Warner Losh

> @LeChasseur good point. If there is no permission to modify the work and to distribute the work in modified form, this license is not conformant to the open source...

I'd document when / how existing licenses are updated: 1. To correct errors in original submission (eg, the wrong version was submitted or a slightly corrupted version was submitted and...

> I think this is more process-related stuff, than historical, but some food for thought here: > > * not sure what you mean by 1 Have we ever had...

This looks to be a hardware problem and/or subtle incompatibility with the ahci driver. Commands are timing out, including the most basic command identify. Assuming the hard disk's firmware isn't...

In a nutshell, what's happening... We get a timeout for the drive, we decide (for reasons I still don't know) that the drive is AFU and we should drop it......

So, threading through the code, we're invalidating the periph for some reason: 1. We tried to send an INQUIRY command to the device that failed in an unexpected way. We...

I can give patches to the kernel to provide more data if someone can recreate it.

The malloc from a bad place thing is about 4 things downstream of the cause, btw. Even if we fixed it, we'd still have the periph departing and arriving (from...

It might help (it wouldn't hurt to try it). If that's it, it would fix the root cause... But the problem is we're generating a AC_LOST_DEVICE event (where I'm still...

> This panic is very reproducible for me Maybe you should just send me the system it happens on (wink wink wink). > So I’m happy to try patches. Well,...