linux-stable/drivers/net/ethernet/intel/ice/ice_ethtool.c

1688 lines
52 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2018, Intel Corporation. */
/* ethtool support for ice */
#include "ice.h"
struct ice_stats {
char stat_string[ETH_GSTRING_LEN];
int sizeof_stat;
int stat_offset;
};
#define ICE_STAT(_type, _name, _stat) { \
.stat_string = _name, \
.sizeof_stat = FIELD_SIZEOF(_type, _stat), \
.stat_offset = offsetof(_type, _stat) \
}
#define ICE_VSI_STAT(_name, _stat) \
ICE_STAT(struct ice_vsi, _name, _stat)
#define ICE_PF_STAT(_name, _stat) \
ICE_STAT(struct ice_pf, _name, _stat)
static int ice_q_stats_len(struct net_device *netdev)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
return ((np->vsi->alloc_txq + np->vsi->alloc_rxq) *
(sizeof(struct ice_q_stats) / sizeof(u64)));
}
#define ICE_PF_STATS_LEN ARRAY_SIZE(ice_gstrings_pf_stats)
#define ICE_VSI_STATS_LEN ARRAY_SIZE(ice_gstrings_vsi_stats)
#define ICE_ALL_STATS_LEN(n) (ICE_PF_STATS_LEN + ICE_VSI_STATS_LEN + \
ice_q_stats_len(n))
static const struct ice_stats ice_gstrings_vsi_stats[] = {
ICE_VSI_STAT("tx_unicast", eth_stats.tx_unicast),
ICE_VSI_STAT("rx_unicast", eth_stats.rx_unicast),
ICE_VSI_STAT("tx_multicast", eth_stats.tx_multicast),
ICE_VSI_STAT("rx_multicast", eth_stats.rx_multicast),
ICE_VSI_STAT("tx_broadcast", eth_stats.tx_broadcast),
ICE_VSI_STAT("rx_broadcast", eth_stats.rx_broadcast),
ICE_VSI_STAT("tx_bytes", eth_stats.tx_bytes),
ICE_VSI_STAT("rx_bytes", eth_stats.rx_bytes),
ICE_VSI_STAT("rx_discards", eth_stats.rx_discards),
ICE_VSI_STAT("tx_errors", eth_stats.tx_errors),
ICE_VSI_STAT("tx_linearize", tx_linearize),
ICE_VSI_STAT("rx_unknown_protocol", eth_stats.rx_unknown_protocol),
ICE_VSI_STAT("rx_alloc_fail", rx_buf_failed),
ICE_VSI_STAT("rx_pg_alloc_fail", rx_page_failed),
};
/* These PF_STATs might look like duplicates of some NETDEV_STATs,
* but they aren't. This device is capable of supporting multiple
* VSIs/netdevs on a single PF. The NETDEV_STATs are for individual
* netdevs whereas the PF_STATs are for the physical function that's
* hosting these netdevs.
*
* The PF_STATs are appended to the netdev stats only when ethtool -S
* is queried on the base PF netdev.
*/
static struct ice_stats ice_gstrings_pf_stats[] = {
ICE_PF_STAT("tx_bytes", stats.eth.tx_bytes),
ICE_PF_STAT("rx_bytes", stats.eth.rx_bytes),
ICE_PF_STAT("tx_unicast", stats.eth.tx_unicast),
ICE_PF_STAT("rx_unicast", stats.eth.rx_unicast),
ICE_PF_STAT("tx_multicast", stats.eth.tx_multicast),
ICE_PF_STAT("rx_multicast", stats.eth.rx_multicast),
ICE_PF_STAT("tx_broadcast", stats.eth.tx_broadcast),
ICE_PF_STAT("rx_broadcast", stats.eth.rx_broadcast),
ICE_PF_STAT("tx_errors", stats.eth.tx_errors),
ICE_PF_STAT("tx_size_64", stats.tx_size_64),
ICE_PF_STAT("rx_size_64", stats.rx_size_64),
ICE_PF_STAT("tx_size_127", stats.tx_size_127),
ICE_PF_STAT("rx_size_127", stats.rx_size_127),
ICE_PF_STAT("tx_size_255", stats.tx_size_255),
ICE_PF_STAT("rx_size_255", stats.rx_size_255),
ICE_PF_STAT("tx_size_511", stats.tx_size_511),
ICE_PF_STAT("rx_size_511", stats.rx_size_511),
ICE_PF_STAT("tx_size_1023", stats.tx_size_1023),
ICE_PF_STAT("rx_size_1023", stats.rx_size_1023),
ICE_PF_STAT("tx_size_1522", stats.tx_size_1522),
ICE_PF_STAT("rx_size_1522", stats.rx_size_1522),
ICE_PF_STAT("tx_size_big", stats.tx_size_big),
ICE_PF_STAT("rx_size_big", stats.rx_size_big),
ICE_PF_STAT("link_xon_tx", stats.link_xon_tx),
ICE_PF_STAT("link_xon_rx", stats.link_xon_rx),
ICE_PF_STAT("link_xoff_tx", stats.link_xoff_tx),
ICE_PF_STAT("link_xoff_rx", stats.link_xoff_rx),
ICE_PF_STAT("tx_dropped_link_down", stats.tx_dropped_link_down),
ICE_PF_STAT("rx_undersize", stats.rx_undersize),
ICE_PF_STAT("rx_fragments", stats.rx_fragments),
ICE_PF_STAT("rx_oversize", stats.rx_oversize),
ICE_PF_STAT("rx_jabber", stats.rx_jabber),
ICE_PF_STAT("rx_csum_bad", hw_csum_rx_error),
ICE_PF_STAT("rx_length_errors", stats.rx_len_errors),
ICE_PF_STAT("rx_dropped", stats.eth.rx_discards),
ICE_PF_STAT("rx_crc_errors", stats.crc_errors),
ICE_PF_STAT("illegal_bytes", stats.illegal_bytes),
ICE_PF_STAT("mac_local_faults", stats.mac_local_faults),
ICE_PF_STAT("mac_remote_faults", stats.mac_remote_faults),
};
static u32 ice_regs_dump_list[] = {
PFGEN_STATE,
PRTGEN_STATUS,
QRX_CTRL(0),
QINT_TQCTL(0),
QINT_RQCTL(0),
PFINT_OICR_ENA,
QRX_ITR(0),
};
/**
* ice_nvm_version_str - format the NVM version strings
* @hw: ptr to the hardware info
*/
static char *ice_nvm_version_str(struct ice_hw *hw)
{
static char buf[ICE_ETHTOOL_FWVER_LEN];
u8 ver, patch;
u32 full_ver;
u16 build;
full_ver = hw->nvm.oem_ver;
ver = (u8)((full_ver & ICE_OEM_VER_MASK) >> ICE_OEM_VER_SHIFT);
build = (u16)((full_ver & ICE_OEM_VER_BUILD_MASK) >>
ICE_OEM_VER_BUILD_SHIFT);
patch = (u8)(full_ver & ICE_OEM_VER_PATCH_MASK);
snprintf(buf, sizeof(buf), "%x.%02x 0x%x %d.%d.%d",
(hw->nvm.ver & ICE_NVM_VER_HI_MASK) >> ICE_NVM_VER_HI_SHIFT,
(hw->nvm.ver & ICE_NVM_VER_LO_MASK) >> ICE_NVM_VER_LO_SHIFT,
hw->nvm.eetrack, ver, build, patch);
return buf;
}
static void
ice_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *drvinfo)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
struct ice_pf *pf = vsi->back;
strlcpy(drvinfo->driver, KBUILD_MODNAME, sizeof(drvinfo->driver));
strlcpy(drvinfo->version, ice_drv_ver, sizeof(drvinfo->version));
strlcpy(drvinfo->fw_version, ice_nvm_version_str(&pf->hw),
sizeof(drvinfo->fw_version));
strlcpy(drvinfo->bus_info, pci_name(pf->pdev),
sizeof(drvinfo->bus_info));
}
static int ice_get_regs_len(struct net_device __always_unused *netdev)
{
return sizeof(ice_regs_dump_list);
}
static void
ice_get_regs(struct net_device *netdev, struct ethtool_regs *regs, void *p)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_pf *pf = np->vsi->back;
struct ice_hw *hw = &pf->hw;
u32 *regs_buf = (u32 *)p;
int i;
regs->version = 1;
for (i = 0; i < ARRAY_SIZE(ice_regs_dump_list); ++i)
regs_buf[i] = rd32(hw, ice_regs_dump_list[i]);
}
static u32 ice_get_msglevel(struct net_device *netdev)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_pf *pf = np->vsi->back;
#ifndef CONFIG_DYNAMIC_DEBUG
if (pf->hw.debug_mask)
netdev_info(netdev, "hw debug_mask: 0x%llX\n",
pf->hw.debug_mask);
#endif /* !CONFIG_DYNAMIC_DEBUG */
return pf->msg_enable;
}
static void ice_set_msglevel(struct net_device *netdev, u32 data)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_pf *pf = np->vsi->back;
#ifndef CONFIG_DYNAMIC_DEBUG
if (ICE_DBG_USER & data)
pf->hw.debug_mask = data;
else
