"Under a highly specific and detailed set of internal timing conditions, the MWAIT instruction may cause a
thread to hang in SMT (Simultaneous Multithreading) Mode."
Since this is a nofix for AMD, this must be fixed by the OS. The AMD document states:
"Suggested Workaround
System software may contain the workaround for this erratum."
So, unfortunately AMD will never solve that hardware error. I suggest they change the first two letters of their company name ;)
This in fact leaves every linux system with a Ryzen CPU instable, so this should be considered serious.
I didn't read a comment from any kernel developer in the history of this bug report. Did I overlook something?
Question: Will the Linux kernel ever implement a workaround (as suggested by AMD??
I think this bug is mostly about the official AMD Errata 1109 "MWAIT Instruction May Hang a Thread".
This is described here ( https:/ /developer. amd.com/ wp-content/ resources/ 55449_1. 12.pdf ) as:
"Under a highly specific and detailed set of internal timing conditions, the MWAIT instruction may cause a
thread to hang in SMT (Simultaneous Multithreading) Mode."
Since this is a nofix for AMD, this must be fixed by the OS. The AMD document states:
"Suggested Workaround
System software may contain the workaround for this erratum."
So, unfortunately AMD will never solve that hardware error. I suggest they change the first two letters of their company name ;)
This in fact leaves every linux system with a Ryzen CPU instable, so this should be considered serious.
I didn't read a comment from any kernel developer in the history of this bug report. Did I overlook something?
Question: Will the Linux kernel ever implement a workaround (as suggested by AMD??