igc driver causes suspend to fail if powersave is enabled

igc driver causes suspend to fail if powersave is enabled

[Srikar Dronamraju]

Hi,

On Lenovo ThinkPad P15 Gen 2i with I225-LM Ethernet card running Fedora 37
kernel 6.1.13-200.fc37.x86_64, running powertop --auto-tune causes suspend
and reboot to fail. Once suspend fails, networking stops working even for
wireless. Infact as a normal user, I cant even start a sudo session after
trying to suspend.

Reboot/shutdown fails since Network-Manager cant be killed. (More below)
The only option left is for the system to be hard-reboot.

I finally found that the problem was because enabling auto or powersave on
ethernet card was causing this issue. i.e

echo 'auto' > '/sys/bus/pci/devices/0000:0b:00.0/power/control'

Pls do note, I was not using ethernet, I was just using wireless even before
suspend.

lspci -v reports
0b:00.0 Ethernet controller: Intel Corporation Ethernet Controller I225-LM (rev 03)
	Subsystem: Lenovo Device 22d8
	Flags: fast devsel, IRQ 16, IOMMU group 17
	Memory at be300000 (32-bit, non-prefetchable) [size=1M]
	Memory at be400000 (32-bit, non-prefetchable) [size=16K]
	Capabilities: [40] Power Management version 3
	Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
	Capabilities: [70] MSI-X: Enable- Count=5 Masked-
	Capabilities: [a0] Express Endpoint, MSI 00
	Capabilities: [100] Advanced Error Reporting
	Capabilities: [140] Device Serial Number 88-a4-c2-ff-ff-5f-09-88
	Capabilities: [1c0] Latency Tolerance Reporting
	Capabilities: [1f0] Precision Time Measurement
	Capabilities: [1e0] L1 PM Substates
	Kernel modules: igc

When suspend fails, we see messages like this

kernel: Freezing of tasks failed after 20.001 seconds (3 tasks refusing to freeze, wq_busy=0):
kernel: task:NetworkManager  state:D stack:0     pid:2014  ppid:1      flags:0x00004006
kernel: Call Trace:
kernel:  <TASK>
kernel:  __schedule+0x35f/0x1360
kernel:  ? asm_sysvec_reschedule_ipi+0x16/0x20
kernel:  schedule+0x5d/0xe0
kernel:  schedule_preempt_disabled+0x14/0x30
kernel:  __mutex_lock.constprop.0+0x390/0x6e0
kernel:  ? __cond_resched+0x1c/0x30
kernel:  igc_resume+0xfc/0x1d0 [igc]
kernel:  ? pci_pm_restore_noirq+0xc0/0xc0
kernel:  __rpm_callback+0x41/0x170
kernel:  rpm_callback+0x35/0x70
kernel:  ? pci_pm_restore_noirq+0xc0/0xc0
kernel:  rpm_resume+0x5bb/0x800
kernel:  __pm_runtime_resume+0x47/0x80
kernel:  dev_ethtool+0x124/0x2eb0
kernel:  ? avc_has_extended_perms+0x22a/0x520
kernel:  ? inet_ioctl+0xd8/0x1e0
kernel:  dev_ioctl+0x156/0x520
kernel:  sock_do_ioctl+0xda/0x120
kernel:  sock_ioctl+0xed/0x330
kernel:  ? security_file_ioctl+0x39/0x60
kernel:  __x64_sys_ioctl+0x8d/0xd0
kernel:  do_syscall_64+0x58/0x80
kernel:  ? __x64_sys_ioctl+0xa8/0xd0
kernel:  ? syscall_exit_to_user_mode+0x17/0x40
kernel:  ? do_syscall_64+0x67/0x80
kernel:  ? do_syscall_64+0x67/0x80
kernel:  ? do_syscall_64+0x67/0x80
kernel:  ? do_syscall_64+0x67/0x80
kernel:  entry_SYSCALL_64_after_hwframe+0x63/0xcd
kernel: RIP: 0033:0x7ff964b93d6f
kernel: RSP: 002b:00007ffc96c6f540 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
kernel: RAX: ffffffffffffffda RBX: 00005608c0ba8040 RCX: 00007ff964b93d6f
kernel: RDX: 00007ffc96c6f640 RSI: 0000000000008946 RDI: 000000000000001e
kernel: RBP: 00007ffc96c6f7c0 R08: 0000000000000000 R09: 0000000000000000
kernel: R10: 0000000000000021 R11: 0000000000000246 R12: 0000000000000000
kernel: R13: 00007ffc96c6f640 R14: 00007ffc96c6f620 R15: 00007ffc96c6f620
kernel:  </TASK>