pf->msg_enable = data;
#else
pf->msg_enable = data;
#endif /* !CONFIG_DYNAMIC_DEBUG */
}
static void ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
char *p = (char *)data;
unsigned int i;
switch (stringset) {
case ETH_SS_STATS:
for (i = 0; i < ICE_VSI_STATS_LEN; i++) {
snprintf(p, ETH_GSTRING_LEN, "%s",
ice_gstrings_vsi_stats[i].stat_string);
p += ETH_GSTRING_LEN;
}
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
ice_for_each_alloc_txq(vsi, i) {
snprintf(p, ETH_GSTRING_LEN,
"tx-queue-%u.tx_packets", i);
p += ETH_GSTRING_LEN;
snprintf(p, ETH_GSTRING_LEN, "tx-queue-%u.tx_bytes", i);
p += ETH_GSTRING_LEN;
}
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
ice_for_each_alloc_rxq(vsi, i) {
snprintf(p, ETH_GSTRING_LEN,
"rx-queue-%u.rx_packets", i);
p += ETH_GSTRING_LEN;
snprintf(p, ETH_GSTRING_LEN, "rx-queue-%u.rx_bytes", i);
p += ETH_GSTRING_LEN;
}
if (vsi->type != ICE_VSI_PF)
return;
for (i = 0; i < ICE_PF_STATS_LEN; i++) {
snprintf(p, ETH_GSTRING_LEN, "port.%s",
ice_gstrings_pf_stats[i].stat_string);
p += ETH_GSTRING_LEN;
}
break;
default:
break;
}
}
static int ice_get_sset_count(struct net_device *netdev, int sset)
{
switch (sset) {
case ETH_SS_STATS:
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
/* The number (and order) of strings reported *must* remain
* constant for a given netdevice. This function must not
* report a different number based on run time parameters
* (such as the number of queues in use, or the setting of
* a private ethtool flag). This is due to the nature of the
* ethtool stats API.
*
* User space programs such as ethtool must make 3 separate
* ioctl requests, one for size, one for the strings, and
* finally one for the stats. Since these cross into
* user space, changes to the number or size could result in
* undefined memory access or incorrect string<->value
* correlations for statistics.
*
* Even if it appears to be safe, changes to the size or
* order of strings will suffer from race conditions and are
* not safe.
*/
return ICE_ALL_STATS_LEN(netdev);
default:
return -EOPNOTSUPP;
}
}
static void
ice_get_ethtool_stats(struct net_device *netdev,
struct ethtool_stats __always_unused *stats, u64 *data)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
struct ice_pf *pf = vsi->back;
struct ice_ring *ring;
unsigned int j = 0;
int i = 0;
char *p;
for (j = 0; j < ICE_VSI_STATS_LEN; j++) {
p = (char *)vsi + ice_gstrings_vsi_stats[j].stat_offset;
data[i++] = (ice_gstrings_vsi_stats[j].sizeof_stat ==
sizeof(u64)) ? *(u64 *)p : *(u32 *)p;
}
/* populate per queue stats */
rcu_read_lock();
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
ice_for_each_alloc_txq(vsi, j) {
ring = READ_ONCE(vsi->tx_rings[j]);
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
if (ring) {
data[i++] = ring->stats.pkts;
data[i++] = ring->stats.bytes;
} else {
data[i++] = 0;
data[i++] = 0;
}
}
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
ice_for_each_alloc_rxq(vsi, j) {
ring = READ_ONCE(vsi->rx_rings[j]);
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
if (ring) {
data[i++] = ring->stats.pkts;
data[i++] = ring->stats.bytes;
} else {
data[i++] = 0;
data[i++] = 0;
}
}
rcu_read_unlock();
if (vsi->type != ICE_VSI_PF)
return;
for (j = 0; j < ICE_PF_STATS_LEN; j++) {
p = (char *)pf + ice_gstrings_pf_stats[j].stat_offset;
data[i++] = (ice_gstrings_pf_stats[j].sizeof_stat ==
sizeof(u64)) ? *(u64 *)p : *(u32 *)p;
}
}
/**
* ice_phy_type_to_ethtool - convert the phy_types to ethtool link modes
* @netdev: network interface device structure
* @ks: ethtool link ksettings struct to fill out
*/
static void ice_phy_type_to_ethtool(struct net_device *netdev,
struct ethtool_link_ksettings *ks)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_link_status *hw_link_info;
struct ice_vsi *vsi = np->vsi;
u64 phy_types_low;
hw_link_info = &vsi->port_info->phy.link_info;
phy_types_low = vsi->port_info->phy.phy_type_low;
ethtool_link_ksettings_zero_link_mode(ks, supported);
ethtool_link_ksettings_zero_link_mode(ks, advertising);
if (phy_types_low & ICE_PHY_TYPE_LOW_100BASE_TX ||
phy_types_low & ICE_PHY_TYPE_LOW_100M_SGMII) {
ethtool_link_ksettings_add_link_mode(ks, supported,
100baseT_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_100MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
100baseT_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_1G_SGMII) {
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseT_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_1000MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
1000baseT_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_KX) {
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseKX_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_1000MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
1000baseKX_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_SX ||
phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_LX) {
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseX_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_1000MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
1000baseX_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_T) {
ethtool_link_ksettings_add_link_mode(ks, supported,
2500baseT_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_2500MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
2500baseT_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_X ||
phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_KX) {
ethtool_link_ksettings_add_link_mode(ks, supported,
2500baseX_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_2500MB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
2500baseX_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_KR) {
ethtool_link_ksettings_add_link_mode(ks, supported,
5000baseT_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_5GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
5000baseT_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_10G_SFI_DA ||
phy_types_low & ICE_PHY_TYPE_LOW_10G_SFI_AOC_ACC ||
phy_types_low & ICE_PHY_TYPE_LOW_10G_SFI_C2C) {
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseT_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_10GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseT_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_KR_CR1) {
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseKR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_10GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseKR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_SR) {
