On 1/22/2017 2:08 PM, Sinan Kaya wrote:
> [snip]
>
>>>
>>> Bjorn
>>>
>>
>> / # lspci
>> 03:00:00.0 Class 0604: 17cb:0401
>> 03:01:00.0 Class 0200: 8086:1572
>> 03:01:00.1 Class 0200: 8086:1572
>>
>>
>> It looks like ASPM is only getting configured for the root port following
>> insertion.
>>
>> Removal:
>> [ 32.142474] pci 0003:01:00.1: pcie_aspm_exit_link_state:662
>> [ 34.142378] pci 0003:01:00.0: pcie_aspm_exit_link_state:659
>>
>> Insertion:
>> [ 35.910342] pcieport 0003:00:00.0: pcie_aspm_init_link_state:565
>> [ 35.931227] pcieport 0003:00:00.0: pcie_aspm_init_link_state:617
>>
it looks like this is what is happening.
hotplug insertion
pcie_aspm_init_link_state
pcie_config_aspm_path
default policy, do nothing
[ 128.691521] pcieport 0003:00:00.0: pcie_aspm_init_link_state:570
[ 128.712402] pcie_config_aspm_link:439 state=0
430 static void pcie_config_aspm_link(struct pcie_link_state *link, u32 state)
431 {
432 u32 upstream = 0, dwstream = 0;
433 struct pci_dev *child, *parent = link->pdev;
434 struct pci_bus *linkbus = parent->subordinate;
435
436 /* Nothing to do if the link is already in the requested state */
437 state &= (link->aspm_capable & ~link->aspm_disable);
438 if (link->aspm_enabled == state) {
439 pr_info("%s:%d state=%d\n", __func__, __LINE__, state);
440 return;
441 }
Should we add a new policy that comes through the hotplug insertion path?
Regarding your question on why this is not being done in configure_device,
probably the reason is that L0s and L1 needs to be enabled in a specific
order before enabling on the root port.
That must be the reason for attaching this code behavior.
I think the real problem is the assumption default policy on the hotplug
insertion path.
--
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
