Skip to contentSkip to conversation

e1000e: Enable the GPT clock before sending message to the CSME

PATCH Superseded Sasha Neftin 3 participants 4 messages May 2, 2022 – May 3, 2022

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 e1000e: Enable the GPT clock before sending message to the CSME

Enable the dynamic GPT clock. The clock is always ticking is the guarantee CSME receive the H2ME message and exit from the DMoff state. This clock cleared upon HW initialization (D3 -> D0 transition).

Fixes 3e55d231716e ("e1000e: Add handshake with the CSME to support s0ix")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=214821

---
 drivers/net/ethernet/intel/e1000e/netdev.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/ethernet/intel/e1000e/netdev.c +4 -0
index fa06f68c8c80..e29a718469ee 100644
@@ -6494,6 +6494,10 @@ static void e1000e_s0ix_exit_flow(struct e1000_adapter *adapter)
 
 	if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID &&
 	    hw->mac.type >= e1000_pch_adp) {
+		/* Keep the gpt_clk_enable_d clock for CSME*/
+		mac_data = er32(FEXTNVM);
+		mac_data |= BIT(3);
+		ew32(FEXTNVM, mac_data);
 		/* Request ME unconfigure the device from S0ix */
 		mac_data = er32(H2ME);
 		mac_data &= ~E1000_H2ME_START_DPG;
Signature · 1 line
2.30.2
#1 lore raw

PATCH 1/1 e1000e: Enable the GPT clock before sending message to the CSME

Dear Sasha,

Thank you for your patch.

Am 02.05.22 um 15:15 schrieb Sasha Neftin:

You could remove the articles *the* from the summary to make it shorter.

Quoted 3 lines

Enable the dynamic GPT clock. The clock is always ticking is the guarantee CSME receive the H2ME message and exit from the DMoff state. This clock cleared upon HW initialization (D3 -> D0 transition).

Please do not break the line, just because a sentence ends.

*is* cleared?

Please start the commit message by describing the problem, and also give instructions how to reproduce it.

Quoted 12 lines
Fixes 3e55d231716e ("e1000e: Add handshake with the CSME to support s0ix")

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=214821

---
  drivers/net/ethernet/intel/e1000e/netdev.c | 4 ++++
  1 file changed, 4 insertions(+)
drivers/net/ethernet/intel/e1000e/netdev.c +0 -0
index fa06f68c8c80..e29a718469ee 100644

drivers/net/ethernet/intel/e1000e/netdev.c static void e1000e_s0ix_exit_flow(struct e1000_adapter *adapter) · line 6497

Quoted 4 lines
drivers/net/ethernet/intel/e1000e/netdev.c +1 -0
@@ -6494,6 +6494,10 @@ static void e1000e_s0ix_exit_flow(struct e1000_adapter *adapter)
 
  	if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID &&
  	    hw->mac.type >= e1000_pch_adp) {
+		/* Keep the gpt_clk_enable_d clock for CSME*/

Missing space before the closing */.

Why is gpt_clk_enable_d spelled that way?

drivers/net/ethernet/intel/e1000e/netdev.c static void e1000e_s0ix_exit_flow(struct e1000_adapter *adapter) · line 6503

Quoted 6 lines
drivers/net/ethernet/intel/e1000e/netdev.c +3 -0
+		mac_data = er32(FEXTNVM);
+		mac_data |= BIT(3);
+		ew32(FEXTNVM, mac_data);
  		/* Request ME unconfigure the device from S0ix */
  		mac_data = er32(H2ME);
  		mac_data &= ~E1000_H2ME_START_DPG;

Kind regards,

Paul

#2 lore raw

PATCH 1/1 e1000e: Enable the GPT clock before sending message to the CSME

Quoted 9 lines · Paul Menzel

Dear Sasha,

Thank you for your patch.

Am 02.05.22 um 15:15 schrieb Sasha Neftin:

You could remove the articles *the* from the summary to make it shorter.

Quoted 8 lines
Quoted 4 lines

Enable the dynamic GPT clock. The clock is always ticking is the guarantee
CSME receive the H2ME message and exit from the DMoff state. This clock cleared upon HW initialization (D3 -> D0 transition).

Please do not break the line, just because a sentence ends.

*is* cleared?

ok.