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseSR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_10GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseSR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_LR) {
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseLR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_10GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseLR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR_S ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR1 ||
phy_types_low & ICE_PHY_TYPE_LOW_25G_AUI_AOC_ACC ||
phy_types_low & ICE_PHY_TYPE_LOW_25G_AUI_C2C) {
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseCR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_25GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
25000baseCR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_SR ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_LR) {
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseSR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_25GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
25000baseSR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR_S ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR1) {
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseKR_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_25GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
25000baseKR_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_KR4) {
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseKR4_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_40GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseKR4_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_CR4 ||
phy_types_low & ICE_PHY_TYPE_LOW_40G_XLAUI_AOC_ACC ||
phy_types_low & ICE_PHY_TYPE_LOW_40G_XLAUI) {
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseCR4_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_40GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseCR4_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_SR4) {
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseSR4_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_40GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseSR4_Full);
}
if (phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_LR4) {
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseLR4_Full);
if (hw_link_info->req_speeds & ICE_AQ_LINK_SPEED_40GB)
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseLR4_Full);
}
/* Autoneg PHY types */
if (phy_types_low & ICE_PHY_TYPE_LOW_100BASE_TX ||
phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_1000BASE_KX ||
phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_2500BASE_KX ||
phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_5GBASE_KR ||
phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_10GBASE_KR_CR1 ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_T ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR_S ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_CR1 ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR_S ||
phy_types_low & ICE_PHY_TYPE_LOW_25GBASE_KR1 ||
phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_CR4 ||
phy_types_low & ICE_PHY_TYPE_LOW_40GBASE_KR4) {
ethtool_link_ksettings_add_link_mode(ks, supported,
Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
Autoneg);
}
}
#define TEST_SET_BITS_TIMEOUT 50
#define TEST_SET_BITS_SLEEP_MAX 2000
#define TEST_SET_BITS_SLEEP_MIN 1000
/**
* ice_get_settings_link_up - Get Link settings for when link is up
* @ks: ethtool ksettings to fill in
* @netdev: network interface device structure
*/
static void ice_get_settings_link_up(struct ethtool_link_ksettings *ks,
struct net_device *netdev)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ethtool_link_ksettings cap_ksettings;
struct ice_link_status *link_info;
struct ice_vsi *vsi = np->vsi;
bool unrecog_phy_low = false;
link_info = &vsi->port_info->phy.link_info;
/* Initialize supported and advertised settings based on phy settings */
switch (link_info->phy_type_low) {
case ICE_PHY_TYPE_LOW_100BASE_TX:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
100baseT_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
100baseT_Full);
break;
case ICE_PHY_TYPE_LOW_100M_SGMII:
ethtool_link_ksettings_add_link_mode(ks, supported,
100baseT_Full);
break;
case ICE_PHY_TYPE_LOW_1000BASE_T:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseT_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
1000baseT_Full);
break;
case ICE_PHY_TYPE_LOW_1G_SGMII:
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseT_Full);
break;
case ICE_PHY_TYPE_LOW_1000BASE_SX:
case ICE_PHY_TYPE_LOW_1000BASE_LX:
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseX_Full);
break;
case ICE_PHY_TYPE_LOW_1000BASE_KX:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
1000baseKX_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
1000baseKX_Full);
break;
case ICE_PHY_TYPE_LOW_2500BASE_T:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
2500baseT_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising,
2500baseT_Full);
break;
case ICE_PHY_TYPE_LOW_2500BASE_X:
ethtool_link_ksettings_add_link_mode(ks, supported,
2500baseX_Full);
break;
case ICE_PHY_TYPE_LOW_2500BASE_KX:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
2500baseX_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
2500baseX_Full);
break;
case ICE_PHY_TYPE_LOW_5GBASE_T:
case ICE_PHY_TYPE_LOW_5GBASE_KR:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
5000baseT_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
5000baseT_Full);
break;
case ICE_PHY_TYPE_LOW_10GBASE_T:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseT_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseT_Full);
break;
case ICE_PHY_TYPE_LOW_10G_SFI_DA:
case ICE_PHY_TYPE_LOW_10G_SFI_AOC_ACC:
case ICE_PHY_TYPE_LOW_10G_SFI_C2C:
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseT_Full);
break;
case ICE_PHY_TYPE_LOW_10GBASE_SR:
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseSR_Full);
break;
case ICE_PHY_TYPE_LOW_10GBASE_LR:
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseLR_Full);
break;
case ICE_PHY_TYPE_LOW_10GBASE_KR_CR1:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
10000baseKR_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
10000baseKR_Full);
break;
case ICE_PHY_TYPE_LOW_25GBASE_T:
case ICE_PHY_TYPE_LOW_25GBASE_CR:
case ICE_PHY_TYPE_LOW_25GBASE_CR_S:
case ICE_PHY_TYPE_LOW_25GBASE_CR1:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseCR_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
25000baseCR_Full);
break;
case ICE_PHY_TYPE_LOW_25G_AUI_AOC_ACC:
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseCR_Full);
break;
case ICE_PHY_TYPE_LOW_25GBASE_SR:
case ICE_PHY_TYPE_LOW_25GBASE_LR:
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseSR_Full);
break;
case ICE_PHY_TYPE_LOW_25GBASE_KR:
case ICE_PHY_TYPE_LOW_25GBASE_KR1:
case ICE_PHY_TYPE_LOW_25GBASE_KR_S:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
25000baseKR_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
25000baseKR_Full);
break;
case ICE_PHY_TYPE_LOW_40GBASE_CR4:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseCR4_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseCR4_Full);
break;
case ICE_PHY_TYPE_LOW_40G_XLAUI_AOC_ACC:
case ICE_PHY_TYPE_LOW_40G_XLAUI:
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseCR4_Full);
break;
case ICE_PHY_TYPE_LOW_40GBASE_SR4:
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseSR4_Full);
break;
case ICE_PHY_TYPE_LOW_40GBASE_LR4:
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseLR4_Full);
break;
case ICE_PHY_TYPE_LOW_40GBASE_KR4:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported,
40000baseKR4_Full);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
40000baseKR4_Full);
break;
default:
unrecog_phy_low = true;
}
if (unrecog_phy_low) {
/* if we got here and link is up something bad is afoot */
netdev_info(netdev, "WARNING: Unrecognized PHY_Low (0x%llx).\n",
(u64)link_info->phy_type_low);
}
/* Now that we've worked out everything that could be supported by the
* current PHY type, get what is supported by the NVM and intersect
* them to get what is truly supported
*/
memset(&cap_ksettings, 0, sizeof(struct ethtool_link_ksettings));
ice_phy_type_to_ethtool(netdev, &cap_ksettings);
ethtool_intersect_link_masks(ks, &cap_ksettings);
switch (link_info->link_speed) {
case ICE_AQ_LINK_SPEED_40GB:
ks->base.speed = SPEED_40000;
break;
case ICE_AQ_LINK_SPEED_25GB:
ks->base.speed = SPEED_25000;
break;
case ICE_AQ_LINK_SPEED_20GB:
ks->base.speed = SPEED_20000;
break;
case ICE_AQ_LINK_SPEED_10GB:
ks->base.speed = SPEED_10000;
break;
case ICE_AQ_LINK_SPEED_5GB:
ks->base.speed = SPEED_5000;
break;
case ICE_AQ_LINK_SPEED_2500MB:
ks->base.speed = SPEED_2500;
break;
case ICE_AQ_LINK_SPEED_1000MB:
ks->base.speed = SPEED_1000;
break;
case ICE_AQ_LINK_SPEED_100MB:
ks->base.speed = SPEED_100;
break;
default:
netdev_info(netdev,
"WARNING: Unrecognized link_speed (0x%x).\n",
link_info->link_speed);
break;
}
ks->base.duplex = DUPLEX_FULL;
}
/**
* ice_get_settings_link_down - Get the Link settings when link is down
* @ks: ethtool ksettings to fill in
* @netdev: network interface device structure
*
* Reports link settings that can be determined when link is down
*/
static void
ice_get_settings_link_down(struct ethtool_link_ksettings *ks,
struct net_device __always_unused *netdev)
{
/* link is down and the driver needs to fall back on
* supported phy types to figure out what info to display
*/
ice_phy_type_to_ethtool(netdev, ks);
/* With no link, speed and duplex are unknown */
ks->base.speed = SPEED_UNKNOWN;
ks->base.duplex = DUPLEX_UNKNOWN;
}
/**
* ice_get_link_ksettings - Get Link Speed and Duplex settings
* @netdev: network interface device structure
* @ks: ethtool ksettings
*
* Reports speed/duplex settings based on media_type
*/
static int ice_get_link_ksettings(struct net_device *netdev,
struct ethtool_link_ksettings *ks)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_link_status *hw_link_info;
struct ice_vsi *vsi = np->vsi;
ethtool_link_ksettings_zero_link_mode(ks, supported);
ethtool_link_ksettings_zero_link_mode(ks, advertising);
hw_link_info = &vsi->port_info->phy.link_info;
/* set speed and duplex */
if (hw_link_info->link_info & ICE_AQ_LINK_UP)
ice_get_settings_link_up(ks, netdev);
else
ice_get_settings_link_down(ks, netdev);
/* set autoneg settings */
ks->base.autoneg = (hw_link_info->an_info & ICE_AQ_AN_COMPLETED) ?
AUTONEG_ENABLE : AUTONEG_DISABLE;
/* set media type settings */
switch (vsi->port_info->phy.media_type) {
case ICE_MEDIA_FIBER:
ethtool_link_ksettings_add_link_mode(ks, supported, FIBRE);
ks->base.port = PORT_FIBRE;
break;
case ICE_MEDIA_BASET:
ethtool_link_ksettings_add_link_mode(ks, supported, TP);
ethtool_link_ksettings_add_link_mode(ks, advertising, TP);
ks->base.port = PORT_TP;
break;
case ICE_MEDIA_BACKPLANE:
ethtool_link_ksettings_add_link_mode(ks, supported, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, supported, Backplane);
ethtool_link_ksettings_add_link_mode(ks, advertising, Autoneg);
ethtool_link_ksettings_add_link_mode(ks, advertising,
Backplane);
ks->base.port = PORT_NONE;
break;
case ICE_MEDIA_DA:
ethtool_link_ksettings_add_link_mode(ks, supported, FIBRE);
ethtool_link_ksettings_add_link_mode(ks, advertising, FIBRE);
ks->base.port = PORT_DA;
break;
default:
ks->base.port = PORT_OTHER;
break;
}
/* flow control is symmetric and always supported */
ethtool_link_ksettings_add_link_mode(ks, supported, Pause);
switch (vsi->port_info->fc.req_mode) {
case ICE_FC_FULL:
ethtool_link_ksettings_add_link_mode(ks, advertising, Pause);
break;
case ICE_FC_TX_PAUSE:
ethtool_link_ksettings_add_link_mode(ks, advertising,
Asym_Pause);
break;
case ICE_FC_RX_PAUSE:
ethtool_link_ksettings_add_link_mode(ks, advertising, Pause);
ethtool_link_ksettings_add_link_mode(ks, advertising,
Asym_Pause);
break;
case ICE_FC_PFC:
default:
ethtool_link_ksettings_del_link_mode(ks, advertising, Pause);
ethtool_link_ksettings_del_link_mode(ks, advertising,
Asym_Pause);
break;
}
return 0;
}
/**
* ice_ksettings_find_adv_link_speed - Find advertising link speed
* @ks: ethtool ksettings
*/
static u16
ice_ksettings_find_adv_link_speed(const struct ethtool_link_ksettings *ks)
{
u16 adv_link_speed = 0;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
100baseT_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_100MB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
1000baseX_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_1000MB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
1000baseT_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
1000baseKX_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_1000MB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