We have seen similar problems even with people running other distros like
Arch Linux and RHEL 8 (which is 4.18 based)distros. So this is not distro
specific or kernel specific too.

Even installing tlp package from https://linrunner.de/tlp/index.html causes
similar issues.

-- 
Thanks and Regards
Srikar Dronamraju

Re: [Intel-wired-lan] igc driver causes suspend to fail if powersave is enabled

[Srikar Dronamraju]

* Neftin, Sasha <sasha.neftin@intel.com> [2023-03-15 21:04:07]:

> On 3/15/2023 09:20, Srikar Dronamraju wrote:
> > Hi,
> >
> > On Lenovo ThinkPad P15 Gen 2i with I225-LM Ethernet card running Fedora 37
> > kernel 6.1.13-200.fc37.x86_64, running powertop --auto-tune causes suspend
> > and reboot to fail. Once suspend fails, networking stops working even for
> > wireless. Infact as a normal user, I cant even start a sudo session after
> > trying to suspend.
> Hello,
> Is the i225 card on board (under PCH downstream port) or connected via
> thunderbolt?
> Did you see the same problem on other platforms?
> I will forward this inquiry to our PAE.>

I dont think its connected to thunderbolt. But I am not sure.
So, can you let me know how do I figure this out?

Some additional information if that helps

$ lspci -vt
-[0000:00]-+-00.0  Intel Corporation 11th Gen Core Processor Host Bridge/DRAM Registers
           +-01.0-[01]--+-00.0  NVIDIA Corporation TU117GLM [T1200 Laptop GPU]
           |            \-00.1  NVIDIA Corporation Device 10fa
           +-02.0  Intel Corporation TigerLake-H GT1 [UHD Graphics]
           +-04.0  Intel Corporation TigerLake-LP Dynamic Tuning Processor Participant
           +-06.0-[04]----00.0  Micron Technology Inc Device 5407
           +-07.0-[20-49]--
           +-07.1-[50-79]--
           +-0d.0  Intel Corporation Tiger Lake-H Thunderbolt 4 USB Controller
           +-0d.2  Intel Corporation Tiger Lake-H Thunderbolt 4 NHI #0
           +-14.0  Intel Corporation Tiger Lake-H USB 3.2 Gen 2x1 xHCI Host Controller
           +-14.2  Intel Corporation Tiger Lake-H Shared SRAM
           +-16.0  Intel Corporation Tiger Lake-H Management Engine Interface
           +-16.3  Intel Corporation Device 43e3
           +-1c.0-[1c]--
           +-1c.4-[0b]----00.0  Intel Corporation Ethernet Controller I225-LM
           +-1c.5-[09]----00.0  Intel Corporation Wi-Fi 6 AX210/AX211/AX411 160MHz
           +-1c.7-[0a]----00.0  Genesys Logic, Inc GL9755 SD Host Controller
           +-1f.0  Intel Corporation Device 4389
           +-1f.3  Intel Corporation Tiger Lake-H HD Audio Controller
           +-1f.4  Intel Corporation Tiger Lake-H SMBus Controller
           \-1f.5  Intel Corporation Tiger Lake-H SPI Controller

I can use ethernet even after removing thunderbolt module.

Even when loaded, o/p of
$ lsmod | grep -e thunderbolt -e igc
thunderbolt           397312  0
igc                   163840  0

I dont have any other platform with the same Ethernet Controller device.

--
Thanks and Regards
Srikar Dronamraju