* [PATCH iwl-net 0/2] e1000e: Remove unnecessary workarounds from Nova Lake platform
@ 2026-11-18 16:34 Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 1/2] e1000e: introduce new board type for Nova Lake PCH Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake Dima Ruinskiy
0 siblings, 2 replies; 4+ messages in thread
From: Dima Ruinskiy @ 2026-11-18 16:34 UTC (permalink / raw)
To: intel-wired-lan; +Cc: dima.ruinskiy, anthony.l.nguyen
Introduce a dedicated Nova Lake board type and associate its device IDs
with the correct MAC type. Then restrict the K1 workaround to Meteor Lake
through Panther Lake, excluding Nova Lake where it is unnecessary and can
degrade performance.
Dima Ruinskiy (2):
e1000e: introduce new board type for Nova Lake PCH
e1000e: remove K1 workaround from Nova Lake
drivers/net/ethernet/intel/e1000e/e1000.h | 4 +++-
drivers/net/ethernet/intel/e1000e/hw.h | 4 ++--
drivers/net/ethernet/intel/e1000e/ich8lan.c | 22 ++++++++++++++++++++-
drivers/net/ethernet/intel/e1000e/netdev.c | 17 ++++++++--------
4 files changed, 35 insertions(+), 12 deletions(-)
--
2.44.0
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH iwl-net 1/2] e1000e: introduce new board type for Nova Lake PCH
2026-11-18 16:34 [PATCH iwl-net 0/2] e1000e: Remove unnecessary workarounds from Nova Lake platform Dima Ruinskiy
@ 2026-11-18 16:34 ` Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake Dima Ruinskiy
1 sibling, 0 replies; 4+ messages in thread
From: Dima Ruinskiy @ 2026-11-18 16:34 UTC (permalink / raw)
To: intel-wired-lan; +Cc: dima.ruinskiy, anthony.l.nguyen
Commit 1fe4f45ea461 ("e1000e: Add support for the next LOM generation")
introduced the e1000_pch_nvp MAC type and related device IDs. However,
without a new board type, the device IDs were not associated with the
correct MAC type.
Fix this by introducing a new board type, to distinguish between Nova Lake
and previous generations of LAN devices. This allows limiting specific
configuration flows to their respective device families. For example -
the PHY access retry workaround, the PLL clock gate timeout change and the
forced 38400KHz system clock (SYSCFI).
Additionally, remap device IDs 0x5510 and 0x5511 to the new board type.
These IDs were originally reserved to Lunar Lake LAN, but never used in
the field and reallocated to Nova Lake platforms.
Fixes: 1fe4f45ea461 ("e1000e: Add support for the next LOM generation")
Signed-off-by: Dima Ruinskiy <dima.ruinskiy@intel.com>
---
drivers/net/ethernet/intel/e1000e/e1000.h | 4 +++-
drivers/net/ethernet/intel/e1000e/hw.h | 4 ++--
drivers/net/ethernet/intel/e1000e/ich8lan.c | 20 ++++++++++++++++++++
drivers/net/ethernet/intel/e1000e/netdev.c | 17 +++++++++--------
4 files changed, 34 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/intel/e1000e/e1000.h b/drivers/net/ethernet/intel/e1000e/e1000.h
index 704284188..b662ff2b7 100644
--- a/drivers/net/ethernet/intel/e1000e/e1000.h
+++ b/drivers/net/ethernet/intel/e1000e/e1000.h
@@ -118,7 +118,8 @@ enum e1000_boards {
board_pch_tgp,
board_pch_adp,
board_pch_mtp,
- board_pch_ptp
+ board_pch_ptp,
+ board_pch_nvp
};
struct e1000_ps_page {
@@ -530,6 +531,7 @@ extern const struct e1000_info e1000_pch_tgp_info;
extern const struct e1000_info e1000_pch_adp_info;
extern const struct e1000_info e1000_pch_mtp_info;
extern const struct e1000_info e1000_pch_ptp_info;
+extern const struct e1000_info e1000_pch_nvp_info;
extern const struct e1000_info e1000_es2_info;
void e1000e_ptp_init(struct e1000_adapter *adapter);
diff --git a/drivers/net/ethernet/intel/e1000e/hw.h b/drivers/net/ethernet/intel/e1000e/hw.h
index c7ac599e5..259b5fb8d 100644
--- a/drivers/net/ethernet/intel/e1000e/hw.h
+++ b/drivers/net/ethernet/intel/e1000e/hw.h
@@ -112,14 +112,14 @@ struct e1000_hw;
#define E1000_DEV_ID_PCH_ADP_I219_V19 0x550D
#define E1000_DEV_ID_PCH_LNP_I219_LM20 0x550E
#define E1000_DEV_ID_PCH_LNP_I219_V20 0x550F
-#define E1000_DEV_ID_PCH_LNP_I219_LM21 0x5510