2500baseT_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_2500MB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
2500baseX_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_2500MB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
5000baseT_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_5GB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
10000baseT_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
10000baseKR_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_10GB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
10000baseSR_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
10000baseLR_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_10GB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
25000baseCR_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
25000baseSR_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
25000baseKR_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_25GB;
if (ethtool_link_ksettings_test_link_mode(ks, advertising,
40000baseCR4_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
40000baseSR4_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
40000baseLR4_Full) ||
ethtool_link_ksettings_test_link_mode(ks, advertising,
40000baseKR4_Full))
adv_link_speed |= ICE_AQ_LINK_SPEED_40GB;
return adv_link_speed;
}
/**
* ice_setup_autoneg
* @p: port info
* @ks: ethtool_link_ksettings
* @config: configuration that will be sent down to FW
* @autoneg_enabled: autonegotiation is enabled or not
* @autoneg_changed: will there a change in autonegotiation
* @netdev: network interface device structure
*
* Setup PHY autonegotiation feature
*/
static int
ice_setup_autoneg(struct ice_port_info *p, struct ethtool_link_ksettings *ks,
struct ice_aqc_set_phy_cfg_data *config,
u8 autoneg_enabled, u8 *autoneg_changed,
struct net_device *netdev)
{
int err = 0;
*autoneg_changed = 0;
/* Check autoneg */
if (autoneg_enabled == AUTONEG_ENABLE) {
/* If autoneg was not already enabled */
if (!(p->phy.link_info.an_info & ICE_AQ_AN_COMPLETED)) {
/* If autoneg is not supported, return error */
if (!ethtool_link_ksettings_test_link_mode(ks,
supported,
Autoneg)) {
netdev_info(netdev, "Autoneg not supported on this phy.\n");
err = -EINVAL;
} else {
/* Autoneg is allowed to change */
config->caps |= ICE_AQ_PHY_ENA_AUTO_LINK_UPDT;
*autoneg_changed = 1;
}
}
} else {
/* If autoneg is currently enabled */
if (p->phy.link_info.an_info & ICE_AQ_AN_COMPLETED) {
/* If autoneg is supported 10GBASE_T is the only phy
* that can disable it, so otherwise return error
*/
if (ethtool_link_ksettings_test_link_mode(ks,
supported,
Autoneg)) {
netdev_info(netdev, "Autoneg cannot be disabled on this phy\n");
err = -EINVAL;
} else {
/* Autoneg is allowed to change */
config->caps &= ~ICE_AQ_PHY_ENA_AUTO_LINK_UPDT;
*autoneg_changed = 1;
}
}
}
return err;
}
/**
* ice_set_link_ksettings - Set Speed and Duplex
* @netdev: network interface device structure
* @ks: ethtool ksettings
*
* Set speed/duplex per media_types advertised/forced
*/
static int ice_set_link_ksettings(struct net_device *netdev,
const struct ethtool_link_ksettings *ks)
{
u8 autoneg, timeout = TEST_SET_BITS_TIMEOUT, lport = 0;
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ethtool_link_ksettings safe_ks, copy_ks;
struct ice_aqc_get_phy_caps_data *abilities;
u16 adv_link_speed, curr_link_speed, idx;
struct ice_aqc_set_phy_cfg_data config;
struct ice_pf *pf = np->vsi->back;
struct ice_port_info *p;
u8 autoneg_changed = 0;
enum ice_status status;
u64 phy_type_low;
int err = 0;
bool linkup;
p = np->vsi->port_info;
if (!p)
return -EOPNOTSUPP;
/* Check if this is lan vsi */
for (idx = 0 ; idx < pf->num_alloc_vsi ; idx++) {
if (pf->vsi[idx]->type == ICE_VSI_PF) {
if (np->vsi != pf->vsi[idx])
return -EOPNOTSUPP;
break;
}
}
if (p->phy.media_type != ICE_MEDIA_BASET &&
p->phy.media_type != ICE_MEDIA_FIBER &&
p->phy.media_type != ICE_MEDIA_BACKPLANE &&
p->phy.media_type != ICE_MEDIA_DA &&
p->phy.link_info.link_info & ICE_AQ_LINK_UP)
return -EOPNOTSUPP;
/* copy the ksettings to copy_ks to avoid modifying the original */
memcpy(&copy_ks, ks, sizeof(struct ethtool_link_ksettings));
/* save autoneg out of ksettings */
autoneg = copy_ks.base.autoneg;
memset(&safe_ks, 0, sizeof(safe_ks));
/* Get link modes supported by hardware.*/
ice_phy_type_to_ethtool(netdev, &safe_ks);
/* and check against modes requested by user.
* Return an error if unsupported mode was set.
*/
if (!bitmap_subset(copy_ks.link_modes.advertising,
safe_ks.link_modes.supported,
__ETHTOOL_LINK_MODE_MASK_NBITS))
return -EINVAL;
/* get our own copy of the bits to check against */
memset(&safe_ks, 0, sizeof(struct ethtool_link_ksettings));
safe_ks.base.cmd = copy_ks.base.cmd;
safe_ks.base.link_mode_masks_nwords =
copy_ks.base.link_mode_masks_nwords;
ice_get_link_ksettings(netdev, &safe_ks);
/* set autoneg back to what it currently is */
copy_ks.base.autoneg = safe_ks.base.autoneg;
/* we don't compare the speed */
copy_ks.base.speed = safe_ks.base.speed;
/* If copy_ks.base and safe_ks.base are not the same now, then they are
* trying to set something that we do not support.
*/
if (memcmp(&copy_ks.base, &safe_ks.base,
sizeof(struct ethtool_link_settings)))
return -EOPNOTSUPP;
while (test_and_set_bit(__ICE_CFG_BUSY, pf->state)) {
timeout--;
if (!timeout)
return -EBUSY;
usleep_range(TEST_SET_BITS_SLEEP_MIN, TEST_SET_BITS_SLEEP_MAX);
}
abilities = devm_kzalloc(&pf->pdev->dev, sizeof(*abilities),
GFP_KERNEL);
if (!abilities)
return -ENOMEM;
/* Get the current phy config */
status = ice_aq_get_phy_caps(p, false, ICE_AQC_REPORT_SW_CFG, abilities,
NULL);
if (status) {
err = -EAGAIN;
goto done;
}
/* Copy abilities to config in case autoneg is not set below */
memset(&config, 0, sizeof(struct ice_aqc_set_phy_cfg_data));
config.caps = abilities->caps & ~ICE_AQC_PHY_AN_MODE;
if (abilities->caps & ICE_AQC_PHY_AN_MODE)
config.caps |= ICE_AQ_PHY_ENA_AUTO_LINK_UPDT;
/* Check autoneg */
err = ice_setup_autoneg(p, &safe_ks, &config, autoneg, &autoneg_changed,
netdev);
if (err)
goto done;
/* Call to get the current link speed */
p->phy.get_link_info = true;
status = ice_get_link_status(p, &linkup);
if (status) {
err = -EAGAIN;
goto done;
}
curr_link_speed = p->phy.link_info.link_speed;
adv_link_speed = ice_ksettings_find_adv_link_speed(ks);
/* If speed didn't get set, set it to what it currently is.
* This is needed because if advertise is 0 (as it is when autoneg
* is disabled) then speed won't get set.
*/
if (!adv_link_speed)
adv_link_speed = curr_link_speed;
/* Convert the advertise link speeds to their corresponded PHY_TYPE */
ice_update_phy_type(&phy_type_low, adv_link_speed);
if (!autoneg_changed && adv_link_speed == curr_link_speed) {
netdev_info(netdev, "Nothing changed, exiting without setting anything.\n");
goto done;
}
/* copy over the rest of the abilities */
config.low_power_ctrl = abilities->low_power_ctrl;
config.eee_cap = abilities->eee_cap;
config.eeer_value = abilities->eeer_value;
config.link_fec_opt = abilities->link_fec_options;
/* save the requested speeds */
p->phy.link_info.req_speeds = adv_link_speed;
/* set link and auto negotiation so changes take effect */
config.caps |= ICE_AQ_PHY_ENA_LINK;
if (phy_type_low) {
config.phy_type_low = cpu_to_le64(phy_type_low) &
abilities->phy_type_low;
} else {
err = -EAGAIN;
netdev_info(netdev, "Nothing changed. No PHY_TYPE is corresponded to advertised link speed.\n");
goto done;
}
/* If link is up put link down */
if (p->phy.link_info.link_info & ICE_AQ_LINK_UP) {
/* Tell the OS link is going down, the link will go
* back up when fw says it is ready asynchronously
*/
ice_print_link_msg(np->vsi, false);
netif_carrier_off(netdev);
netif_tx_stop_all_queues(netdev);
}
/* make the aq call */
status = ice_aq_set_phy_cfg(&pf->hw, lport, &config, NULL);
if (status) {
netdev_info(netdev, "Set phy config failed,\n");
err = -EAGAIN;
}
done:
devm_kfree(&pf->pdev->dev, abilities);
clear_bit(__ICE_CFG_BUSY, pf->state);
return err;
}
/**
* ice_get_rxnfc - command to get RX flow classification rules
* @netdev: network interface device structure
* @cmd: ethtool rxnfc command
* @rule_locs: buffer to rturn Rx flow classification rules
*
* Returns Success if the command is supported.
*/
static int ice_get_rxnfc(struct net_device *netdev, struct ethtool_rxnfc *cmd,
u32 __always_unused *rule_locs)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
int ret = -EOPNOTSUPP;
switch (cmd->cmd) {
case ETHTOOL_GRXRINGS:
cmd->data = vsi->rss_size;
ret = 0;
break;
default:
break;
}
return ret;
}
static void
ice_get_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
ring->rx_max_pending = ICE_MAX_NUM_DESC;
ring->tx_max_pending = ICE_MAX_NUM_DESC;
ring->rx_pending = vsi->rx_rings[0]->count;
ring->tx_pending = vsi->tx_rings[0]->count;
ring->rx_mini_pending = ICE_MIN_NUM_DESC;
ring->rx_mini_max_pending = 0;
ring->rx_jumbo_max_pending = 0;
ring->rx_jumbo_pending = 0;
}
static int
ice_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring)
{
struct ice_ring *tx_rings = NULL, *rx_rings = NULL;
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
struct ice_pf *pf = vsi->back;
int i, timeout = 50, err = 0;
u32 new_rx_cnt, new_tx_cnt;
if (ring->tx_pending > ICE_MAX_NUM_DESC ||
ring->tx_pending < ICE_MIN_NUM_DESC ||
ring->rx_pending > ICE_MAX_NUM_DESC ||
ring->rx_pending < ICE_MIN_NUM_DESC) {
netdev_err(netdev, "Descriptors requested (Tx: %d / Rx: %d) out of range [%d-%d]\n",
ring->tx_pending, ring->rx_pending,
ICE_MIN_NUM_DESC, ICE_MAX_NUM_DESC);
return -EINVAL;
}
new_tx_cnt = ALIGN(ring->tx_pending, ICE_REQ_DESC_MULTIPLE);
new_rx_cnt = ALIGN(ring->rx_pending, ICE_REQ_DESC_MULTIPLE);
/* if nothing to do return success */
if (new_tx_cnt == vsi->tx_rings[0]->count &&
new_rx_cnt == vsi->rx_rings[0]->count) {
netdev_dbg(netdev, "Nothing to change, descriptor count is same as requested\n");
return 0;
}
while (test_and_set_bit(__ICE_CFG_BUSY, pf->state)) {
timeout--;
if (!timeout)
return -EBUSY;
usleep_range(1000, 2000);
}
/* set for the next time the netdev is started */
if (!netif_running(vsi->netdev)) {
for (i = 0; i < vsi->alloc_txq; i++)
vsi->tx_rings[i]->count = new_tx_cnt;
for (i = 0; i < vsi->alloc_rxq; i++)
vsi->rx_rings[i]->count = new_rx_cnt;
netdev_dbg(netdev, "Link is down, descriptor count change happens when link is brought up\n");
goto done;
}
if (new_tx_cnt == vsi->tx_rings[0]->count)
goto process_rx;
/* alloc updated Tx resources */
netdev_info(netdev, "Changing Tx descriptor count from %d to %d\n",
vsi->tx_rings[0]->count, new_tx_cnt);
tx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_txq,
sizeof(struct ice_ring), GFP_KERNEL);
if (!tx_rings) {
err = -ENOMEM;
goto done;
}
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
for (i = 0; i < vsi->alloc_txq; i++) {
/* clone ring and setup updated count */
tx_rings[i] = *vsi->tx_rings[i];
tx_rings[i].count = new_tx_cnt;
tx_rings[i].desc = NULL;
tx_rings[i].tx_buf = NULL;
err = ice_setup_tx_ring(&tx_rings[i]);
if (err) {
while (i) {
i--;
ice_clean_tx_ring(&tx_rings[i]);
}
devm_kfree(&pf->pdev->dev, tx_rings);
goto done;
}
}
process_rx:
if (new_rx_cnt == vsi->rx_rings[0]->count)
goto process_link;
/* alloc updated Rx resources */
netdev_info(netdev, "Changing Rx descriptor count from %d to %d\n",
vsi->rx_rings[0]->count, new_rx_cnt);
rx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_rxq,
sizeof(struct ice_ring), GFP_KERNEL);
if (!rx_rings) {
err = -ENOMEM;
goto done;
}
ice: Report stats for allocated queues via ethtool stats It is not safe to have the string table for statistics change order or size over the lifetime of a given netdevice. This is because of the nature of the 3-step process for obtaining stats. First, user space performs a request for the size of the strings table. Second it performs a separate request for the strings themselves, after allocating space for the table. Third, it requests the stats themselves, also allocating space for the table. If the size decreased, there is potential to see garbage data or stats values. In the worst case, we could potentially see stats values become mis-aligned with their strings, so that it looks like a statistic is being reported differently than it actually is. Even worse, if the size increased, there is potential that the strings table or stats table was not allocated large enough and the stats code could access and write to memory it should not, potentially resulting in undefined behavior and system crashes. It isn't even safe if the size always changes under the RTNL lock. This is because the calls take place over multiple user space commands, so it is not possible to hold the RTNL lock for the entire duration of obtaining strings and stats. Further, not all consumers of the ethtool API are the user space ethtool program, and it is possible that one assumes the strings will not change (valid under the current contract), and thus only requests the stats values when requesting stats in a loop. Finally, it's not possible in the general case to detect when the size changes, because it is quite possible that one value which could impact the stat size increased, while another decreased. This would result in the same total number of stats, but reordering them so that stats no longer line up with the strings they belong to. Since only size changes aren't enough, we would need some sort of hash or token to determine when the strings no longer match. This would require extending the ethtool stats commands, but there is no more space in the relevant structures. The real solution to resolve this would be to add a completely new API for stats, probably over netlink. In the ice driver, the only thing impacting the stats that is not constant is the number of queues. Instead of reporting stats for each used queue, report stats for each allocated queue. We do not change the number of queues allocated for a given netdevice, as we pass this into the alloc_etherdev_mq() function to set the num_tx_queues and num_rx_queues. This resolves the potential bugs at the slight cost of displaying many queue statistics which will not be activated. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com> Tested-by: Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-08-09 13:28:54 +00:00
