(In reply to Nayan Deshmukh from comment #32)
> Created attachment 128064 [details]
> dmesg_with_debug_options

These logs do not show the "resume failed" messages that appeared in the
original logs. Could you try to trigger this?

The logs also shows no \_SB.PCI0.RP05.PC05._ON (or _OFF) method invocations, so
pcie_port_pm=off (or force) is likely not going to make a difference here.

Btw, to increase the kernel log size from 256K you could pass log_buf_len=1M to
your cmdline. Could you try obtaining a log where the "resume failed" messages
are shown?

These messages also look suspicious, any idea what it could be? (Alex?)
[  221.378665] VI should always have 2 performance levels
[  221.380787]  failed to send message 18a ret is 255 
[  221.380794]  failed to send pre message 145 ret is 255 
[  221.621346]  failed to send message 146 ret is 0 
[  221.862589]  failed to send pre message 148 ret is 0 

They also appear in bug 97240 and bug 97548.