-#define E1000_DEV_ID_PCH_LNP_I219_V21 0x5511
#define E1000_DEV_ID_PCH_ARL_I219_LM24 0x57A0
#define E1000_DEV_ID_PCH_ARL_I219_V24 0x57A1
#define E1000_DEV_ID_PCH_PTP_I219_LM25 0x57B3
#define E1000_DEV_ID_PCH_PTP_I219_V25 0x57B4
#define E1000_DEV_ID_PCH_PTP_I219_LM27 0x57B7
#define E1000_DEV_ID_PCH_PTP_I219_V27 0x57B8
+#define E1000_DEV_ID_PCH_NVL_I219_LM28 0x5510
+#define E1000_DEV_ID_PCH_NVL_I219_V28 0x5511
#define E1000_DEV_ID_PCH_NVL_I219_LM29 0x57B9
#define E1000_DEV_ID_PCH_NVL_I219_V29 0x57BA
diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index 87cd1cdc6..a412475f5 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -6240,3 +6240,23 @@ const struct e1000_info e1000_pch_ptp_info = {
.phy_ops = &ich8_phy_ops,
.nvm_ops = &spt_nvm_ops,
};
+
+const struct e1000_info e1000_pch_nvp_info = {
+ .mac = e1000_pch_nvp,
+ .flags = FLAG_IS_ICH
+ | FLAG_HAS_WOL
+ | FLAG_HAS_HW_TIMESTAMP
+ | FLAG_HAS_CTRLEXT_ON_LOAD
+ | FLAG_HAS_AMT
+ | FLAG_HAS_FLASH
+ | FLAG_HAS_JUMBO_FRAMES
+ | FLAG_APME_IN_WUC,
+ .flags2 = FLAG2_HAS_PHY_STATS
+ | FLAG2_HAS_EEE,
+ .pba = 26,
+ .max_hw_frame_size = 9022,
+ .get_variants = e1000_get_variants_ich8lan,
+ .mac_ops = &ich8_mac_ops,
+ .phy_ops = &ich8_phy_ops,
+ .nvm_ops = &spt_nvm_ops,
+};
diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index c3e944ee4..13c26144f 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -57,6 +57,7 @@ static const struct e1000_info *e1000_info_tbl[] = {
[board_pch_adp] = &e1000_pch_adp_info,
[board_pch_mtp] = &e1000_pch_mtp_info,
[board_pch_ptp] = &e1000_pch_ptp_info,
+ [board_pch_nvp] = &e1000_pch_nvp_info,
};
static const struct dmi_system_id disable_k1_list[] = {
@@ -8196,12 +8197,6 @@ static const struct pci_device_id e1000_pci_tbl[] = {
}, {
PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_LNP_I219_V20),
.driver_data = board_pch_mtp,
- }, {
- PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_LNP_I219_LM21),
- .driver_data = board_pch_mtp,
- }, {
- PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_LNP_I219_V21),
- .driver_data = board_pch_mtp,
}, {
PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_ARL_I219_LM24),
.driver_data = board_pch_mtp,
@@ -8220,12 +8215,18 @@ static const struct pci_device_id e1000_pci_tbl[] = {
}, {
PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_PTP_I219_V27),
.driver_data = board_pch_ptp,
+ }, {
+ PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_NVL_I219_LM28),
+ .driver_data = board_pch_nvp,
+ }, {
+ PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_NVL_I219_V28),
+ .driver_data = board_pch_nvp,
}, {
PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_NVL_I219_LM29),
- .driver_data = board_pch_ptp,
+ .driver_data = board_pch_nvp,
}, {
PCI_VDEVICE(INTEL, E1000_DEV_ID_PCH_NVL_I219_V29),
- .driver_data = board_pch_ptp
+ .driver_data = board_pch_nvp
},
{ } /* terminate list */
--
2.44.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake
2026-11-18 16:34 [PATCH iwl-net 0/2] e1000e: Remove unnecessary workarounds from Nova Lake platform Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 1/2] e1000e: introduce new board type for Nova Lake PCH Dima Ruinskiy
@ 2026-11-18 16:34 ` Dima Ruinskiy
2026-11-18 16:51 ` sashiko-bot
1 sibling, 1 reply; 4+ messages in thread
From: Dima Ruinskiy @ 2026-11-18 16:34 UTC (permalink / raw)
To: intel-wired-lan; +Cc: dima.ruinskiy, anthony.l.nguyen, Michael Edri
Commit 3c7bf5af2196 ("e1000e: Introduce private flag to disable K1")
implemented a workaround for packet loss issues when the K1 low-power link
state is enabled on the integrated LAN on several generations of platforms,
between Meteor Lake and Panther Lake. On Nova Lake platforms, it is not
needed, and may cause noticeable performance degradations.
Now that board_pch_nvp and e1000_pch_nvp are distinct types, add another
MAC type check in e1000_reconfigure_k1_params() so that it never applies
to MAC types before Meteor Lake or after Panther Lake.