for (i = 0; i < vsi->alloc_rxq; i++) {
/* clone ring and setup updated count */
rx_rings[i] = *vsi->rx_rings[i];
rx_rings[i].count = new_rx_cnt;
rx_rings[i].desc = NULL;
rx_rings[i].rx_buf = NULL;
/* this is to allow wr32 to have something to write to
* during early allocation of Rx buffers
*/
rx_rings[i].tail = vsi->back->hw.hw_addr + PRTGEN_STATUS;
err = ice_setup_rx_ring(&rx_rings[i]);
if (err)
goto rx_unwind;
/* allocate Rx buffers */
err = ice_alloc_rx_bufs(&rx_rings[i],
ICE_DESC_UNUSED(&rx_rings[i]));
rx_unwind:
if (err) {
while (i) {
i--;
ice_free_rx_ring(&rx_rings[i]);
}
devm_kfree(&pf->pdev->dev, rx_rings);
err = -ENOMEM;
goto free_tx;
}
}
process_link:
/* Bring interface down, copy in the new ring info, then restore the
* interface. if VSI is up, bring it down and then back up
*/
if (!test_and_set_bit(__ICE_DOWN, vsi->state)) {
ice_down(vsi);
if (tx_rings) {
for (i = 0; i < vsi->alloc_txq; i++) {
ice_free_tx_ring(vsi->tx_rings[i]);
*vsi->tx_rings[i] = tx_rings[i];
}
devm_kfree(&pf->pdev->dev, tx_rings);
}
if (rx_rings) {
for (i = 0; i < vsi->alloc_rxq; i++) {
ice_free_rx_ring(vsi->rx_rings[i]);
/* copy the real tail offset */
rx_rings[i].tail = vsi->rx_rings[i]->tail;
/* this is to fake out the allocation routine
* into thinking it has to realloc everything
* but the recycling logic will let us re-use
* the buffers allocated above
*/
rx_rings[i].next_to_use = 0;
rx_rings[i].next_to_clean = 0;
rx_rings[i].next_to_alloc = 0;
*vsi->rx_rings[i] = rx_rings[i];
}
devm_kfree(&pf->pdev->dev, rx_rings);
}
ice_up(vsi);
}
goto done;
free_tx:
/* error cleanup if the Rx allocations failed after getting Tx */
if (tx_rings) {
for (i = 0; i < vsi->alloc_txq; i++)
ice_free_tx_ring(&tx_rings[i]);
devm_kfree(&pf->pdev->dev, tx_rings);
}
done:
clear_bit(__ICE_CFG_BUSY, pf->state);
return err;
}
static int ice_nway_reset(struct net_device *netdev)
{
/* restart autonegotiation */
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_link_status *hw_link_info;
struct ice_vsi *vsi = np->vsi;
struct ice_port_info *pi;
enum ice_status status;
bool link_up;
pi = vsi->port_info;
hw_link_info = &pi->phy.link_info;
link_up = hw_link_info->link_info & ICE_AQ_LINK_UP;
status = ice_aq_set_link_restart_an(pi, link_up, NULL);
if (status) {
netdev_info(netdev, "link restart failed, err %d aq_err %d\n",
status, pi->hw->adminq.sq_last_status);
return -EIO;
}
return 0;
}
/**
* ice_get_pauseparam - Get Flow Control status
* @netdev: network interface device structure
* @pause: ethernet pause (flow control) parameters
*/
static void
ice_get_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_port_info *pi;
pi = np->vsi->port_info;
pause->autoneg =
((pi->phy.link_info.an_info & ICE_AQ_AN_COMPLETED) ?
AUTONEG_ENABLE : AUTONEG_DISABLE);
if (pi->fc.current_mode == ICE_FC_RX_PAUSE) {
pause->rx_pause = 1;
} else if (pi->fc.current_mode == ICE_FC_TX_PAUSE) {
pause->tx_pause = 1;
} else if (pi->fc.current_mode == ICE_FC_FULL) {
pause->rx_pause = 1;
pause->tx_pause = 1;
}
}
/**
* ice_set_pauseparam - Set Flow Control parameter
* @netdev: network interface device structure
* @pause: return tx/rx flow control status
*/
static int
ice_set_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_link_status *hw_link_info;
struct ice_pf *pf = np->vsi->back;
struct ice_vsi *vsi = np->vsi;
struct ice_hw *hw = &pf->hw;
struct ice_port_info *pi;
enum ice_status status;
u8 aq_failures;
bool link_up;
int err = 0;
pi = vsi->port_info;
hw_link_info = &pi->phy.link_info;
link_up = hw_link_info->link_info & ICE_AQ_LINK_UP;
/* Changing the port's flow control is not supported if this isn't the
* PF VSI
*/
if (vsi->type != ICE_VSI_PF) {
netdev_info(netdev, "Changing flow control parameters only supported for PF VSI\n");
return -EOPNOTSUPP;
}
if (pause->autoneg != (hw_link_info->an_info & ICE_AQ_AN_COMPLETED)) {
netdev_info(netdev, "To change autoneg please use: ethtool -s <dev> autoneg <on|off>\n");
return -EOPNOTSUPP;
}
/* If we have link and don't have autoneg */
if (!test_bit(__ICE_DOWN, pf->state) &&
!(hw_link_info->an_info & ICE_AQ_AN_COMPLETED)) {
/* Send message that it might not necessarily work*/
netdev_info(netdev, "Autoneg did not complete so changing settings may not result in an actual change.\n");
}
if (pause->rx_pause && pause->tx_pause)
pi->fc.req_mode = ICE_FC_FULL;
else if (pause->rx_pause && !pause->tx_pause)
pi->fc.req_mode = ICE_FC_RX_PAUSE;
else if (!pause->rx_pause && pause->tx_pause)
pi->fc.req_mode = ICE_FC_TX_PAUSE;
else if (!pause->rx_pause && !pause->tx_pause)
pi->fc.req_mode = ICE_FC_NONE;
else
return -EINVAL;
/* Tell the OS link is going down, the link will go back up when fw
* says it is ready asynchronously
*/
ice_print_link_msg(vsi, false);
netif_carrier_off(netdev);
netif_tx_stop_all_queues(netdev);
/* Set the FC mode and only restart AN if link is up */
status = ice_set_fc(pi, &aq_failures, link_up);
if (aq_failures & ICE_SET_FC_AQ_FAIL_GET) {
netdev_info(netdev, "Set fc failed on the get_phy_capabilities call with err %d aq_err %d\n",
status, hw->adminq.sq_last_status);
err = -EAGAIN;
} else if (aq_failures & ICE_SET_FC_AQ_FAIL_SET) {
netdev_info(netdev, "Set fc failed on the set_phy_config call with err %d aq_err %d\n",
status, hw->adminq.sq_last_status);
err = -EAGAIN;
} else if (aq_failures & ICE_SET_FC_AQ_FAIL_UPDATE) {
netdev_info(netdev, "Set fc failed on the get_link_info call with err %d aq_err %d\n",
status, hw->adminq.sq_last_status);
err = -EAGAIN;
}
if (!test_bit(__ICE_DOWN, pf->state)) {
/* Give it a little more time to try to come back */
msleep(75);
if (!test_bit(__ICE_DOWN, pf->state))
return ice_nway_reset(netdev);
}
return err;
}
/**
* ice_get_rxfh_key_size - get the RSS hash key size
* @netdev: network interface device structure
*
* Returns the table size.
*/
static u32 ice_get_rxfh_key_size(struct net_device __always_unused *netdev)
{
return ICE_VSIQF_HKEY_ARRAY_SIZE;
}
/**
* ice_get_rxfh_indir_size - get the rx flow hash indirection table size
* @netdev: network interface device structure
*
* Returns the table size.
*/
static u32 ice_get_rxfh_indir_size(struct net_device *netdev)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
return np->vsi->rss_table_size;
}
/**
* ice_get_rxfh - get the rx flow hash indirection table
* @netdev: network interface device structure
* @indir: indirection table
* @key: hash key
* @hfunc: hash function
*
* Reads the indirection table directly from the hardware.
*/
static int
ice_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key, u8 *hfunc)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
struct ice_pf *pf = vsi->back;
int ret = 0, i;
u8 *lut;
if (hfunc)
*hfunc = ETH_RSS_HASH_TOP;
if (!indir)
return 0;
if (!test_bit(ICE_FLAG_RSS_ENA, pf->flags)) {
/* RSS not supported return error here */
netdev_warn(netdev, "RSS is not configured on this VSI!\n");
return -EIO;
}
lut = devm_kzalloc(&pf->pdev->dev, vsi->rss_table_size, GFP_KERNEL);
if (!lut)
return -ENOMEM;
if (ice_get_rss(vsi, key, lut, vsi->rss_table_size)) {
ret = -EIO;
goto out;
}
for (i = 0; i < vsi->rss_table_size; i++)
indir[i] = (u32)(lut[i]);
out:
devm_kfree(&pf->pdev->dev, lut);
return ret;
}
/**
* ice_set_rxfh - set the rx flow hash indirection table
* @netdev: network interface device structure
* @indir: indirection table
* @key: hash key
* @hfunc: hash function
*
* Returns -EINVAL if the table specifies an invalid queue id, otherwise
* returns 0 after programming the table.
*/
static int ice_set_rxfh(struct net_device *netdev, const u32 *indir,
const u8 *key, const u8 hfunc)
{
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_vsi *vsi = np->vsi;
struct ice_pf *pf = vsi->back;
u8 *seed = NULL;
if (hfunc != ETH_RSS_HASH_NO_CHANGE && hfunc != ETH_RSS_HASH_TOP)
return -EOPNOTSUPP;
if (!test_bit(ICE_FLAG_RSS_ENA, pf->flags)) {
/* RSS not supported return error here */
netdev_warn(netdev, "RSS is not configured on this VSI!\n");
return -EIO;
}
if (key) {
if (!vsi->rss_hkey_user) {
vsi->rss_hkey_user =
devm_kzalloc(&pf->pdev->dev,
ICE_VSIQF_HKEY_ARRAY_SIZE,
GFP_KERNEL);
if (!vsi->rss_hkey_user)
return -ENOMEM;
}
memcpy(vsi->rss_hkey_user, key, ICE_VSIQF_HKEY_ARRAY_SIZE);
seed = vsi->rss_hkey_user;
}
if (!vsi->rss_lut_user) {
vsi->rss_lut_user = devm_kzalloc(&pf->pdev->dev,
vsi->rss_table_size,
GFP_KERNEL);
if (!vsi->rss_lut_user)
return -ENOMEM;
}
/* Each 32 bits pointed by 'indir' is stored with a lut entry */
if (indir) {
int i;
for (i = 0; i < vsi->rss_table_size; i++)
vsi->rss_lut_user[i] = (u8)(indir[i]);
} else {
ice_fill_rss_lut(vsi->rss_lut_user, vsi->rss_table_size,
vsi->rss_size);
}
if (ice_set_rss(vsi, seed, vsi->rss_lut_user, vsi->rss_table_size))
return -EIO;
return 0;
}
static const struct ethtool_ops ice_ethtool_ops = {
.get_link_ksettings = ice_get_link_ksettings,
.set_link_ksettings = ice_set_link_ksettings,
.get_drvinfo = ice_get_drvinfo,
.get_regs_len = ice_get_regs_len,
.get_regs = ice_get_regs,
.get_msglevel = ice_get_msglevel,
.set_msglevel = ice_set_msglevel,
.get_link = ethtool_op_get_link,
.get_strings = ice_get_strings,
.get_ethtool_stats = ice_get_ethtool_stats,
.get_sset_count = ice_get_sset_count,
.get_rxnfc = ice_get_rxnfc,
.get_ringparam = ice_get_ringparam,
.set_ringparam = ice_set_ringparam,
.nway_reset = ice_nway_reset,
.get_pauseparam = ice_get_pauseparam,
.set_pauseparam = ice_set_pauseparam,
.get_rxfh_key_size = ice_get_rxfh_key_size,
.get_rxfh_indir_size = ice_get_rxfh_indir_size,
.get_rxfh = ice_get_rxfh,
.set_rxfh = ice_set_rxfh,
};
/**
* ice_set_ethtool_ops - setup netdev ethtool ops
* @netdev: network interface device structure
*
* setup netdev ethtool ops with ice specific ops
*/
void ice_set_ethtool_ops(struct net_device *netdev)
{
netdev->ethtool_ops = &ice_ethtool_ops;
}