Quoted 3 lines

Please start the commit message by describing the problem, and also give instructions how to reproduce it.

please, refer to the link below (bugzilla)

Quoted 9 lines
Quoted 9 lines
Fixes 3e55d231716e ("e1000e: Add handshake with the CSME to support

s0ix")
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=214821

---
? drivers/net/ethernet/intel/e1000e/netdev.c | 4 ++++ ? 1 file changed, 4 insertions(+)

Quoted 12 lines
Quoted 10 lines
drivers/net/ethernet/intel/e1000e/netdev.c +0 -0
index fa06f68c8c80..e29a718469ee 100644
@@ -6494,6 +6494,10 @@ static void e1000e_s0ix_exit_flow(struct 
 e1000_adapter *adapter)

????? if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID && ????????? hw->mac.type >= e1000_pch_adp) { +??????? /* Keep the gpt_clk_enable_d clock for CSME*/

Missing space before the closing */.

Thanks - will fix in v2

Quoted 2 lines

Why is gpt_clk_enable_d spelled that way?

I took it from HW design. I will spell it as 'GPT clock' (more clearly).

Quoted 9 lines
Quoted 7 lines

+??????? mac_data = er32(FEXTNVM);
+??????? mac_data |= BIT(3);
+??????? ew32(FEXTNVM, mac_data);
????????? /* Request ME unconfigure the device from S0ix */ ????????? mac_data = er32(H2ME);
????????? mac_data &= ~E1000_H2ME_START_DPG;

Quoted 3 lines

Kind regards,

Paul

Sasha

#3 lore raw

PATCH 1/1 e1000e: Enable the GPT clock before sending message to the CSME

Dear Sasha,

Am 03.05.22 um 05:14 schrieb Neftin, Sasha:

Quoted 5 lines · Paul Menzel
Quoted 4 lines

Am 02.05.22 um 15:15 schrieb Sasha Neftin:

You could remove the articles *the* from the summary to make it shorter.

As you acked all the other comments, I just want to make sure, you saw this one.

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines

Enable the dynamic GPT clock. The clock is always ticking is the guarantee
CSME receive the H2ME message and exit from the DMoff state. This clock cleared upon HW initialization (D3 -> D0 transition).

Quoted 8 lines
Quoted 3 lines

Please do not break the line, just because a sentence ends.

*is* cleared?

ok.

Quoted 3 lines

Please start the commit message by describing the problem, and also give instructions how to reproduce it.

please, refer to the link below (bugzilla)

I saw that but no, as always commit messages need to contain the motivation and also need to be some kind of self-contained. Reviewers can *not* be expected to read through several comments in a bug report, and people looking at a commit are not always online, and Web service are not always online. So, a summary of the issue should be provided.

Quoted 1 line
Quoted 1 line
Quoted 1 line
Fixes 3e55d231716e ("e1000e: Add handshake with the CSME to support s0ix")

The bug report also does not mention anything about a regression. Did it never work?

Quoted 18 lines
Quoted 6 lines
Quoted 6 lines

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=214821

---
? drivers/net/ethernet/intel/e1000e/netdev.c | 4 ++++ ? 1 file changed, 4 insertions(+)

Quoted 12 lines
Quoted 10 lines
drivers/net/ethernet/intel/e1000e/netdev.c +0 -0
index fa06f68c8c80..e29a718469ee 100644
@@ -6494,6 +6494,10 @@ static void e1000e_s0ix_exit_flow(struct 
 e1000_adapter *adapter)

????? if (er32(FWSM) & E1000_ICH_FWSM_FW_VALID && ????????? hw->mac.type >= e1000_pch_adp) { +??????? /* Keep the gpt_clk_enable_d clock for CSME*/

Missing space before the closing */.

Quoted 11 lines

Thanks - will fix in v2

Quoted 2 lines

Why is gpt_clk_enable_d spelled that way?

I took it from HW design. I will spell it as 'GPT clock' (more clearly).

Quoted 7 lines
Quoted 7 lines

+??????? mac_data = er32(FEXTNVM);
+??????? mac_data |= BIT(3);
+??????? ew32(FEXTNVM, mac_data);
????????? /* Request ME unconfigure the device from S0ix */ ????????? mac_data = er32(H2ME);
????????? mac_data &= ~E1000_H2ME_START_DPG;

Kind regards,

Paul