Tested on an I219-V29 (8086:57ba), K1 enabled, 1 Gb/s link. For each test,
first line is the result without the patch, second line is with the patch.
iperf3 TCP receive, 20 s runs:
16-50 Mbps, 72-208 rx_crc_errors per run
949 Mbps, 0 rx_crc_errors (5 of 5 runs)
UDP 120 Mbps for 120 s, 2 runs:
0.55-0.67 % lost, 1855-2651 rx_crc_errors
0 % lost, 0 rx_crc_errors
HTTPS download of a 4.3 GB and a 3.5 GB file:
5.6 and 10.3 MB/s
74 and 59 MB/s
Fixes: 3c7bf5af2196 ("e1000e: Introduce private flag to disable K1")
Signed-off-by: Dima Ruinskiy <dima.ruinskiy@intel.com>
Co-developed-by: Michael Edri <michael.edri@intel.com>
Signed-off-by: Michael Edri <michael.edri@intel.com>
---
drivers/net/ethernet/intel/e1000e/ich8lan.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index a412475f5..f5f3867ad 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -301,7 +301,7 @@ static s32 e1000_reconfigure_k1_params(struct e1000_hw *hw)
u32 fextnvm12;
s32 ret_val;
- if (hw->mac.type < e1000_pch_mtp) {
+ if (hw->mac.type < e1000_pch_mtp || hw->mac.type > e1000_pch_ptp) {
if (hw->adapter->flags2 & FLAG2_DISABLE_K1)
return e1000_configure_k1_ich8lan(hw, false);
return 0;
--
2.44.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake
2026-11-18 16:34 ` [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake Dima Ruinskiy
@ 2026-11-18 16:51 ` sashiko-bot
0 siblings, 0 replies; 4+ messages in thread
From: sashiko-bot @ 2026-11-18 16:51 UTC (permalink / raw)
To: Dima Ruinskiy, intel-wired-lan
Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] e1000e: inaccurate Fixes tag and commit attribution for K1
workaround
--
commit 5c9c29651f27bff328cda7813a58dbb82ff6d70b
Author: Dima Ruinskiy <dima.ruinskiy@intel.com>
> e1000e: remove K1 workaround from Nova Lake
>
> Commit 3c7bf5af2196 ("e1000e: Introduce private flag to disable K1")
> implemented a workaround for packet loss issues when the K1 low-power link
> state is enabled on the integrated LAN on several generations of platforms,
> between Meteor Lake and Panther Lake. On Nova Lake platforms, it is not
> needed, and may cause noticeable performance degradations.
>
> Now that board_pch_nvp and e1000_pch_nvp are distinct types, add another
> MAC type check in e1000_reconfigure_k1_params() so that it never applies
> to MAC types before Meteor Lake or after Panther Lake.
>
> Tested on an I219-V29 (8086:57ba), K1 enabled, 1 Gb/s link. For each test,
> first line is the result without the patch, second line is with the patch.
>
> iperf3 TCP receive, 20 s runs:
> 16-50 Mbps, 72-208 rx_crc_errors per run
> 949 Mbps, 0 rx_crc_errors (5 of 5 runs)
> UDP 120 Mbps for 120 s, 2 runs:
> 0.55-0.67 % lost, 1855-2651 rx_crc_errors
> 0 % lost, 0 rx_crc_errors
> HTTPS download of a 4.3 GB and a 3.5 GB file:
> 5.6 and 10.3 MB/s
> 74 and 59 MB/s
>
> Fixes: 3c7bf5af2196 ("e1000e: Introduce private flag to disable K1")
> Signed-off-by: Dima Ruinskiy <dima.ruinskiy@intel.com>
> Co-developed-by: Michael Edri <michael.edri@intel.com>
> Signed-off-by: Michael Edri <michael.edri@intel.com>
[Severity: Low]
Should the commit description and Fixes tag reference efaaf344bc29 ("e1000e:
change k1 configuration on MTP and later platforms") instead of 3c7bf5af2196
("e1000e: Introduce private flag to disable K1")?
In e1000_reconfigure_k1_params(), the early return for hw->mac.type >
e1000_pch_ptp still honors FLAG2_DISABLE_K1 (added in 3c7bf5af2196 ("e1000e:
Introduce private flag to disable K1")), while skipping the FEXTNVM12 and
I217_PHY_TIMEOUTS_REG K1 parameter workaround introduced by efaaf344bc29
("e1000e: change k1 configuration on MTP and later platforms").
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261118163447.1117617-1-dima.ruinskiy@intel.com?part=2
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-11-18 16:52 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 16:34 [PATCH iwl-net 0/2] e1000e: Remove unnecessary workarounds from Nova Lake platform Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 1/2] e1000e: introduce new board type for Nova Lake PCH Dima Ruinskiy
2026-11-18 16:34 ` [PATCH iwl-net 2/2] e1000e: remove K1 workaround from Nova Lake Dima Ruinskiy
2026-11-18 16:51 ` sashiko-bot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox