Skip to contentSkip to conversation

ACPI IRQ storm with 6.10

DISCUSSION Jiri Slaby 5 participants 21 messages Aug 14, 2024 – Aug 22, 2024

Subsystems (4)
lore
Share & tools
lore

Jiri Slaby

author
#0 lore raw

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):
9: 20220768 ... IR-IO-APIC 9-fasteoi acpi

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 84 lines

event-0188 ev_fixed_event_detect : Fixed Event Block: Enable 00000020 Status 00000000

evgpe-0673 ev_detect_gpe         : Read registers for GPE 00: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 01: Status=00, Enable=02, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 02: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 03: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 04: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 05: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 06: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 07: Status=00, Enable=00, RunEnable=02, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 08-0F: RunEnable=00, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 10: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 11: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 12: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 13: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 14: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 15: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 16: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 17: Status=00, Enable=80, RunEnable=80, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 18-1F: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 20-27: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 28-2F: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 30-37: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 38-3F: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 40-47: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 48-4F: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 50-57: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 58-5F: RunEnable=00, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 60: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 61: Status=00, Enable=02, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 62: Status=00, Enable=04, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 63: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 64: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 65: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 66: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 67: Status=00, Enable=00, RunEnable=06, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 68: Status=00, Enable=00, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 69: Status=00, Enable=02, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6A: Status=00, Enable=00, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6B: Status=00, Enable=08, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6C: Status=00, Enable=00, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6D: Status=20, Enable=00, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6E: Status=40, Enable=40, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 6F: Status=00, Enable=00, RunEnable=4A, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 70: Status=00, Enable=00, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 71: Status=00, Enable=02, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 72: Status=00, Enable=04, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 73: Status=00, Enable=08, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 74: Status=00, Enable=00, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 75: Status=00, Enable=00, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 76: Status=00, Enable=00, RunEnable=0E, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 77: Status=00, Enable=00, RunEnable=0E, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 78-7F: RunEnable=00, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 80: Status=00, Enable=01, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 81: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 82: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 83: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 84: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 85: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 86: Status=40, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 87: Status=00, Enable=00, RunEnable=01, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 88: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 89: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8A: Status=00, Enable=04, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8B: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8C: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8D: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8E: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE 8F: Status=00, Enable=00, RunEnable=04, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 90-97: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE 98-9F: RunEnable=00, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A0: Status=00, Enable=01, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A1: Status=00, Enable=02, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A2: Status=00, Enable=04, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A3: Status=00, Enable=00, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A4: Status=00, Enable=00, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A5: Status=00, Enable=00, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A6: Status=00, Enable=00, RunEnable=87, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE A7: Status=00, Enable=80, RunEnable=87, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE A8-AF: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE B0-B7: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE B8-BF: RunEnable=00, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C0: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C1: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C2: Status=00, Enable=00, RunEnable=80, WakeEnable=00
Quoted 8 lines
evgpe-0673 ev_detect_gpe         : Read registers for GPE C3: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C4: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C5: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C6: Status=00, Enable=00, RunEnable=80, WakeEnable=00
evgpe-0673 ev_detect_gpe         : Read registers for GPE C7: Status=00, Enable=80, RunEnable=80, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE C8-CF: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE D0-D7: RunEnable=00, WakeEnable=00
evgpe-0396 ev_gpe_detect         : Ignore disabled registers for GPE D8-DF: RunEnable=00, WakeEnable=00

Longer (yet still trimmed due to pace of the messages) log:
https://bugzilla.suse.com/attachment.cgi?id=876664

Any ideas?

thanks,

Signature · 2 lines
js
suse labs
author
#1 lore raw
Quoted 5 lines · Jiri Slaby

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):
   9:   20220768          ...  IR-IO-APIC    9-fasteoi   acpi

Quoted 7 lines

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 1 line

evgpe-0673 ev_detect_gpe         : Read registers for GPE 6D: Status=20, Enable=00, RunEnable=4A, WakeEnable=00

0x6d seems to count excessively (10 snapshots every 1 second):

Quoted 11 lines
/sys/firmware/acpi/interrupts/gpe6D:   82066  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   86536  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   90990     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   95468  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  100282  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  105187     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  110014     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  114852     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  119682     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  124194     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  128641  EN STS enabled      unmasked

acpidump:
https://bugzilla.suse.com/attachment.cgi?id=876677

DSDT:
https://bugzilla.suse.com/attachment.cgi?id=876678

Quoted 3 lines

Any ideas?

thanks,--

js

#2 lore raw
Quoted 31 lines · Jiri Slaby
Quoted 6 lines · Jiri Slaby

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):

    9:   20220768          ...  IR-IO-APIC    9-fasteoi   acpi
Quoted 9 lines

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 2 lines

evgpe-0673 ev_detect_gpe : Read registers for GPE 6D:
Status=20, Enable=00, RunEnable=4A, WakeEnable=00

0x6d seems to count excessively (10 snapshots every 1 second):

Quoted 11 lines
/sys/firmware/acpi/interrupts/gpe6D:   82066  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   86536  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   90990     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   95468  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  100282  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  105187     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  110014     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  114852     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  119682     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  124194     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  128641  EN STS enabled      unmasked
Quoted 7 lines

acpidump:
https://bugzilla.suse.com/attachment.cgi?id=876677

DSDT:
https://bugzilla.suse.com/attachment.cgi?id=876678

Quoted 1 line

Any ideas?

GPE 6D is listed in _PRW for some devices, so maybe one of them continues to trigger wakeup events?

You can ask the reporter to mask that GPE via "echo mask > /sys/firmware/acpi/interrupts/gpe6D" and see if the storm goes away then.

The only ACPI core issue introduced between 6.9 and 6.10 I'm aware of is the one addressed by this series

https://lore.kernel.org/linux-acpi/[email protected]/

but this is about the EC and the problem here doesn't appear to be EC-related. It may be worth trying anyway, though.

#3 lore raw

Dne 16. 08. 24 v 20:29 Rafael J. Wysocki napsal(a):

Quoted 38 lines · Jiri Slaby
Quoted 37 lines · Jiri Slaby
Quoted 6 lines

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):

     9:   20220768          ...  IR-IO-APIC    9-fasteoi   acpi
Quoted 9 lines

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 2 lines

evgpe-0673 ev_detect_gpe : Read registers for GPE 6D:
Status=20, Enable=00, RunEnable=4A, WakeEnable=00

0x6d seems to count excessively (10 snapshots every 1 second):

Quoted 11 lines
/sys/firmware/acpi/interrupts/gpe6D:   82066  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   86536  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   90990     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   95468  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  100282  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  105187     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  110014     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  114852     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  119682     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  124194     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  128641  EN STS enabled      unmasked

acpidump:
https://bugzilla.suse.com/attachment.cgi?id=876677

DSDT:
https://bugzilla.suse.com/attachment.cgi?id=876678

Quoted 9 lines
Quoted 1 line
Quoted 1 line

Any ideas?

GPE 6D is listed in _PRW for some devices, so maybe one of them continues to trigger wakeup events?

You can ask the reporter to mask that GPE via "echo mask > /sys/firmware/acpi/interrupts/gpe6D" and see if the storm goes away then.

It works, thank you. High CPU usage by irq/9-acpi is immediately gone.

Quoted 8 lines

The only ACPI core issue introduced between 6.9 and 6.10 I'm aware of is the one addressed by this series

https://lore.kernel.org/linux-acpi/[email protected]/

but this is about the EC and the problem here doesn't appear to be EC-related. It may be worth trying anyway, though.

#4 lore raw
In reply to Rafael J. Wysocki

Dne 16. 08. 24 v 20:29 Rafael J. Wysocki napsal(a):

Quoted 35 lines · Jiri Slaby
Quoted 34 lines · Jiri Slaby
Quoted 6 lines

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):

     9:   20220768          ...  IR-IO-APIC    9-fasteoi   acpi
Quoted 9 lines

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 2 lines

evgpe-0673 ev_detect_gpe : Read registers for GPE 6D:
Status=20, Enable=00, RunEnable=4A, WakeEnable=00

0x6d seems to count excessively (10 snapshots every 1 second):

Quoted 11 lines
/sys/firmware/acpi/interrupts/gpe6D:   82066  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   86536  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   90990     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:   95468  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  100282  EN STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  105187     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  110014     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  114852     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  119682     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  124194     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  128641  EN STS enabled      unmasked

acpidump:
https://bugzilla.suse.com/attachment.cgi?id=876677

Quoted 8 lines
Quoted 4 lines

DSDT:
https://bugzilla.suse.com/attachment.cgi?id=876678

Quoted 1 line

Any ideas?

GPE 6D is listed in _PRW for some devices, so maybe one of them continues to trigger wakeup events?

Disabling powertop service (which calls /usr/sbin/powertop --auto-tune) solves problem completely. After some search I have found this is the cause:

# causes IRQ storm on 6.10.x
# kernel 6.9.9 is immune
echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

lspci | grep 1f.6
00:1f.6 Ethernet controller: Intel Corporation Device 550b (rev 20)

journalctl -b | grep 1f.6
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: [8086:550b] type 00 class 0x020000 conventional PCI endpoint
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: BAR 0 [mem 0x9c300000-0x9c31ffff]
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: PME# supported from D0 D3hot D3cold
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: Adding to iommu group 12
srp 17 19:44:19 e14 kernel: e1000e 0000:00:1f.6: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
srp 17 19:44:19 e14 kernel: e1000e 0000:00:1f.6 0000:00:1f.6 (uninitialized): registered PHC clock
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: (PCI Express:2.5GT/s:Width x1) fc:5c:ee:b0:13:74
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: Intel(R) PRO/1000 Network Connection
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: MAC: 16, PHY: 12, PBA No: FFFFFF-0FF
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 enp0s31f6: renamed from eth0
srp 17 19:44:24 e14 ModemManager[1434]: <info> [base-manager] couldn't check support for device '/sys/devices/pci0000:00/0000:00:1f.6': not supported by any plugin

Quoted 12 lines

You can ask the reporter to mask that GPE via "echo mask > /sys/firmware/acpi/interrupts/gpe6D" and see if the storm goes away then.

The only ACPI core issue introduced between 6.9 and 6.10 I'm aware of is the one addressed by this series

https://lore.kernel.org/linux-acpi/[email protected]/

but this is about the EC and the problem here doesn't appear to be EC-related. It may be worth trying anyway, though.

author
#5 lore raw

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Quoted 47 lines · Petr Valenta

Dne 16. 08. 24 v 20:29 Rafael J. Wysocki napsal(a):

Quoted 35 lines · Jiri Slaby
Quoted 34 lines · Jiri Slaby
Quoted 6 lines

Hi,

one openSUSE's user reported that with 6.10, he sees one CPU under an IRQ storm from ACPI (sci_interrupt):
     9:   20220768          ...  IR-IO-APIC    9-fasteoi   acpi

Quoted 9 lines

At:
https://bugzilla.suse.com/show_bug.cgi?id=1229085

6.9 was OK.

With acpi.debug_level=0x08000000 acpi.debug_layer=0xffffffff, there is a repeated load of:

Quoted 2 lines

evgpe-0673 ev_detect_gpe         : Read registers for GPE 6D:
Status=20, Enable=00, RunEnable=4A, WakeEnable=00

0x6d seems to count excessively (10 snapshots every 1 second):

Quoted 11 lines

/sys/firmware/acpi/interrupts/gpe6D:   82066  EN STS enabled unmasked
/sys/firmware/acpi/interrupts/gpe6D:   86536  EN STS enabled unmasked
/sys/firmware/acpi/interrupts/gpe6D:   90990     STS enabled unmasked
/sys/firmware/acpi/interrupts/gpe6D:   95468  EN STS enabled unmasked
/sys/firmware/acpi/interrupts/gpe6D:  100282  EN STS enabled unmasked

/sys/firmware/acpi/interrupts/gpe6D:  105187     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  110014     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  114852     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  119682     STS enabled      unmasked
/sys/firmware/acpi/interrupts/gpe6D:  124194     STS enabled      unmasked

/sys/firmware/acpi/interrupts/gpe6D:  128641  EN STS enabled unmasked

acpidump:
https://bugzilla.suse.com/attachment.cgi?id=876677

Quoted 8 lines
Quoted 4 lines

DSDT:
https://bugzilla.suse.com/attachment.cgi?id=876678

Quoted 1 line

Any ideas?

GPE 6D is listed in _PRW for some devices, so maybe one of them continues to trigger wakeup events?

Quoted 5 lines

Disabling powertop service (which calls /usr/sbin/powertop --auto-tune) solves problem completely. After some search I have found this is the cause:

# causes IRQ storm on 6.10.x
# kernel 6.9.9 is immune
echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control
$ git log --no-merges  --oneline v6.9..v6.10 drivers/net/ethernet/intel/e1000e/
76a0a3f9cc2f e1000e: fix force smbus during suspend flow
c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems
bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function
6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates
1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu()
b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops
75a3f93b5383 net: intel: implement modern PM ops declarations

The last two play with PM ^^. I cannot immediately see if the issue can be caused by any of those, though.

If there are no ideas, possibly giving revert of both a try?

Quoted 30 lines

lspci | grep 1f.6
00:1f.6 Ethernet controller: Intel Corporation Device 550b (rev 20)

journalctl -b | grep 1f.6
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: [8086:550b] type 00 class 0x020000 conventional PCI endpoint
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: BAR 0 [mem 0x9c300000-0x9c31ffff]
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: PME# supported from D0 D3hot D3cold
srp 17 19:44:17 e14 kernel: pci 0000:00:1f.6: Adding to iommu group 12
srp 17 19:44:19 e14 kernel: e1000e 0000:00:1f.6: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
srp 17 19:44:19 e14 kernel: e1000e 0000:00:1f.6 0000:00:1f.6 (uninitialized): registered PHC clock
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: (PCI Express:2.5GT/s:Width x1) fc:5c:ee:b0:13:74
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: Intel(R) PRO/1000 Network Connection
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 eth0: MAC: 16, PHY: 12, PBA No: FFFFFF-0FF
srp 17 19:44:20 e14 kernel: e1000e 0000:00:1f.6 enp0s31f6: renamed from eth0
srp 17 19:44:24 e14 ModemManager[1434]: <info>  [base-manager] couldn't check support for device '/sys/devices/pci0000:00/0000:00:1f.6': not supported by any plugin

Quoted 12 lines

You can ask the reporter to mask that GPE via "echo mask > /sys/firmware/acpi/interrupts/gpe6D" and see if the storm goes away then.

The only ACPI core issue introduced between 6.9 and 6.10 I'm aware of is the one addressed by this series

https://lore.kernel.org/linux-acpi/[email protected]/

but this is about the EC and the problem here doesn't appear to be EC-related.  It may be worth trying anyway, though.

Signature · 2 lines
js
suse labs
author
#6 lore raw
Quoted 1 line · Jiri Slaby

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device *net dev,

         * duplex is forced.
         */
        if (cmd->base.eth_tp_mdix_ctrl) {

- if (hw->phy.media_type != e1000_media_type_copper) {
- ret_val = -EOPNOTSUPP;
- goto out;
- }
+ if (hw->phy.media_type != e1000_media_type_copper)
+ return -EOPNOTSUPP;

                if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                    (cmd->base.autoneg != AUTONEG_ENABLE)) {
                        e_err("forcing MDI/MDI-X state is not supported 
when lin
k speed and/or duplex are forced\n");
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                }
        }

@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device *netde v,

        }

out:

-       pm_runtime_put_sync(netdev->dev.parent);
        clear_bit(__E1000_RESETTING, &adapter->state);
        return ret_val;

}
==========================================

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

thanks,

Signature · 2 lines
js
suse labs
#7 lore raw
Quoted 40 lines · Jiri Slaby
Quoted 1 line · Jiri Slaby

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
         * duplex is forced.
         */
        if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                    (cmd->base.autoneg != AUTONEG_ENABLE)) {
                        e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

(quoted diff) +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                }
        }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
        }
 
 out:
-       pm_runtime_put_sync(netdev->dev.parent);
        clear_bit(__E1000_RESETTING, &adapter->state);
        return ret_val;
Quoted 5 lines

}
==========================================

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

No, not intentional, looks like I just blew it, sorry. Will post a fix soon. Thanks a lot for debugging this.

Bjorn

#8 lore raw
In reply to Jiri Slaby

[+to Petr, -cc Jesse, bouncing]

Quoted 40 lines · Jiri Slaby
Quoted 1 line · Jiri Slaby

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
         * duplex is forced.
         */
        if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                    (cmd->base.autoneg != AUTONEG_ENABLE)) {
                        e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

(quoted diff) +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                }
        }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
        }
 
 out:
-       pm_runtime_put_sync(netdev->dev.parent);
        clear_bit(__E1000_RESETTING, &adapter->state);
        return ret_val;
Quoted 5 lines

}
==========================================

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

drivers/net/ethernet/intel/e1000e/ethtool.c +8 -4
index 9364bc2b4eb1..9db36ee71684 100644
@@ -280,7 +280,8 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
 	if (hw->phy.ops.check_reset_block &&
 	    hw->phy.ops.check_reset_block(hw)) {
 		e_err("Cannot change link characteristics when SoL/IDER is active.\n");
-		return -EINVAL;
+		ret_val = -EINVAL;
+		goto out;
 	}
 
 	/* MDI setting is only allowed when autoneg enabled because
@@ -288,13 +289,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
 	 * duplex is forced.
 	 */
 	if (cmd->base.eth_tp_mdix_ctrl) {
-		if (hw->phy.media_type != e1000_media_type_copper)
-			return -EOPNOTSUPP;
+		if (hw->phy.media_type != e1000_media_type_copper) {
+			ret_val = -EOPNOTSUPP;
+			goto out;
+		}
 
 		if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
 		    (cmd->base.autoneg != AUTONEG_ENABLE)) {
 			e_err("forcing MDI/MDI-X state is not supported when link speed and/or duplex are forced\n");
-			return -EINVAL;
+			ret_val = -EINVAL;
+			goto out;
 		}
 	}
 
#9 lore raw

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 46 lines

[+to Petr, -cc Jesse, bouncing]

Quoted 43 lines · Jiri Slaby
Quoted 1 line · Jiri Slaby

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
          * duplex is forced.
          */
         if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                 if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                     (cmd->base.autoneg != AUTONEG_ENABLE)) {
                         e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

(quoted diff) +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                 }
         }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
         }
 
  out:
-       pm_runtime_put_sync(netdev->dev.parent);
         clear_bit(__E1000_RESETTING, &adapter->state);
         return ret_val;

}
==========================================

Quoted 7 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Quoted 34 lines
drivers/net/ethernet/intel/e1000e/ethtool.c +8 -4
index 9364bc2b4eb1..9db36ee71684 100644
@@ -280,7 +280,8 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	if (hw->phy.ops.check_reset_block &&
  	    hw->phy.ops.check_reset_block(hw)) {
  		e_err("Cannot change link characteristics when SoL/IDER is active.\n");
-		return -EINVAL;
+		ret_val = -EINVAL;
+		goto out;
  	}
 
  	/* MDI setting is only allowed when autoneg enabled because
@@ -288,13 +289,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	 * duplex is forced.
  	 */
  	if (cmd->base.eth_tp_mdix_ctrl) {
-		if (hw->phy.media_type != e1000_media_type_copper)
-			return -EOPNOTSUPP;
+		if (hw->phy.media_type != e1000_media_type_copper) {
+			ret_val = -EOPNOTSUPP;
+			goto out;
+		}
 
  		if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
  		    (cmd->base.autoneg != AUTONEG_ENABLE)) {
  			e_err("forcing MDI/MDI-X state is not supported when link speed and/or duplex are forced\n");
-			return -EINVAL;
+			ret_val = -EINVAL;
+			goto out;
  		}
  	}
 
#10 lore raw
Quoted 52 lines · Petr Valenta

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 44 lines · Jiri Slaby
Quoted 43 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
          * duplex is forced.
          */
         if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                 if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                     (cmd->base.autoneg != AUTONEG_ENABLE)) {
                         e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

(quoted diff) +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                 }
         }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
         }
 
  out:
-       pm_runtime_put_sync(netdev->dev.parent);
         clear_bit(__E1000_RESETTING, &adapter->state);
         return ret_val;

}
==========================================

Quoted 6 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

Quoted 3 lines

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

Thank you very much for testing that!

Quoted 3 lines

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Thanks, that matches what I get when applying the patch on v6.10.

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

commit 5e92945ffe5c ("Revert "e1000e: Remove redundant runtime resume for ethtool_ops"")

Author: Bjorn Helgaas <[email protected]>
Date:   Tue Aug 20 16:18:32 2024 -0500

Revert "e1000e: Remove redundant runtime resume for ethtool_ops"

This reverts commit b2c289415b2b2ef112b78d5e73b4acecf5db409e.

drivers/net/ethernet/intel/e1000e/ethtool.c +56 -6
index 9364bc2b4eb1..61fa2f6b3708 100644
@@ -156,7 +156,7 @@ static int e1000_get_link_ksettings(struct net_device *netdev,
 			speed = adapter->link_speed;
 			cmd->base.duplex = adapter->link_duplex - 1;
 		}
-	} else {
+	} else if (!pm_runtime_suspended(netdev->dev.parent)) {
 		u32 status = er32(STATUS);
 
 		if (status & E1000_STATUS_LU) {
@@ -274,13 +274,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
 	ethtool_convert_link_mode_to_legacy_u32(&advertising,
 						cmd->link_modes.advertising);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	/* When SoL/IDER sessions are active, autoneg/speed/duplex
 	 * cannot be changed
 	 */
 	if (hw->phy.ops.check_reset_block &&
 	    hw->phy.ops.check_reset_block(hw)) {
 		e_err("Cannot change link characteristics when SoL/IDER is active.\n");
-		return -EINVAL;
+		ret_val = -EINVAL;
+		goto out;
 	}
 
 	/* MDI setting is only allowed when autoneg enabled because
@@ -288,13 +291,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
 	 * duplex is forced.
 	 */
 	if (cmd->base.eth_tp_mdix_ctrl) {
-		if (hw->phy.media_type != e1000_media_type_copper)
-			return -EOPNOTSUPP;
+		if (hw->phy.media_type != e1000_media_type_copper) {
+			ret_val = -EOPNOTSUPP;
+			goto out;
+		}
 
 		if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
 		    (cmd->base.autoneg != AUTONEG_ENABLE)) {
 			e_err("forcing MDI/MDI-X state is not supported when link speed and/or duplex are forced\n");
-			return -EINVAL;
+			ret_val = -EINVAL;
+			goto out;
 		}
 	}
 
