Re: [E1000-devel] 82754L spontaneous freeze networking woes continue in 2.6.37

[Allan, B. W <bru...@in...>]

>-----Original Message-----
>From: Nix [mailto:ni...@es...]
>Sent: Monday, January 31, 2011 2:43 PM
>To: Allan, Bruce W
>Cc: e10...@li...
>Subject: Re: [E1000-devel] 82754L spontaneous freeze networking woes continue in
>2.6.37
>
>On 31 Jan 2011, ni...@es... stated:
>
>> On 31 Jan 2011, Bruce W. Allan said:
>>> Have you tried booting with pcie_aspm=off kernel parameter?
>>
>> I didn't know that parameter existe. Added, will reboot shortly: let us
>> see what happens. :)
>
>No change:
>
>LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes Disabled- Retrain- CommClk+
>        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
>
>LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes Disabled- Retrain- CommClk+
>        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
>
>Boot messages include:
>
>[    0.000000] PCIe ASPM is disabled
>[    2.132444] e1000e 0000:03:00.0: Disabling ASPM L0s
>[    2.293944] e1000e 0000:02:00.0: Disabling ASPM L0s
>[    8.489378] e1000e 0000:02:00.0: Disabling ASPM  L1
>
>(the latter is on the gigabit link).
>
>Either lspci is lying to me, or the kernel's attempts to disable ASPM
>are doing nothing at all.
>
>I will find out soon enough which is true, as I'm no longer doing the
>continuous pingflood, so if ASPM is on (or the problem is somewhere
>else), the card will hang again...

Because lspci simply reads the PCI configuration space (IIRC), I doubt it
is reporting incorrect information.  The e1000e driver uses the kernel
API to disable ASPM (when CONFIG_PCIEASPM is enabled in the kernel config
otherwise it writes directly to the PCI configuration space to disable
ASPM).  Assuming your kernel config has CONFIG_PCIEASPM enabled, my guess
at this point would be there is something broken in the kernel.  With ASPM
L0s enabled, the 82574 (and other parts supported by the driver) will most
definitely have issues, so we need to find out what is broke and fix it.