm_data(mvm->trans, mvm->cfg, hw, sw, calib,
regulatory, mac_override, phy_sku,
mvm->fw->valid_tx_ant, mvm->fw->valid_rx_ant,
lar_enabled);
}
/* Loads the NVM data stored in mvm->nvm_sections into the NIC */
int iwl_mvm_load_nvm_to_nic(struct iwl_mvm *mvm)
{
int i, ret = 0;
struct iwl_nvm_section *sections = mvm->nvm_sections;
IWL_DEBUG_EEPROM(mvm->trans->dev, "'Write to NVM\n");
for (i = 0; i < ARRAY_SIZE(mvm->nvm_sections); i++) {
if (!mvm->nvm_sections[i].data || !mvm->nvm_sections[i].length)
continue;
ret = iwl_nvm_write_section(mvm, i, sections[i].data,
sections[i].length);
if (ret < 0) {
IWL_ERR(mvm, "iwl_mvm_send_cmd failed: %d\n", ret);
break;
}
}
return ret;
}
int iwl_nvm_init(struct iwl_mvm *mvm)
{
int ret, section;
u32 size_read = 0;
u8 *nvm_buffer, *temp;
const char *nvm_file_C = mvm->cfg->default_nvm_file_C_step;
if (WARN_ON_ONCE(mvm->cfg->nvm_hw_section_num >= NVM_MAX_NUM_SECTIONS))
return -EINVAL;
/* load NVM values from nic */
/* Read From FW NVM */
IWL_DEBUG_EEPROM(mvm->trans->dev, "Read from NVM\n");
nvm_buffer = kmalloc(mvm->cfg->base_params->eeprom_size,
GFP_KERNEL);
if (!nvm_buffer)
return -ENOMEM;
for (section = 0; section < NVM_MAX_NUM_SECTIONS; section++) {
/* we override the constness for initial read */
ret = iwl_nvm_read_section(mvm, section, nvm_buffer,
size_read);
if (ret < 0)
continue;
size_read += ret;
temp = kmemdup(nvm_buffer, ret, GFP_KERNEL);
if (!temp) {
ret = -ENOMEM;
break;
}
iwl_nvm_fixups(mvm->trans->hw_id, section, temp, ret);
mvm->nvm_sections[section].data = temp;
mvm->nvm_sections[section].length = ret;
#ifdef CONFIG_IWLWIFI_DEBUGFS
switch (section) {
case NVM_SECTION_TYPE_SW:
mvm->nvm_sw_blob.data = temp;
mvm->nvm_sw_blob.size = ret;
break;
case NVM_SECTION_TYPE_CALIBRATION:
mvm->nvm_calib_blob.data = temp;
mvm->nvm_calib_blob.size = ret;
break;
case NVM_SECTION_TYPE_PRODUCTION:
mvm->nvm_prod_blob.data = temp;
mvm->nvm_prod_blob.size = ret;
break;
case NVM_SECTION_TYPE_PHY_SKU:
mvm->nvm_phy_sku_blob.data = temp;
mvm->nvm_phy_sku_blob.size = ret;
break;
default:
if (section == mvm->cfg->nvm_hw_section_num) {
mvm->nvm_hw_blob.data = temp;
mvm->nvm_hw_blob.size = ret;
break;
}
}
#endif
}
if (!size_read)
IWL_ERR(mvm, "OTP is blank\n");
kfree(nvm_buffer);
/* Only if PNVM selected in the mod param - load external NVM */
if (mvm->nvm_file_name) {
/* read External NVM file from the mod param */
ret = iwl_read_external_nvm(mvm->trans, mvm->nvm_file_name,
mvm->nvm_sections);
if (ret) {
mvm->nvm_file_name = nvm_file_C;
if ((ret == -EFAULT || ret == -ENOENT) &&
mvm->nvm_file_name) {
/* in case nvm file was failed try again */
ret = iwl_read_external_nvm(mvm->trans,
mvm->nvm_file_name,
mvm->nvm_sections);
if (ret)
return ret;
} else {
return ret;
}
}
}
/* parse the relevant nvm sections */
mvm->nvm_data = iwl_parse_nvm_sections(mvm);
if (!mvm->nvm_data)
return -ENODATA;
IWL_DEBUG_EEPROM(mvm->trans->dev, "nvm version = %x\n",
mvm->nvm_data->nvm_version);
return 0;
}
struct iwl_mcc_update_resp *
iwl_mvm_update_mcc(struct iwl_mvm *mvm, const char *alpha2,
enum iwl_mcc_source src_id)
{
struct iwl_mcc_update_cmd mcc_update_cmd = {
.mcc = cpu_to_le16(alpha2[0] << 8 | alpha2[1]),
.source_id = (u8)src_id,
};
struct iwl_mcc_update_resp *resp_cp;
struct iwl_rx_packet *pkt;
struct iwl_host_cmd cmd = {
.id = MCC_UPDATE_CMD,
.flags = CMD_WANT_SKB,
.data = { &mcc_update_cmd },
};
int ret;
u32 status;
int resp_len, n_channels;
u16 mcc;
if (WARN_ON_ONCE(!iwl_mvm_is_lar_supported(mvm)))
return ERR_PTR(-EOPNOTSUPP);
cmd.len[0] = sizeof(struct iwl_mcc_update_cmd);
IWL_DEBUG_LAR(mvm, "send MCC update to FW with '%c%c' src = %d\n",
alpha2[0], alpha2[1], src_id);
ret = iwl_mvm_send_cmd(mvm, &cmd);
if (ret)
return ERR_PTR(ret);
pkt = cmd.resp_pkt;