@@ -341,6 +347,7 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
 	}
 
 out:
+	pm_runtime_put_sync(netdev->dev.parent);
 	clear_bit(__E1000_RESETTING, &adapter->state);
 	return ret_val;
 }
@@ -376,6 +383,8 @@ static int e1000_set_pauseparam(struct net_device *netdev,
 	while (test_and_set_bit(__E1000_RESETTING, &adapter->state))
 		usleep_range(1000, 2000);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	if (adapter->fc_autoneg == AUTONEG_ENABLE) {
 		hw->fc.requested_mode = e1000_fc_default;
 		if (netif_running(adapter->netdev)) {
@@ -408,6 +417,7 @@ static int e1000_set_pauseparam(struct net_device *netdev,
 	}
 
 out:
+	pm_runtime_put_sync(netdev->dev.parent);
 	clear_bit(__E1000_RESETTING, &adapter->state);
 	return retval;
 }
@@ -438,6 +448,8 @@ static void e1000_get_regs(struct net_device *netdev,
 	u32 *regs_buff = p;
 	u16 phy_data;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	memset(p, 0, E1000_REGS_LEN * sizeof(u32));
 
 	regs->version = (1u << 24) |
@@ -483,6 +495,8 @@ static void e1000_get_regs(struct net_device *netdev,
 	e1e_rphy(hw, MII_STAT1000, &phy_data);
 	regs_buff[24] = (u32)phy_data;	/* phy local receiver status */
 	regs_buff[25] = regs_buff[24];	/* phy remote receiver status */
+
+	pm_runtime_put_sync(netdev->dev.parent);
 }
 
 static int e1000_get_eeprom_len(struct net_device *netdev)
@@ -515,6 +529,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
 	if (!eeprom_buff)
 		return -ENOMEM;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	if (hw->nvm.type == e1000_nvm_eeprom_spi) {
 		ret_val = e1000_read_nvm(hw, first_word,
 					 last_word - first_word + 1,
@@ -528,6 +544,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
 		}
 	}
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
 	if (ret_val) {
 		/* a read error occurred, throw away the result */
 		memset(eeprom_buff, 0xff, sizeof(u16) *
@@ -577,6 +595,8 @@ static int e1000_set_eeprom(struct net_device *netdev,
 
 	ptr = (void *)eeprom_buff;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	if (eeprom->offset & 1) {
 		/* need read/modify/write of first changed EEPROM word */
 		/* only the second byte of the word is being modified */
@@ -617,6 +637,7 @@ static int e1000_set_eeprom(struct net_device *netdev,
 		ret_val = e1000e_update_nvm_checksum(hw);
 
 out:
+	pm_runtime_put_sync(netdev->dev.parent);
 	kfree(eeprom_buff);
 	return ret_val;
 }
@@ -712,6 +733,8 @@ static int e1000_set_ringparam(struct net_device *netdev,
 		}
 	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	e1000e_down(adapter, true);
 
 	/* We can't just free everything and then setup again, because the
@@ -750,6 +773,7 @@ static int e1000_set_ringparam(struct net_device *netdev,
 		e1000e_free_tx_resources(temp_tx);
 err_setup:
 	e1000e_up(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
 free_temp:
 	vfree(temp_tx);
 	vfree(temp_rx);
@@ -1792,6 +1816,8 @@ static void e1000_diag_test(struct net_device *netdev,
 	u8 autoneg;
 	bool if_running = netif_running(netdev);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	set_bit(__E1000_TESTING, &adapter->state);
 
 	if (!if_running) {
@@ -1877,6 +1903,8 @@ static void e1000_diag_test(struct net_device *netdev,
 	}
 
 	msleep_interruptible(4 * 1000);
+
+	pm_runtime_put_sync(netdev->dev.parent);
 }
 
 static void e1000_get_wol(struct net_device *netdev,
@@ -2018,11 +2046,15 @@ static int e1000_set_coalesce(struct net_device *netdev,
 		adapter->itr_setting = adapter->itr & ~3;
 	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	if (adapter->itr_setting != 0)
 		e1000e_write_itr(adapter, adapter->itr);
 	else
 		e1000e_write_itr(adapter, 0);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
 	return 0;
 }
 
@@ -2036,7 +2068,9 @@ static int e1000_nway_reset(struct net_device *netdev)
 	if (!adapter->hw.mac.autoneg)
 		return -EINVAL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
 	e1000e_reinit_locked(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
 
 	return 0;
 }
@@ -2050,8 +2084,12 @@ static void e1000_get_ethtool_stats(struct net_device *netdev,
 	int i;
 	char *p = NULL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	dev_get_stats(netdev, &net_stats);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
 	for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) {
 		switch (e1000_gstrings_stats[i].type) {
 		case NETDEV_STATS:
@@ -2108,7 +2146,9 @@ static int e1000_get_rxnfc(struct net_device *netdev,
 		struct e1000_hw *hw = &adapter->hw;
 		u32 mrqc;
 
+		pm_runtime_get_sync(netdev->dev.parent);
 		mrqc = er32(MRQC);
+		pm_runtime_put_sync(netdev->dev.parent);
 
 		if (!(mrqc & E1000_MRQC_RSS_FIELD_MASK))
 			return 0;
@@ -2171,9 +2211,13 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
 		return -EOPNOTSUPP;
 	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	ret_val = hw->phy.ops.acquire(hw);
-	if (ret_val)
+	if (ret_val) {
+		pm_runtime_put_sync(netdev->dev.parent);
 		return -EBUSY;
+	}
 
 	/* EEE Capability */
 	ret_val = e1000_read_emi_reg_locked(hw, cap_addr, &phy_data);
@@ -2213,6 +2257,8 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
 	if (ret_val)
 		ret_val = -ENODATA;
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
 	return ret_val;
 }
 
@@ -2253,12 +2299,16 @@ static int e1000e_set_eee(struct net_device *netdev, struct ethtool_keee *edata)
 
 	hw->dev_spec.ich8lan.eee_disable = !edata->eee_enabled;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
 	/* reset the link */
 	if (netif_running(netdev))
 		e1000e_reinit_locked(adapter);
 	else
 		e1000e_reset(adapter);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
 	return 0;
 }
 
author
#11 lore raw
Quoted 60 lines · Bjorn Helgaas
Quoted 52 lines · Petr Valenta

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 44 lines · Jiri Slaby
Quoted 43 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
           * duplex is forced.
           */
          if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                  if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                      (cmd->base.autoneg != AUTONEG_ENABLE)) {
                          e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

drivers/net/ethernet/intel/e1000e/ethtool.c +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                  }
          }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
          }
 
   out:
-       pm_runtime_put_sync(netdev->dev.parent);
          clear_bit(__E1000_RESETTING, &adapter->state);
          return ret_val;

}
==========================================

Quoted 6 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

Quoted 3 lines

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

Thank you very much for testing that!

Quoted 8 lines
Quoted 3 lines

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Thanks, that matches what I get when applying the patch on v6.10.

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

FWIW he already tested with b2c289415b2b reverted (I provided him with a built kernel). It behaves the same. So you are not the breaker.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

Or perhaps easier to do:
git bisect v6.10 v6.9 -- drivers/net/ethernet/intel/e1000e/
directly. But that assumes one of the above commits broke it. If they did not, as a last resort, you can still do full bisect (without the "-- drivers" part).

I would take v6.10 suses config.
Would boot 6.10.
do lsmod > /tmp/lsmod
make LSMOD=/tmp/lsmod localyesconfig
make bzImage
and use that bzImage.

Note that graphics, wireless and other stuff will be defunct unless you build in firmwares for them (EXTRA_FIRMWARE config). Alternatively use localmodconfig and build and install also modules (now limited to your machine).

thanks,

Signature · 2 lines
js
suse labs
#12 lore raw
Quoted 100 lines · Jiri Slaby
Quoted 62 lines · Bjorn Helgaas
Quoted 54 lines · Petr Valenta

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 46 lines · Jiri Slaby
Quoted 45 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +0 -0
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct 

net_device
*net
dev,
           * duplex is forced.
           */
          if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) { -                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper) +                       return -EOPNOTSUPP;

                  if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) && (cmd->base.autoneg != AUTONEG_ENABLE)) { e_err("forcing MDI/MDI-X state is not supported when lin
k speed and/or duplex are forced\n");
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                  }
          }

(quoted diff) +0 -0
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct 

net_device
*netde
v,
          }

   out:
-       pm_runtime_put_sync(netdev->dev.parent); clear_bit(__E1000_RESETTING, &adapter->state); return ret_val;
   }
==========================================

Quoted 6 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional.  Petr, do you have the ability to test the patch below?  I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

Quoted 3 lines

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

Thank you very much for testing that!

Quoted 8 lines
Quoted 3 lines

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Thanks, that matches what I get when applying the patch on v6.10.

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

FWIW he already tested with b2c289415b2b reverted (I provided him with a built kernel). It behaves the same. So you are not the breaker.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

   76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

Or perhaps easier to do:
  git bisect v6.10 v6.9 -- drivers/net/ethernet/intel/e1000e/
directly. But that assumes one of the above commits broke it. If they did not, as a last resort, you can still do full bisect (without the "-- drivers" part).

I would take v6.10 suses config.
Would boot 6.10.
do lsmod > /tmp/lsmod
make LSMOD=/tmp/lsmod localyesconfig
make bzImage
and use that bzImage.

Quoted 3 lines

Note that graphics, wireless and other stuff will be defunct unless you build in firmwares for them (EXTRA_FIRMWARE config). Alternatively use localmodconfig and build and install also modules (now limited to your machine).

thanks,

From all the data in the Suse Bugzilla I understood that the issue happens with a cable disconnected.

Does it reproduce with a connected cable?

Also, I suspect that it might be related to wake up interrupts that the I219 device gets which might cause the IRQ storm.

I would like to ask you to enable the debug prints in the e1000e driver and share the dmesg log. I would like to see if we can get the values of WUFC and WUS on on resume.

For getting these values please comment out the following in __e1000_resume function in netdev.c file:

        /* report the system wakeup cause from S3/S4 */ //      if (adapter->flags2 & FLAG2_HAS_PHY_WAKEUP) { u16 phy_data;

                e1e_rphy(&adapter->hw, BM_WUS, &phy_data);
                if (phy_data) {
                        e_info("PHY Wakeup cause - %s\n", phy_data & E1000_WUS_EX ? "Unicast Packet" :
                               phy_data & E1000_WUS_MC ? "Multicast Packet" :
                               phy_data & E1000_WUS_BC ? "Broadcast Packet" :
                               phy_data & E1000_WUS_MAG ? "Magic Packet" :
                               phy_data & E1000_WUS_LNKC ?
                               "Link Status Change" : "other");
                }
                e1e_wphy(&adapter->hw, BM_WUS, ~0);
//      } else {
                u32 wus = er32(WUS);

                if (wus) {
                        e_info("MAC Wakeup cause - %s\n",
                               wus & E1000_WUS_EX ? "Unicast Packet" :
                               wus & E1000_WUS_MC ? "Multicast Packet" :
                               wus & E1000_WUS_BC ? "Broadcast Packet" :
                               wus & E1000_WUS_MAG ? "Magic Packet" :
                               wus & E1000_WUS_LNKC ? "Link Status Change" :
                               "other");
//              }
                ew32(WUS, ~0);

To enable the debug prints you can run:

echo "module e1000e +p" | sudo tee /sys/kernel/debug/dynamic_debug/control

#13 lore raw
In reply to Bjorn Helgaas

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 59 lines · Petr Valenta
Quoted 52 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 44 lines · Jiri Slaby
Quoted 43 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
           * duplex is forced.
           */
          if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                  if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                      (cmd->base.autoneg != AUTONEG_ENABLE)) {
                          e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

drivers/net/ethernet/intel/e1000e/ethtool.c +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                  }
          }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
          }
 
   out:
-       pm_runtime_put_sync(netdev->dev.parent);
          clear_bit(__E1000_RESETTING, &adapter->state);
          return ret_val;

}
==========================================

Quoted 6 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

Quoted 3 lines

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

Thank you very much for testing that!

Quoted 8 lines
Quoted 3 lines

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Thanks, that matches what I get when applying the patch on v6.10.

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 12 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Quoted 717 lines

commit 5e92945ffe5c ("Revert "e1000e: Remove redundant runtime resume for ethtool_ops"")

Author: Bjorn Helgaas <[email protected]>
Date:   Tue Aug 20 16:18:32 2024 -0500

Revert "e1000e: Remove redundant runtime resume for ethtool_ops"

This reverts commit b2c289415b2b2ef112b78d5e73b4acecf5db409e.

drivers/net/ethernet/intel/e1000e/ethtool.c +56 -6
index 9364bc2b4eb1..61fa2f6b3708 100644
@@ -156,7 +156,7 @@ static int e1000_get_link_ksettings(struct net_device *netdev,
  			speed = adapter->link_speed;
  			cmd->base.duplex = adapter->link_duplex - 1;
  		}
-	} else {
+	} else if (!pm_runtime_suspended(netdev->dev.parent)) {
  		u32 status = er32(STATUS);
 
  		if (status & E1000_STATUS_LU) {
@@ -274,13 +274,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	ethtool_convert_link_mode_to_legacy_u32(&advertising,
  						cmd->link_modes.advertising);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	/* When SoL/IDER sessions are active, autoneg/speed/duplex
  	 * cannot be changed
  	 */
  	if (hw->phy.ops.check_reset_block &&
  	    hw->phy.ops.check_reset_block(hw)) {
  		e_err("Cannot change link characteristics when SoL/IDER is active.\n");
-		return -EINVAL;
+		ret_val = -EINVAL;
+		goto out;
  	}
 
  	/* MDI setting is only allowed when autoneg enabled because
@@ -288,13 +291,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	 * duplex is forced.
  	 */
  	if (cmd->base.eth_tp_mdix_ctrl) {
-		if (hw->phy.media_type != e1000_media_type_copper)
-			return -EOPNOTSUPP;
+		if (hw->phy.media_type != e1000_media_type_copper) {
+			ret_val = -EOPNOTSUPP;
+			goto out;
+		}
 
  		if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
  		    (cmd->base.autoneg != AUTONEG_ENABLE)) {
  			e_err("forcing MDI/MDI-X state is not supported when link speed and/or duplex are forced\n");
-			return -EINVAL;
+			ret_val = -EINVAL;
+			goto out;
  		}
  	}
 
@@ -341,6 +347,7 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	}
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	clear_bit(__E1000_RESETTING, &adapter->state);
  	return ret_val;
  }
@@ -376,6 +383,8 @@ static int e1000_set_pauseparam(struct net_device *netdev,
  	while (test_and_set_bit(__E1000_RESETTING, &adapter->state))
  		usleep_range(1000, 2000);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (adapter->fc_autoneg == AUTONEG_ENABLE) {
  		hw->fc.requested_mode = e1000_fc_default;
  		if (netif_running(adapter->netdev)) {
@@ -408,6 +417,7 @@ static int e1000_set_pauseparam(struct net_device *netdev,
  	}
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	clear_bit(__E1000_RESETTING, &adapter->state);
  	return retval;
  }
@@ -438,6 +448,8 @@ static void e1000_get_regs(struct net_device *netdev,
  	u32 *regs_buff = p;
  	u16 phy_data;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	memset(p, 0, E1000_REGS_LEN * sizeof(u32));
 
  	regs->version = (1u << 24) |
@@ -483,6 +495,8 @@ static void e1000_get_regs(struct net_device *netdev,
  	e1e_rphy(hw, MII_STAT1000, &phy_data);
  	regs_buff[24] = (u32)phy_data;	/* phy local receiver status */
  	regs_buff[25] = regs_buff[24];	/* phy remote receiver status */
+
+	pm_runtime_put_sync(netdev->dev.parent);
  }
 
  static int e1000_get_eeprom_len(struct net_device *netdev)
@@ -515,6 +529,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
  	if (!eeprom_buff)
  		return -ENOMEM;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (hw->nvm.type == e1000_nvm_eeprom_spi) {
  		ret_val = e1000_read_nvm(hw, first_word,
  					 last_word - first_word + 1,
@@ -528,6 +544,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
  		}
  	}
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	if (ret_val) {
  		/* a read error occurred, throw away the result */
  		memset(eeprom_buff, 0xff, sizeof(u16) *
@@ -577,6 +595,8 @@ static int e1000_set_eeprom(struct net_device *netdev,
 
  	ptr = (void *)eeprom_buff;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (eeprom->offset & 1) {
  		/* need read/modify/write of first changed EEPROM word */
  		/* only the second byte of the word is being modified */
@@ -617,6 +637,7 @@ static int e1000_set_eeprom(struct net_device *netdev,
  		ret_val = e1000e_update_nvm_checksum(hw);
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	kfree(eeprom_buff);
  	return ret_val;
  }
@@ -712,6 +733,8 @@ static int e1000_set_ringparam(struct net_device *netdev,
  		}
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	e1000e_down(adapter, true);
 
  	/* We can't just free everything and then setup again, because the
@@ -750,6 +773,7 @@ static int e1000_set_ringparam(struct net_device *netdev,
  		e1000e_free_tx_resources(temp_tx);
  err_setup:
  	e1000e_up(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
  free_temp:
  	vfree(temp_tx);
  	vfree(temp_rx);
@@ -1792,6 +1816,8 @@ static void e1000_diag_test(struct net_device *netdev,
  	u8 autoneg;
  	bool if_running = netif_running(netdev);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	set_bit(__E1000_TESTING, &adapter->state);
 
  	if (!if_running) {
@@ -1877,6 +1903,8 @@ static void e1000_diag_test(struct net_device *netdev,
  	}
 
  	msleep_interruptible(4 * 1000);
+
+	pm_runtime_put_sync(netdev->dev.parent);
  }
 
  static void e1000_get_wol(struct net_device *netdev,
@@ -2018,11 +2046,15 @@ static int e1000_set_coalesce(struct net_device *netdev,
  		adapter->itr_setting = adapter->itr & ~3;
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (adapter->itr_setting != 0)
  		e1000e_write_itr(adapter, adapter->itr);
  	else
  		e1000e_write_itr(adapter, 0);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return 0;
  }
 
@@ -2036,7 +2068,9 @@ static int e1000_nway_reset(struct net_device *netdev)
  	if (!adapter->hw.mac.autoneg)
  		return -EINVAL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
  	e1000e_reinit_locked(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
 
  	return 0;
  }
@@ -2050,8 +2084,12 @@ static void e1000_get_ethtool_stats(struct net_device *netdev,
  	int i;
  	char *p = NULL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	dev_get_stats(netdev, &net_stats);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) {
  		switch (e1000_gstrings_stats[i].type) {
  		case NETDEV_STATS:
@@ -2108,7 +2146,9 @@ static int e1000_get_rxnfc(struct net_device *netdev,
  		struct e1000_hw *hw = &adapter->hw;
  		u32 mrqc;
 
+		pm_runtime_get_sync(netdev->dev.parent);
  		mrqc = er32(MRQC);
+		pm_runtime_put_sync(netdev->dev.parent);
 
  		if (!(mrqc & E1000_MRQC_RSS_FIELD_MASK))
  			return 0;
@@ -2171,9 +2211,13 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
  		return -EOPNOTSUPP;
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	ret_val = hw->phy.ops.acquire(hw);
-	if (ret_val)
+	if (ret_val) {
+		pm_runtime_put_sync(netdev->dev.parent);
  		return -EBUSY;
+	}
 
  	/* EEE Capability */
  	ret_val = e1000_read_emi_reg_locked(hw, cap_addr, &phy_data);
@@ -2213,6 +2257,8 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
  	if (ret_val)
  		ret_val = -ENODATA;
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return ret_val;
  }
 
@@ -2253,12 +2299,16 @@ static int e1000e_set_eee(struct net_device *netdev, struct ethtool_keee *edata)
 
  	hw->dev_spec.ich8lan.eee_disable = !edata->eee_enabled;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	/* reset the link */
  	if (netif_running(netdev))
  		e1000e_reinit_locked(adapter);
  	else
  		e1000e_reset(adapter);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return 0;
  }
 

From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <[email protected]>
X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on
aws-us-west-2-korg-lkml-1.web.codeaurora.org
Received: from smtp2.osuosl.org (smtp2.osuosl.org [140.211.166.133])

	(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
	(No client certificate requested)
	by smtp.lore.kernel.org (Postfix) with ESMTPS id 22EE6C3DA4A
	for <[email protected]>; Tue, 20 Aug 2024 21:30:44 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
	by smtp2.osuosl.org (Postfix) with ESMTP id E223040474;
	Tue, 20 Aug 2024 21:30:43 +0000 (UTC)

X-Virus-Scanned: amavis at osuosl.org
Received: from smtp2.osuosl.org ([127.0.0.1])
by localhost (smtp2.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id bS7tH8aI4dXw; Tue, 20 Aug 2024 21:30:42 +0000 (UTC)
X-Comment: SPF check N/A for local connections - client-ip=140.211.166.34; helo=ash.osuosl.org; [email protected]; receiver=<UNKNOWN>
DKIM-Filter: OpenDKIM Filter v2.11.0 smtp2.osuosl.org 9F4E940B38
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=osuosl.org;

	s=default; t=1724189442;
	bh=mlj2OU/9jQQlYVgc8xXGFGEbmSA2BUO2s9ZMNvlZa5E=;
	h=Date:From:To:In-Reply-To:Subject:List-Id:List-Unsubscribe:
	 List-Archive:List-Post:List-Help:List-Subscribe:Cc:From;
	b=gRsw4VpLnbvTsoxcHbTQgLUM2iVdECmRSG1uEylrYmPsdWJDin8aVjhm4QrxSXuMO
	 g+6njzmv+w6950L+rcQlJiK6ry1jbhFrNzTmFOyCvctsUeLmaqGu1jrLvOp2XhP3uJ
	 dOpebDahCjpzldI2iCU6Oy/iNQsLx2KSsxzZgtlNqOx3rXCjDFXINVRKkL9Sz2kv5b
	 etzK/KOTa/xffZ7mnp9ZelV0v0YGUFmgFKvt6/QLqeTpqpXphMr3MntldSJLu7BrFm
	 4EJloku9spGmIB2ZBt4wRHhPvuoKG4HPu8DszG7VJAOAj52QaC1bmnQ19OKwKM+Rlz
	 C0+0QilD6XFkQ==
Received: from ash.osuosl.org (ash.osuosl.org [140.211.166.34])
	by smtp2.osuosl.org (Postfix) with ESMTP id 9F4E940B38;
	Tue, 20 Aug 2024 21:30:42 +0000 (UTC)