/* Extract MCC response */
if (fw_has_capa(&mvm->fw->ucode_capa,
IWL_UCODE_TLV_CAPA_MCC_UPDATE_11AX_SUPPORT)) {
struct iwl_mcc_update_resp *mcc_resp = (void *)pkt->data;
n_channels = __le32_to_cpu(mcc_resp->n_channels);
resp_len = sizeof(struct iwl_mcc_update_resp) +
n_channels * sizeof(__le32);
resp_cp = kmemdup(mcc_resp, resp_len, GFP_KERNEL);
if (!resp_cp) {
resp_cp = ERR_PTR(-ENOMEM);
goto exit;
}
} else {
struct iwl_mcc_update_resp_v3 *mcc_resp_v3 = (void *)pkt->data;
n_channels = __le32_to_cpu(mcc_resp_v3->n_channels);
resp_len = sizeof(struct iwl_mcc_update_resp) +
n_channels * sizeof(__le32);
resp_cp = kzalloc(resp_len, GFP_KERNEL);
if (!resp_cp) {
resp_cp = ERR_PTR(-ENOMEM);
goto exit;
}
resp_cp->status = mcc_resp_v3->status;
resp_cp->mcc = mcc_resp_v3->mcc;
resp_cp->cap = cpu_to_le16(mcc_resp_v3->cap);
resp_cp->source_id = mcc_resp_v3->source_id;
resp_cp->time = mcc_resp_v3->time;
resp_cp->geo_info = mcc_resp_v3->geo_info;
resp_cp->n_channels = mcc_resp_v3->n_channels;
memcpy(resp_cp->channels, mcc_resp_v3->channels,
n_channels * sizeof(__le32));
}
status = le32_to_cpu(resp_cp->status);
mcc = le16_to_cpu(resp_cp->mcc);
/* W/A for a FW/NVM issue - returns 0x00 for the world domain */
if (mcc == 0) {
mcc = 0x3030; /* "00" - world */
resp_cp->mcc = cpu_to_le16(mcc);
}
IWL_DEBUG_LAR(mvm,
"MCC response status: 0x%x. new MCC: 0x%x ('%c%c') change: %d n_chans: %d\n",
status, mcc, mcc >> 8, mcc & 0xff,
!!(status == MCC_RESP_NEW_CHAN_PROFILE), n_channels);
exit:
iwl_free_resp(&cmd);
return resp_cp;
}
int iwl_mvm_init_mcc(struct iwl_mvm *mvm)
{
bool tlv_lar;
bool nvm_lar;
int retval;
struct ieee80211_regdomain *regd;
char mcc[3];
if (mvm->cfg->nvm_type == IWL_NVM_EXT) {
tlv_lar = fw_has_capa(&mvm->fw->ucode_capa,
IWL_UCODE_TLV_CAPA_LAR_SUPPORT);
nvm_lar = mvm->nvm_data->lar_enabled;
if (tlv_lar != nvm_lar)
IWL_INFO(mvm,
"Conflict between TLV & NVM regarding enabling LAR (TLV = %s NVM =%s)\n",
tlv_lar ? "enabled" : "disabled",
nvm_lar ? "enabled" : "disabled");
}
if (!iwl_mvm_is_lar_supported(mvm))
return 0;
/*
* try to replay the last set MCC to FW. If it doesn't exist,
* queue an update to cfg80211 to retrieve the default alpha2 from FW.
*/
retval = iwl_mvm_init_fw_regd(mvm);
if (retval != -ENOENT)
return retval;
/*
* Driver regulatory hint for initial update, this also informs the
* firmware we support wifi location updates.
* Disallow scans that might crash the FW while the LAR regdomain
* is not set.
*/
mvm->lar_regdom_set = false;
regd = iwl_mvm_get_current_regdomain(mvm, NULL);
if (IS_ERR_OR_NULL(regd))
return -EIO;
if (iwl_mvm_is_wifi_mcc_supported(mvm) &&
!iwl_acpi_get_mcc(mvm->dev, mcc)) {
kfree(regd);
regd = iwl_mvm_get_regdomain(mvm->hw->wiphy, mcc,
MCC_SOURCE_BIOS, NULL);
if (IS_ERR_OR_NULL(regd))
return -EIO;
}
retval = regulatory_set_wiphy_regd_sync_rtnl(mvm->hw->wiphy, regd);
kfree(regd);
return retval;
}
void iwl_mvm_rx_chub_update_mcc(struct iwl_mvm *mvm,
struct iwl_rx_cmd_buffer *rxb)
{
struct iwl_rx_packet *pkt = rxb_addr(rxb);
struct iwl_mcc_chub_notif *notif = (void *)pkt->data;
enum iwl_mcc_source src;
char mcc[3];
struct ieee80211_regdomain *regd;
lockdep_assert_held(&mvm->mutex);
if (iwl_mvm_is_vif_assoc(mvm) && notif->source_id == MCC_SOURCE_WIFI) {
IWL_DEBUG_LAR(mvm, "Ignore mcc update while associated\n");
return;
}
if (WARN_ON_ONCE(!iwl_mvm_is_lar_supported(mvm)))
return;
mcc[0] = le16_to_cpu(notif->mcc) >> 8;
mcc[1] = le16_to_cpu(notif->mcc) & 0xff;
mcc[2] = '\0';
src = notif->source_id;
IWL_DEBUG_LAR(mvm,
"RX: received chub update mcc cmd (mcc '%s' src %d)\n",
mcc, src);
regd = iwl_mvm_get_regdomain(mvm->hw->wiphy, mcc, src, NULL);
if (IS_ERR_OR_NULL(regd))
return;
regulatory_set_wiphy_regd(mvm->hw->wiphy, regd);
kfree(regd);
}