Received: from smtp4.osuosl.org (smtp4.osuosl.org [140.211.166.137])
by ash.osuosl.org (Postfix) with ESMTP id B70001BF3B8
for <[email protected]>; Tue, 20 Aug 2024 21:30:40 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
by smtp4.osuosl.org (Postfix) with ESMTP id B04CA4067F
for <[email protected]>; Tue, 20 Aug 2024 21:30:40 +0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp4.osuosl.org ([127.0.0.1])
by localhost (smtp4.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP id 6VOvnWK04xp7 for <[email protected]>;
Tue, 20 Aug 2024 21:30:39 +0000 (UTC)
Received-SPF: Pass (mailfrom) identity=mailfrom;
client-ip=2604:1380:4641:c500::1; helo=dfw.source.kernel.org; [email protected]; receiver=<UNKNOWN>
DMARC-Filter: OpenDMARC Filter v1.4.2 smtp4.osuosl.org 3098940679
DKIM-Filter: OpenDKIM Filter v2.11.0 smtp4.osuosl.org 3098940679
Received: from dfw.source.kernel.org (dfw.source.kernel.org
[IPv6:2604:1380:4641:c500::1])
by smtp4.osuosl.org (Postfix) with ESMTPS id 3098940679
for <[email protected]>; Tue, 20 Aug 2024 21:30:39 +0000 (UTC)
Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58])
by dfw.source.kernel.org (Postfix) with ESMTP id EF93D60DF6;
Tue, 20 Aug 2024 21:30:37 +0000 (UTC)
Received: by smtp.kernel.org (Postfix) with ESMTPSA id 4A00BC4AF09;
Tue, 20 Aug 2024 21:30:37 +0000 (UTC)
Date: Tue, 20 Aug 2024 16:30:35 -0500
From: Bjorn Helgaas <[email protected]>
To: Petr Valenta <[email protected]>
Message-ID: <20240820213035.GA226181@bhelgaas>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <[email protected]> X-Mailman-Original-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple;
d=kernel.org; s=k20201202; t=1724189437; bh=0t79HOH2kYVclDTTF8FvHw6OfaaM4JS6jorC17zXI3k=; h=Date:From:To:Cc:Subject:In-Reply-To:From; b=GhFyFhraHJm9t4ey15N3XxG5Bc8U2z1N5/4pJSFm4vNlndjUOQltzb39mj8r5tzdO 0Qrk2iUfn9usdQTfr3eHowUDbvxoFI6z/iFJUe0kPts5eaue6Z5J1ESezm6daA6gwZ yJtcho1EXrvEH0dGBx7xhKIKBezS0BAXW2lbUFKuTB9YSHJvprlgy0VZjDNjg/G3WL 4nV3cEtNJ3EiMYMMNl9JNQtJDhvg/VIeHBFzQPJHHzBh+gZ4MmwNPJSkTNCDceZicW OJIm3Vprxxg/osSezx/ysd0VAXqV7ywftKaXHshPID8u49v2SiJtBafOh/+7jMlh+U CASP86+QHj9sA==
X-Mailman-Original-Authentication-Results: smtp4.osuosl.org;
dmarc=pass (p=quarantine dis=none)
header.from=kernel.org
X-Mailman-Original-Authentication-Results: smtp4.osuosl.org;
dkim=pass (2048-bit key,
unprotected) header.d=kernel.org [email protected] header.a=rsa-sha256 header.s=k20201202 header.b=GhFyFhra
Subject: Re: [Intel-wired-lan] ACPI IRQ storm with 6.10
X-BeenThere: [email protected]
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Intel Wired Ethernet Linux Kernel Driver Development
<intel-wired-lan.osuosl.org>
List-Unsubscribe: <https://lists.osuosl.org/mailman/options/intel-wired-lan>,
<mailto:[email protected]?subject=unsubscribe>
List-Archive: <http://lists.osuosl.org/pipermail/intel-wired-lan/>
List-Post: <mailto:[email protected]>
List-Help: <mailto:[email protected]?subject=help>
List-Subscribe: <https://lists.osuosl.org/mailman/listinfo/intel-wired-lan>,
<mailto:[email protected]?subject=subscribe>
Cc: Linux regressions mailing list <[email protected]>,
"Rafael J. Wysocki" <[email protected]>, [email protected], Linux kernel mailing list <[email protected]>, "[email protected]" <[email protected]>,
Tony Nguyen <[email protected]>, Bjorn Helgaas <[email protected]>, [email protected], Jiri Slaby <[email protected]>, Len Brown <[email protected]>
Errors-To: [email protected]
Sender: "Intel-wired-lan" <[email protected]>

Quoted 52 lines · Petr Valenta

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 44 lines · Jiri Slaby
Quoted 43 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
           * duplex is forced.
           */
          if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                  if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                      (cmd->base.autoneg != AUTONEG_ENABLE)) {
                          e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

drivers/net/ethernet/intel/e1000e/ethtool.c +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                  }
          }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
          }
 
   out:
-       pm_runtime_put_sync(netdev->dev.parent);
          clear_bit(__E1000_RESETTING, &adapter->state);
          return ret_val;

}
==========================================

Quoted 6 lines
Quoted 2 lines

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

Not intentional. Petr, do you have the ability to test the patch below? I'm not sure it's the correct fix, but it reverts the pieces of b2c289415b2b that Jiri pointed out.

Quoted 3 lines

I tested the patch below but it didn't help. After the first boot with new kernel it looked promising as the irq storm only appeared for a few seconds, but with subsequent reboots it was the same as without the patch.

Thank you very much for testing that!

Quoted 3 lines

To be sure, I also send the md5sum of ethtool.c after applying the patch:

a25c003257538f16994b4d7c4260e894 ethtool.c

Thanks, that matches what I get when applying the patch on v6.10.

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

commit 5e92945ffe5c ("Revert "e1000e: Remove redundant runtime resume for ethtool_ops"")

Author: Bjorn Helgaas <[email protected]>
Date:   Tue Aug 20 16:18:32 2024 -0500

Revert "e1000e: Remove redundant runtime resume for ethtool_ops"

This reverts commit b2c289415b2b2ef112b78d5e73b4acecf5db409e.

drivers/net/ethernet/intel/e1000e/ethtool.c +56 -6
index 9364bc2b4eb1..61fa2f6b3708 100644
@@ -156,7 +156,7 @@ static int e1000_get_link_ksettings(struct net_device *netdev,
  			speed = adapter->link_speed;
  			cmd->base.duplex = adapter->link_duplex - 1;
  		}
-	} else {
+	} else if (!pm_runtime_suspended(netdev->dev.parent)) {
  		u32 status = er32(STATUS);
 
  		if (status & E1000_STATUS_LU) {
@@ -274,13 +274,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	ethtool_convert_link_mode_to_legacy_u32(&advertising,
  						cmd->link_modes.advertising);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	/* When SoL/IDER sessions are active, autoneg/speed/duplex
  	 * cannot be changed
  	 */
  	if (hw->phy.ops.check_reset_block &&
  	    hw->phy.ops.check_reset_block(hw)) {
  		e_err("Cannot change link characteristics when SoL/IDER is active.\n");
-		return -EINVAL;
+		ret_val = -EINVAL;
+		goto out;
  	}
 
  	/* MDI setting is only allowed when autoneg enabled because
@@ -288,13 +291,16 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	 * duplex is forced.
  	 */
  	if (cmd->base.eth_tp_mdix_ctrl) {
-		if (hw->phy.media_type != e1000_media_type_copper)
-			return -EOPNOTSUPP;
+		if (hw->phy.media_type != e1000_media_type_copper) {
+			ret_val = -EOPNOTSUPP;
+			goto out;
+		}
 
  		if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
  		    (cmd->base.autoneg != AUTONEG_ENABLE)) {
  			e_err("forcing MDI/MDI-X state is not supported when link speed and/or duplex are forced\n");
-			return -EINVAL;
+			ret_val = -EINVAL;
+			goto out;
  		}
  	}
 
@@ -341,6 +347,7 @@ static int e1000_set_link_ksettings(struct net_device *netdev,
  	}
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	clear_bit(__E1000_RESETTING, &adapter->state);
  	return ret_val;
  }
@@ -376,6 +383,8 @@ static int e1000_set_pauseparam(struct net_device *netdev,
  	while (test_and_set_bit(__E1000_RESETTING, &adapter->state))
  		usleep_range(1000, 2000);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (adapter->fc_autoneg == AUTONEG_ENABLE) {
  		hw->fc.requested_mode = e1000_fc_default;
  		if (netif_running(adapter->netdev)) {
@@ -408,6 +417,7 @@ static int e1000_set_pauseparam(struct net_device *netdev,
  	}
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	clear_bit(__E1000_RESETTING, &adapter->state);
  	return retval;
  }
@@ -438,6 +448,8 @@ static void e1000_get_regs(struct net_device *netdev,
  	u32 *regs_buff = p;
  	u16 phy_data;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	memset(p, 0, E1000_REGS_LEN * sizeof(u32));
 
  	regs->version = (1u << 24) |
@@ -483,6 +495,8 @@ static void e1000_get_regs(struct net_device *netdev,
  	e1e_rphy(hw, MII_STAT1000, &phy_data);
  	regs_buff[24] = (u32)phy_data;	/* phy local receiver status */
  	regs_buff[25] = regs_buff[24];	/* phy remote receiver status */
+
+	pm_runtime_put_sync(netdev->dev.parent);
  }
 
  static int e1000_get_eeprom_len(struct net_device *netdev)
@@ -515,6 +529,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
  	if (!eeprom_buff)
  		return -ENOMEM;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (hw->nvm.type == e1000_nvm_eeprom_spi) {
  		ret_val = e1000_read_nvm(hw, first_word,
  					 last_word - first_word + 1,
@@ -528,6 +544,8 @@ static int e1000_get_eeprom(struct net_device *netdev,
  		}
  	}
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	if (ret_val) {
  		/* a read error occurred, throw away the result */
  		memset(eeprom_buff, 0xff, sizeof(u16) *
@@ -577,6 +595,8 @@ static int e1000_set_eeprom(struct net_device *netdev,
 
  	ptr = (void *)eeprom_buff;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (eeprom->offset & 1) {
  		/* need read/modify/write of first changed EEPROM word */
  		/* only the second byte of the word is being modified */
@@ -617,6 +637,7 @@ static int e1000_set_eeprom(struct net_device *netdev,
  		ret_val = e1000e_update_nvm_checksum(hw);
 
  out:
+	pm_runtime_put_sync(netdev->dev.parent);
  	kfree(eeprom_buff);
  	return ret_val;
  }
@@ -712,6 +733,8 @@ static int e1000_set_ringparam(struct net_device *netdev,
  		}
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	e1000e_down(adapter, true);
 
  	/* We can't just free everything and then setup again, because the
@@ -750,6 +773,7 @@ static int e1000_set_ringparam(struct net_device *netdev,
  		e1000e_free_tx_resources(temp_tx);
  err_setup:
  	e1000e_up(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
  free_temp:
  	vfree(temp_tx);
  	vfree(temp_rx);
@@ -1792,6 +1816,8 @@ static void e1000_diag_test(struct net_device *netdev,
  	u8 autoneg;
  	bool if_running = netif_running(netdev);
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	set_bit(__E1000_TESTING, &adapter->state);
 
  	if (!if_running) {
@@ -1877,6 +1903,8 @@ static void e1000_diag_test(struct net_device *netdev,
  	}
 
  	msleep_interruptible(4 * 1000);
+
+	pm_runtime_put_sync(netdev->dev.parent);
  }
 
  static void e1000_get_wol(struct net_device *netdev,
@@ -2018,11 +2046,15 @@ static int e1000_set_coalesce(struct net_device *netdev,
  		adapter->itr_setting = adapter->itr & ~3;
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	if (adapter->itr_setting != 0)
  		e1000e_write_itr(adapter, adapter->itr);
  	else
  		e1000e_write_itr(adapter, 0);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return 0;
  }
 
@@ -2036,7 +2068,9 @@ static int e1000_nway_reset(struct net_device *netdev)
  	if (!adapter->hw.mac.autoneg)
  		return -EINVAL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
  	e1000e_reinit_locked(adapter);
+	pm_runtime_put_sync(netdev->dev.parent);
 
  	return 0;
  }
@@ -2050,8 +2084,12 @@ static void e1000_get_ethtool_stats(struct net_device *netdev,
  	int i;
  	char *p = NULL;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	dev_get_stats(netdev, &net_stats);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	for (i = 0; i < E1000_GLOBAL_STATS_LEN; i++) {
  		switch (e1000_gstrings_stats[i].type) {
  		case NETDEV_STATS:
@@ -2108,7 +2146,9 @@ static int e1000_get_rxnfc(struct net_device *netdev,
  		struct e1000_hw *hw = &adapter->hw;
  		u32 mrqc;
 
+		pm_runtime_get_sync(netdev->dev.parent);
  		mrqc = er32(MRQC);
+		pm_runtime_put_sync(netdev->dev.parent);
 
  		if (!(mrqc & E1000_MRQC_RSS_FIELD_MASK))
  			return 0;
@@ -2171,9 +2211,13 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
  		return -EOPNOTSUPP;
  	}
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	ret_val = hw->phy.ops.acquire(hw);
-	if (ret_val)
+	if (ret_val) {
+		pm_runtime_put_sync(netdev->dev.parent);
  		return -EBUSY;
+	}
 
  	/* EEE Capability */
  	ret_val = e1000_read_emi_reg_locked(hw, cap_addr, &phy_data);
@@ -2213,6 +2257,8 @@ static int e1000e_get_eee(struct net_device *netdev, struct ethtool_keee *edata)
  	if (ret_val)
  		ret_val = -ENODATA;
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return ret_val;
  }
 
@@ -2253,12 +2299,16 @@ static int e1000e_set_eee(struct net_device *netdev, struct ethtool_keee *edata)
 
  	hw->dev_spec.ich8lan.eee_disable = !edata->eee_enabled;
 
+	pm_runtime_get_sync(netdev->dev.parent);
+
  	/* reset the link */
  	if (netif_running(netdev))
  		e1000e_reinit_locked(adapter);
  	else
  		e1000e_reset(adapter);
 
+	pm_runtime_put_sync(netdev->dev.parent);
+
  	return 0;
  }
 
#14 lore raw

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 32 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing. It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1. I added the authors of all three in case they have any insights.

Bjorn

#15 lore raw
Quoted 42 lines · Bjorn Helgaas

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing. It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1. I added the authors of all three in case they have any insights.

Bjorn

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

#16 lore raw

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 51 lines · Bjorn Helgaas
Quoted 42 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss. You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing. It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1. I added the authors of all three in case they have any insights.

Bjorn

Quoted 7 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

Quoted 2 lines
#17 lore raw

Dne 22. 08. 24 v 9:44 Petr Valenta napsal(a):

Quoted 245 lines

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 51 lines · Bjorn Helgaas
Quoted 42 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

Quoted 7 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

Quoted 2 lines

From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <[email protected]>
X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on
    aws-us-west-2-korg-lkml-1.web.codeaurora.org
Received: from smtp2.osuosl.org (smtp2.osuosl.org [140.211.166.133])
    (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
    (No client certificate requested)
    by smtp.lore.kernel.org (Postfix) with ESMTPS id 7319CC531DF
    for <[email protected]>; Thu, 22 Aug 2024
07:44:59 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
    by smtp2.osuosl.org (Postfix) with ESMTP id 2EE99404B8;
    Thu, 22 Aug 2024 07:44:59 +0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp2.osuosl.org ([127.0.0.1])
by localhost (smtp2.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id VRgkrPDlq_WW; Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
X-Comment: SPF check N/A for local connections - client-ip=140.211.166.34; helo=ash.osuosl.org; [email protected]; receiver=<UNKNOWN> DKIM-Filter: OpenDKIM Filter v2.11.0 smtp2.osuosl.org 53F64405BA
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=osuosl.org;
    s=default; t=1724312696;
    bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
    h=Date:To:References:From:In-Reply-To:Subject:List-Id:
     List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe:
     Cc:From;
    b=ZIudOsHGSDoQvtekseiE4SUpOKofnvHlxj7aT3f7bLvqCDMOCfygsO6tctN23YgSh
     xYqnq4yBSB4/JQ4v7Juyg0P/wqTcr+XFqhORTc2qBku9GCA+Y4wRKbRUeH4/AUNthL
     cf/zG7uEOFEKz4YALwviQFqR5E+HW9gD+YnXahtGUVqYiTjB01HuESDZdYI5huiCLI
     eHnQDw/SSwM1YmkjLzQgICjlxtIRVYjUL+shaltRg9f7t4otZa9bvrvLptzw5Mrfc0
     GLvrNRmHckPFKEJOXgmIeQI40IOHckD3MX2dkQ2dQ0VCrkl9JIgtuSRuS3IpB1dr65
     TatTrq9Onm26w==
Received: from ash.osuosl.org (ash.osuosl.org [140.211.166.34])
    by smtp2.osuosl.org (Postfix) with ESMTP id 53F64405BA;
    Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
Received: from smtp1.osuosl.org (smtp1.osuosl.org [140.211.166.138])
by ash.osuosl.org (Postfix) with ESMTP id 81E351BF322
for <[email protected]>; Thu, 22 Aug 2024 07:44:54 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
by smtp1.osuosl.org (Postfix) with ESMTP id 79A0C80A82
for <[email protected]>; Thu, 22 Aug 2024 07:44:54 +0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp1.osuosl.org ([127.0.0.1])
by localhost (smtp1.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id m9sJJpu9kR7y for <[email protected]>;
Thu, 22 Aug 2024 07:44:53 +0000 (UTC)
Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=136.143.188.52;
helo=sender4-of-o52.zoho.com; [email protected];
receiver=<UNKNOWN> DMARC-Filter: OpenDMARC Filter v1.4.2 smtp1.osuosl.org 3674B80A59
DKIM-Filter: OpenDKIM Filter v2.11.0 smtp1.osuosl.org 3674B80A59
Received: from sender4-of-o52.zoho.com (sender4-of-o52.zoho.com
[136.143.188.52])
by smtp1.osuosl.org (Postfix) with ESMTPS id 3674B80A59
for <[email protected]>; Thu, 22 Aug 2024 07:44:51 +0000 (UTC)
ARC-Seal: i=1; a=rsa-sha256; t=1724312671; cv=none; d=zohomail.com; s=zohoarc; b=B0wnUG3UHEcTRfbjC9HSfLJG+WBnpU18yag7r0240QuMQMnP/cHcj9e4oJU2FgxRPLpt6OGnlZOiPNE2GUFgnkBzKBPwzxb7eTHFwW4P8cW+1IrIOQ6jZWd2rhOIyWcRKYMydfCbMPM04Z+RwKVyRlrLTYL5UDBYYKKHOG08Ikc=
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com;
s=zohoarc; t=1724312671;
h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:MIME-Version:Message-ID:References:Subject:Subject:To:To:Message-Id:Reply-To;
bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=; b=VvZAc/xKVy85rZpNNCAwUwpxquk4r4Xw2QjZmePGlnINwOvJf6oilR9lqx2WDMezV20iKTW9f3dauO4jIjp363HOdh7P21UFfa66a0oK63RODo7IQMHSCqaCwAEoO1PKHfDfTMwz0/BShU1dt+nhtAeSeKwbG7G1qizCcoXTdjo= ARC-Authentication-Results: i=1; mx.zohomail.com;
dkim=pass  header.i=jevklidu.cz;
spf=pass  [email protected];
dmarc=pass header.from=<[email protected]>
Received: by mx.zohomail.com with SMTPS id 1724312669862808.3168476405893;
Thu, 22 Aug 2024 00:44:29 -0700 (PDT)
Message-ID: <[email protected]>
Date: Thu, 22 Aug 2024 09:44:22 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
To: Vitaly Lifshits <[email protected]>,
Bjorn Helgaas <[email protected]>, Dima Ruinskiy <[email protected]>,
Hui Wang <[email protected]>
References: <20240821145959.GA248604@bhelgaas>
<[email protected]>
Content-Language: cs-CZ, en-US
From: Petr Valenta <[email protected]>
In-Reply-To: <[email protected]>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
X-ZohoMailClient: External
X-Mailman-Original-DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt;
c=relaxed/relaxed; t=1724312671; s=zoho; d=jevklidu.cz; [email protected]; h=Message-ID:Date:Date:MIME-Version:Subject:Subject:To:To:Cc:Cc:References:From:From:In-Reply-To:Content-Type:Content-Transfer-Encoding:Message-Id:Reply-To;
bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
b=RSnIQpqoQp2O3ExJNnw4fk9dlt8CX1T5sbtB6GflBDYejiRQJTcrU3zRHn3pRkFq
tm00/cgXr6pF6T5vJFttBkfrHtnRiPiE8cjqni5KsNxCyOXOwri6I5ARAmPcUj42eda
e/xHQX9E3ayXrWSBQDAsun3Ann63tcXQKwlT7ffI=
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dmarc=none (p=none dis=none)
header.from=jevklidu.cz
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dkim=pass (1024-bit key,
unprotected) header.d=jevklidu.cz [email protected]
header.a=rsa-sha256 header.s=zoho header.b=RSnIQpqo
Subject: Re: [Intel-wired-lan] ACPI IRQ storm with 6.10
X-BeenThere: [email protected]
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Intel Wired Ethernet Linux Kernel Driver Development
<intel-wired-lan.osuosl.org>
List-Unsubscribe: <https://lists.osuosl.org/mailman/options/intel-wired-lan>, <mailto:[email protected]?subject=unsubscribe>
List-Archive: <http://lists.osuosl.org/pipermail/intel-wired-lan/>
List-Post: <mailto:[email protected]>
List-Help: <mailto:[email protected]?subject=help>
List-Subscribe: <https://lists.osuosl.org/mailman/listinfo/intel-wired-lan>,
<mailto:[email protected]?subject=subscribe>
Cc: Linux regressions mailing list <[email protected]>,
"Rafael J. Wysocki" <[email protected]>, [email protected],
Linux kernel mailing list <[email protected]>,
"[email protected]" <[email protected]>,
Tony Nguyen <[email protected]>, Bjorn Helgaas <[email protected]>,
[email protected], Jiri Slaby <[email protected]>,
Len Brown <[email protected]>
Errors-To: [email protected]
Sender: "Intel-wired-lan" <[email protected]>

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 57 lines · Bjorn Helgaas
Quoted 42 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

Quoted 8 lines
Quoted 1 line

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

I have tested now how it behaves with kernel 6.9.9. There is a new finding. After running "echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control" network over cable works but irq storm arrises. I have never tested this before because I don't use cable with this laptop at all. After unplugging cable irq storm is gone.

A possible workaround would be to turn off power control for the e1000e at the kernel level (if is it possible) so that utilities like powertop don't cause irq storm or broken network.

Quoted 3 lines
Quoted 2 lines
#18 lore raw
Quoted 260 lines · Petr Valenta

Dne 22. 08. 24 v 9:44 Petr Valenta napsal(a):

Quoted 257 lines

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 52 lines · Bjorn Helgaas
Quoted 43 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

Quoted 8 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

Quoted 2 lines

 From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <[email protected]>
X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on
     aws-us-west-2-korg-lkml-1.web.codeaurora.org
Received: from smtp2.osuosl.org (smtp2.osuosl.org [140.211.166.133])
     (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
     (No client certificate requested)
     by smtp.lore.kernel.org (Postfix) with ESMTPS id 7319CC531DF
     for <[email protected]>; Thu, 22 Aug 2024 07:44:59 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
     by smtp2.osuosl.org (Postfix) with ESMTP id 2EE99404B8;
     Thu, 22 Aug 2024 07:44:59 +0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp2.osuosl.org ([127.0.0.1])
by localhost (smtp2.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id VRgkrPDlq_WW; Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
X-Comment: SPF check N/A for local connections - client-ip=140.211.166.34; helo=ash.osuosl.org; [email protected]; receiver=<UNKNOWN> DKIM-Filter: OpenDKIM Filter v2.11.0 smtp2.osuosl.org 53F64405BA
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=osuosl.org;
     s=default; t=1724312696;
     bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
     h=Date:To:References:From:In-Reply-To:Subject:List-Id:
List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe:
      Cc:From;
     b=ZIudOsHGSDoQvtekseiE4SUpOKofnvHlxj7aT3f7bLvqCDMOCfygsO6tctN23YgSh
xYqnq4yBSB4/JQ4v7Juyg0P/wqTcr+XFqhORTc2qBku9GCA+Y4wRKbRUeH4/AUNthL
cf/zG7uEOFEKz4YALwviQFqR5E+HW9gD+YnXahtGUVqYiTjB01HuESDZdYI5huiCLI
eHnQDw/SSwM1YmkjLzQgICjlxtIRVYjUL+shaltRg9f7t4otZa9bvrvLptzw5Mrfc0
GLvrNRmHckPFKEJOXgmIeQI40IOHckD3MX2dkQ2dQ0VCrkl9JIgtuSRuS3IpB1dr65
      TatTrq9Onm26w==
Received: from ash.osuosl.org (ash.osuosl.org [140.211.166.34])
     by smtp2.osuosl.org (Postfix) with ESMTP id 53F64405BA;
     Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
Received: from smtp1.osuosl.org (smtp1.osuosl.org [140.211.166.138])
by ash.osuosl.org (Postfix) with ESMTP id 81E351BF322
for <[email protected]>; Thu, 22 Aug 2024 07:44:54
+0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
by smtp1.osuosl.org (Postfix) with ESMTP id 79A0C80A82
for <[email protected]>; Thu, 22 Aug 2024 07:44:54
+0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp1.osuosl.org ([127.0.0.1])
by localhost (smtp1.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id m9sJJpu9kR7y for <[email protected]>;
Thu, 22 Aug 2024 07:44:53 +0000 (UTC)
Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=136.143.188.52;
helo=sender4-of-o52.zoho.com; [email protected];
receiver=<UNKNOWN> DMARC-Filter: OpenDMARC Filter v1.4.2 smtp1.osuosl.org 3674B80A59
DKIM-Filter: OpenDKIM Filter v2.11.0 smtp1.osuosl.org 3674B80A59
Received: from sender4-of-o52.zoho.com (sender4-of-o52.zoho.com
[136.143.188.52])
by smtp1.osuosl.org (Postfix) with ESMTPS id 3674B80A59
for <[email protected]>; Thu, 22 Aug 2024 07:44:51
+0000 (UTC)
ARC-Seal: i=1; a=rsa-sha256; t=1724312671; cv=none; d=zohomail.com; s=zohoarc; b=B0wnUG3UHEcTRfbjC9HSfLJG+WBnpU18yag7r0240QuMQMnP/cHcj9e4oJU2FgxRPLpt6OGnlZOiPNE2GUFgnkBzKBPwzxb7eTHFwW4P8cW+1IrIOQ6jZWd2rhOIyWcRKYMydfCbMPM04Z+RwKVyRlrLTYL5UDBYYKKHOG08Ikc=
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com;
s=zohoarc; t=1724312671;
h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:MIME-Version:Message-ID:References:Subject:Subject:To:To:Message-Id:Reply-To;

bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=; b=VvZAc/xKVy85rZpNNCAwUwpxquk4r4Xw2QjZmePGlnINwOvJf6oilR9lqx2WDMezV20iKTW9f3dauO4jIjp363HOdh7P21UFfa66a0oK63RODo7IQMHSCqaCwAEoO1PKHfDfTMwz0/BShU1dt+nhtAeSeKwbG7G1qizCcoXTdjo= ARC-Authentication-Results: i=1; mx.zohomail.com;
dkim=pass  header.i=jevklidu.cz;
spf=pass  [email protected];
dmarc=pass header.from=<[email protected]>
Received: by mx.zohomail.com with SMTPS id 1724312669862808.3168476405893;
Thu, 22 Aug 2024 00:44:29 -0700 (PDT)
Message-ID: <[email protected]>
Date: Thu, 22 Aug 2024 09:44:22 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
To: Vitaly Lifshits <[email protected]>,
Bjorn Helgaas <[email protected]>, Dima Ruinskiy <[email protected]>,
Hui Wang <[email protected]>
References: <20240821145959.GA248604@bhelgaas>
<[email protected]>
Content-Language: cs-CZ, en-US
From: Petr Valenta <[email protected]>
In-Reply-To: <[email protected]>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
X-ZohoMailClient: External
X-Mailman-Original-DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt;
c=relaxed/relaxed; t=1724312671; s=zoho; d=jevklidu.cz; [email protected]; h=Message-ID:Date:Date:MIME-Version:Subject:Subject:To:To:Cc:Cc:References:From:From:In-Reply-To:Content-Type:Content-Transfer-Encoding:Message-Id:Reply-To;

bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
b=RSnIQpqoQp2O3ExJNnw4fk9dlt8CX1T5sbtB6GflBDYejiRQJTcrU3zRHn3pRkFq tm00/cgXr6pF6T5vJFttBkfrHtnRiPiE8cjqni5KsNxCyOXOwri6I5ARAmPcUj42eda
e/xHQX9E3ayXrWSBQDAsun3Ann63tcXQKwlT7ffI=
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dmarc=none (p=none dis=none)
header.from=jevklidu.cz
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dkim=pass (1024-bit key,
unprotected) header.d=jevklidu.cz [email protected]
header.a=rsa-sha256 header.s=zoho header.b=RSnIQpqo
Subject: Re: [Intel-wired-lan] ACPI IRQ storm with 6.10
X-BeenThere: [email protected]
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Intel Wired Ethernet Linux Kernel Driver Development
<intel-wired-lan.osuosl.org>
List-Unsubscribe: <https://lists.osuosl.org/mailman/options/intel-wired-lan>, <mailto:[email protected]?subject=unsubscribe>
List-Archive: <http://lists.osuosl.org/pipermail/intel-wired-lan/>
List-Post: <mailto:[email protected]>
List-Help: <mailto:[email protected]?subject=help>
List-Subscribe: <https://lists.osuosl.org/mailman/listinfo/intel-wired-lan>,
<mailto:[email protected]?subject=subscribe>
Cc: Linux regressions mailing list <[email protected]>,
"Rafael J. Wysocki" <[email protected]>, [email protected],
Linux kernel mailing list <[email protected]>,
"[email protected]" <[email protected]>,
Tony Nguyen <[email protected]>, Bjorn Helgaas <[email protected]>,
[email protected], Jiri Slaby <[email protected]>,
Len Brown <[email protected]>
Errors-To: [email protected]
Sender: "Intel-wired-lan" <[email protected]>

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 52 lines · Bjorn Helgaas
Quoted 43 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

Quoted 8 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

Quoted 9 lines
Quoted 4 lines

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

I have tested now how it behaves with kernel 6.9.9. There is a new finding. After running "echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control" network over cable works but irq storm arrises. I have never tested this before because I don't use cable with this laptop at all. After unplugging cable irq storm is gone.

A possible workaround would be to turn off power control for the e1000e at the kernel level (if is it possible) so that utilities like powertop don't cause irq storm or broken network.

I would like to suggest the following for now as a triage:

1. revert the following patches:

   i.  e1000e: fix force smbus during suspend flow (76a0a3f9cc2f)

   ii.   e1000e: move force SMBUS near the end of enable_ulp function (bfd546a552e1)

   iii. "e1000e: move force SMBUS from enable ulp function to avoid PHY loss issue ( 861e8086029e)

2. apply this patch:
https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I expect this configuration not to have neither the IRQ storm nor the link indication issue (>"Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.")

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines
#19 lore raw

Dne 22. 08. 24 v 11:18 Vitaly Lifshits napsal(a):

Quoted 278 lines · Petr Valenta
Quoted 254 lines

Dne 22. 08. 24 v 9:44 Petr Valenta napsal(a):

Quoted 251 lines

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 51 lines · Bjorn Helgaas
Quoted 42 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

Quoted 7 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

Quoted 2 lines

 From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <[email protected]>
X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on
     aws-us-west-2-korg-lkml-1.web.codeaurora.org
Received: from smtp2.osuosl.org (smtp2.osuosl.org [140.211.166.133])
     (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
     (No client certificate requested)
     by smtp.lore.kernel.org (Postfix) with ESMTPS id 7319CC531DF
     for <[email protected]>; Thu, 22 Aug 2024 07:44:59 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
     by smtp2.osuosl.org (Postfix) with ESMTP id 2EE99404B8;
     Thu, 22 Aug 2024 07:44:59 +0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp2.osuosl.org ([127.0.0.1])
by localhost (smtp2.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id VRgkrPDlq_WW; Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
X-Comment: SPF check N/A for local connections - client-ip=140.211.166.34; helo=ash.osuosl.org; [email protected]; receiver=<UNKNOWN> DKIM-Filter: OpenDKIM Filter v2.11.0 smtp2.osuosl.org 53F64405BA
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=osuosl.org;
     s=default; t=1724312696;
     bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
     h=Date:To:References:From:In-Reply-To:Subject:List-Id:
List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe:
      Cc:From;
     b=ZIudOsHGSDoQvtekseiE4SUpOKofnvHlxj7aT3f7bLvqCDMOCfygsO6tctN23YgSh
xYqnq4yBSB4/JQ4v7Juyg0P/wqTcr+XFqhORTc2qBku9GCA+Y4wRKbRUeH4/AUNthL
cf/zG7uEOFEKz4YALwviQFqR5E+HW9gD+YnXahtGUVqYiTjB01HuESDZdYI5huiCLI
eHnQDw/SSwM1YmkjLzQgICjlxtIRVYjUL+shaltRg9f7t4otZa9bvrvLptzw5Mrfc0
GLvrNRmHckPFKEJOXgmIeQI40IOHckD3MX2dkQ2dQ0VCrkl9JIgtuSRuS3IpB1dr65
      TatTrq9Onm26w==
Received: from ash.osuosl.org (ash.osuosl.org [140.211.166.34])
     by smtp2.osuosl.org (Postfix) with ESMTP id 53F64405BA;
     Thu, 22 Aug 2024 07:44:56 +0000 (UTC)
Received: from smtp1.osuosl.org (smtp1.osuosl.org [140.211.166.138])
by ash.osuosl.org (Postfix) with ESMTP id 81E351BF322
for <[email protected]>; Thu, 22 Aug 2024 07:44:54
+0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
by smtp1.osuosl.org (Postfix) with ESMTP id 79A0C80A82
for <[email protected]>; Thu, 22 Aug 2024 07:44:54
+0000 (UTC)
X-Virus-Scanned: amavis at osuosl.org
Received: from smtp1.osuosl.org ([127.0.0.1])
by localhost (smtp1.osuosl.org [127.0.0.1]) (amavis, port 10024) with ESMTP
id m9sJJpu9kR7y for <[email protected]>;
Thu, 22 Aug 2024 07:44:53 +0000 (UTC)
Received-SPF: Pass (mailfrom) identity=mailfrom; client-ip=136.143.188.52;
helo=sender4-of-o52.zoho.com; [email protected];
receiver=<UNKNOWN> DMARC-Filter: OpenDMARC Filter v1.4.2 smtp1.osuosl.org 3674B80A59
DKIM-Filter: OpenDKIM Filter v2.11.0 smtp1.osuosl.org 3674B80A59
Received: from sender4-of-o52.zoho.com (sender4-of-o52.zoho.com
[136.143.188.52])
by smtp1.osuosl.org (Postfix) with ESMTPS id 3674B80A59
for <[email protected]>; Thu, 22 Aug 2024 07:44:51
+0000 (UTC)
ARC-Seal: i=1; a=rsa-sha256; t=1724312671; cv=none; d=zohomail.com; s=zohoarc; b=B0wnUG3UHEcTRfbjC9HSfLJG+WBnpU18yag7r0240QuMQMnP/cHcj9e4oJU2FgxRPLpt6OGnlZOiPNE2GUFgnkBzKBPwzxb7eTHFwW4P8cW+1IrIOQ6jZWd2rhOIyWcRKYMydfCbMPM04Z+RwKVyRlrLTYL5UDBYYKKHOG08Ikc=
ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com;
s=zohoarc; t=1724312671;
h=Content-Type:Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:MIME-Version:Message-ID:References:Subject:Subject:To:To:Message-Id:Reply-To;
bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=; b=VvZAc/xKVy85rZpNNCAwUwpxquk4r4Xw2QjZmePGlnINwOvJf6oilR9lqx2WDMezV20iKTW9f3dauO4jIjp363HOdh7P21UFfa66a0oK63RODo7IQMHSCqaCwAEoO1PKHfDfTMwz0/BShU1dt+nhtAeSeKwbG7G1qizCcoXTdjo= ARC-Authentication-Results: i=1; mx.zohomail.com;
dkim=pass  header.i=jevklidu.cz;
spf=pass  [email protected];
dmarc=pass header.from=<[email protected]>
Received: by mx.zohomail.com with SMTPS id 1724312669862808.3168476405893;
Thu, 22 Aug 2024 00:44:29 -0700 (PDT)
Message-ID: <[email protected]>
Date: Thu, 22 Aug 2024 09:44:22 +0200
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
To: Vitaly Lifshits <[email protected]>,
Bjorn Helgaas <[email protected]>, Dima Ruinskiy <[email protected]>,
Hui Wang <[email protected]>
References: <20240821145959.GA248604@bhelgaas>
<[email protected]>
Content-Language: cs-CZ, en-US
From: Petr Valenta <[email protected]>
In-Reply-To: <[email protected]>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
X-ZohoMailClient: External
X-Mailman-Original-DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt;
c=relaxed/relaxed; t=1724312671; s=zoho; d=jevklidu.cz; [email protected]; h=Message-ID:Date:Date:MIME-Version:Subject:Subject:To:To:Cc:Cc:References:From:From:In-Reply-To:Content-Type:Content-Transfer-Encoding:Message-Id:Reply-To;
bh=y3v3IIFARTszfLWu7n/j8Du29EOi4VTxMDP3GF4qp7E=;
b=RSnIQpqoQp2O3ExJNnw4fk9dlt8CX1T5sbtB6GflBDYejiRQJTcrU3zRHn3pRkFq
tm00/cgXr6pF6T5vJFttBkfrHtnRiPiE8cjqni5KsNxCyOXOwri6I5ARAmPcUj42eda
e/xHQX9E3ayXrWSBQDAsun3Ann63tcXQKwlT7ffI=
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dmarc=none (p=none dis=none)
header.from=jevklidu.cz
X-Mailman-Original-Authentication-Results: smtp1.osuosl.org;
dkim=pass (1024-bit key,
unprotected) header.d=jevklidu.cz [email protected]
header.a=rsa-sha256 header.s=zoho header.b=RSnIQpqo
Subject: Re: [Intel-wired-lan] ACPI IRQ storm with 6.10
X-BeenThere: [email protected]
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Intel Wired Ethernet Linux Kernel Driver Development
<intel-wired-lan.osuosl.org>
List-Unsubscribe: <https://lists.osuosl.org/mailman/options/intel-wired-lan>, <mailto:[email protected]?subject=unsubscribe>
List-Archive: <http://lists.osuosl.org/pipermail/intel-wired-lan/>
List-Post: <mailto:[email protected]>
List-Help: <mailto:[email protected]?subject=help>
List-Subscribe: <https://lists.osuosl.org/mailman/listinfo/intel-wired-lan>,
<mailto:[email protected]?subject=subscribe>
Cc: Linux regressions mailing list <[email protected]>,
"Rafael J. Wysocki" <[email protected]>, [email protected],
Linux kernel mailing list <[email protected]>,
"[email protected]" <[email protected]>,
Tony Nguyen <[email protected]>, Bjorn Helgaas <[email protected]>,
[email protected], Jiri Slaby <[email protected]>,
Len Brown <[email protected]>
Errors-To: [email protected]
Sender: "Intel-wired-lan" <[email protected]>

Dne 21. 08. 24 v 17:17 Vitaly Lifshits napsal(a):

Quoted 51 lines · Bjorn Helgaas
Quoted 42 lines

[+to Dima, Vitaly, Hui; beginning of thread at
https://lore.kernel.org/r/[email protected]]

Quoted 29 lines · Petr Valenta

Dne 20. 08. 24 v 23:30 Bjorn Helgaas napsal(a):

Quoted 5 lines · Petr Valenta
Quoted 4 lines

Dne 20. 08. 24 v 20:09 Bjorn Helgaas napsal(a):

Quoted 3 lines · Jiri Slaby
Quoted 2 lines · Jiri Slaby
Quoted 1 line

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

...

Quoted 2 lines

I'm at a loss.  You could try reverting the entire b2c289415b2b commit (patch for that is below).

This patch didn't help, so I reverted it back.

Quoted 11 lines

If that doesn't help, I guess you could try reverting the other commits Jiri mentioned:

    76a0a3f9cc2f e1000e: fix force smbus during suspend flow c93a6f62cb1b e1000e: Fix S0ix residency on corporate systems bfd546a552e1 e1000e: move force SMBUS near the end of enable_ulp function 6918107e2540 net: e1000e & ixgbe: Remove PCI_HEADER_TYPE_MFD duplicates 1eb2cded45b3 net: annotate writes on dev->mtu from ndo_change_mtu() b2c289415b2b e1000e: Remove redundant runtime resume for ethtool_ops 75a3f93b5383 net: intel: implement modern PM ops declarations

Quoted 3 lines

If you do this, I would revert 76a0a3f9cc2f, test, then revert c93a6f62cb1b in addition, test, then revert bfd546a552e1 in addition, etc.

I have created revert patches like this:
git format-patch --stdout -1 76a0a3f9cc2f | interdiff -q /dev/stdin \
/dev/null > revert_76a0a3f9cc2f.patch

Quoted 5 lines

I have applied revert_76a0a3f9cc2f.patch (rebuild and tested), then in addition revert_c93a6f62cb1b.patch and after applying revert_bfd546a552e1.patch irq storm didn't appear.

I have tested it with 3 subsequent reboots and in all those cases it was ok.

Thanks for all this testing.  It sounds like reverting all three of 76a0a3f9cc2f, c93a6f62cb1b, and bfd546a552e1 fixed the IRQ storm, but I'm not clear on the results of other situations.

Quoted 5 lines

It looks like c93a6f62cb1b could be reverted by itself because it's unrelated to 76a0a3f9cc2f and bfd546a552e1.  I added the authors of all three in case they have any insights.

Bjorn

Quoted 7 lines

I doubt that it is related to c93a6f62cb1b, I believe that is more probable to be related to the two other patches.

Apart from what I suggested in the other mailing thread (enabling e1000e debug and to test if it happens with a cable connected),

I suggest to try to apply this patch and see if it fixes the issue:

https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I have applied patch from link above and command bellow really doesn't start irq storm.

Quoted 9 lines
Quoted 4 lines

echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control

Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.

I have tested now how it behaves with kernel 6.9.9. There is a new finding. After running "echo 'auto' > /sys/bus/pci/devices/0000:00:1f.6/power/control" network over cable works but irq storm arrises. I have never tested this before because I don't use cable with this laptop at all. After unplugging cable irq storm is gone.

A possible workaround would be to turn off power control for the e1000e at the kernel level (if is it possible) so that utilities like powertop don't cause irq storm or broken network.

I would like to suggest the following for now as a triage:

1. revert the following patches:

   i.  e1000e: fix force smbus during suspend flow (76a0a3f9cc2f)

   ii.   e1000e: move force SMBUS near the end of enable_ulp function
(bfd546a552e1)

   iii. "e1000e: move force SMBUS from enable ulp function to avoid PHY
loss issue ( 861e8086029e)

Quoted 6 lines

2. apply this patch:
https://patchwork.ozlabs.org/project/intel-wired-lan/patch/[email protected]/

I expect this configuration not to have neither the IRQ storm nor the link indication issue (>"Problem is that after executing this command and plugging cable to ethernet port, kernel is not able to detect link (LED indicate link is on) so network over cable is not working.")

Link detection is fine now but irq storm is still present when cable is plugged. irq/9-acpi utilizes 85% of one cpu core.

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 2 lines
#20 lore raw
In reply to Jiri Slaby

[+to Tony, Przemek for e1000e questions; -cc Jesse]

Quoted 40 lines · Jiri Slaby
Quoted 1 line · Jiri Slaby

CC e1000e guys + Jesse (due to 75a3f93b5383) + Bjorn (due to b2c289415b2b)

Bjorn,

I am confused by these changes:
==========================================

(quoted diff) +2 -4
@@ -291,16 +288,13 @@ static int e1000_set_link_ksettings(struct net_device
 *net
 dev,
         * duplex is forced.
         */
        if (cmd->base.eth_tp_mdix_ctrl) {
-               if (hw->phy.media_type != e1000_media_type_copper) {
-                       ret_val = -EOPNOTSUPP;
-                       goto out;
-               }
+               if (hw->phy.media_type != e1000_media_type_copper)
+                       return -EOPNOTSUPP;
 
                if ((cmd->base.eth_tp_mdix_ctrl != ETH_TP_MDI_AUTO) &&
                    (cmd->base.autoneg != AUTONEG_ENABLE)) {
                        e_err("forcing MDI/MDI-X state is not supported when

lin
k speed and/or duplex are forced\n");

(quoted diff) +1 -3
-                       ret_val = -EINVAL;
-                       goto out;
+                       return -EINVAL;
                }
        }
@@ -347,7 +341,6 @@ static int e1000_set_link_ksettings(struct net_device
 *netde
 v,
        }
 
 out:
-       pm_runtime_put_sync(netdev->dev.parent);
        clear_bit(__E1000_RESETTING, &adapter->state);
        return ret_val;
Quoted 5 lines

}
==========================================

So no more clear_bit(__E1000_RESETTING in the above fail paths. Is that intentional?

I don't remember if it was intentional, but the use of __E1000_RESETTING is a bit subtle and I don't know what is correct.

Here's how it was used before I changed it with b2c289415b2b, i.e., in
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/net/ethernet/intel/e1000e/ethtool.c?id=39f59c72ad3a:

  e1000_set_link_ksettings(...)
  {
    if (hw->phy.ops.check_reset_block(hw)) {
      ret_val = -EINVAL;
      goto out;
    }
    while (test_and_set_bit(__E1000_RESETTING, &adapter->state))
      usleep_range(1000, 2000);
    if (err) {
      ret_val = -EINVAL;
      goto out;
    }
    ...
  out:
    clear_bit(__E1000_RESETTING, &adapter->state);
  }

In this case, we *always* clear __E1000_RESETTING, even if we bail out before the test_and_set_bit(__E1000_RESETTING).

It makes sense to me that we clear __E1000_RESETTING after we've set it via test_and_set_bit() because we know it was set *here*.

But it seems wrong to me that we clear __E1000_RESETTING even when we haven't done the test_and_set_bit() because it may have been set by a concurrent thread executing a different operation.

  e1000_set_ringparam(...)
  {
    if ((ring->rx_mini_pending) || (ring->rx_jumbo_pending))
      return -EINVAL;
    while (test_and_set_bit(__E1000_RESETTING, &adapter->state))
      usleep_range(1000, 2000);
    err = e1000e_setup_tx_resources(...);
    if (err)
      goto out;
    ...
  out:
    clear_bit(__E1000_RESETTING, &adapter->state);
  }

But here, we *don't* clear __E1000_RESETTING if we bail out before the test_and_set_bit(__E1000_RESETTING). This seems like the correct behavior.

In the e1000 driver (not the e1000e driver), e1000_set_link_ksettings() does *not* clear __E1000_RESETTING unless it has already done the test_and_set_bit().

b2c289415b2b changed e1000e to work that way, too.

FWIW, 3ef672ab1862 ("e1000e: ethtool unnecessarily takes device out of RPM suspend") changed e1000e e1000_set_link_ksettings() to clear __E1000_RESETTING even when bailing out before the test_and_set_bit(). That part of 3ef672ab1862 looks possibly buggy to me.

Bjorn