Does MCA handler work correctly?

Hello,

does MCA handler of 2.4.18 work correctly on multiple processors?

While looking into mca.c and mca_asm.S, they have single static stack area for MCA handling, but I could not find synchronization code between processors.

It seems that when MCA occurs on multiple processors, they may break stacks each other.

Best regards.
--
NOMURA, Jun'ichi <[email protected], [email protected]> HPC Operating System Group, 1st Computers Software Division, Computers Software Operations Unit, NEC Solutions.

Good observation - you are absolutely correct that the current implementation is not SMP-safe. The MCA handler is not functional yet - currently, if you do get an MCA the OS handler will return to SAL with a request to reboot the machine, and no attempt is made to recover. (The logging of hardware-corrected errors is functional, however.)

There is an effort in progress to code the OS recovery for selected MCAs. This effort is part of the Atlas project:

http://atlas-64.sourceforge.net

Thanks,
Jenna

-----Original Message-----
From: [email protected] [mailto:[email protected]]
Sent: Tuesday, June 11, 2002 5:01 AM
To: [email protected]
Subject: [Linux-ia64] Does MCA handler work correctly?

Hello,

does MCA handler of 2.4.18 work correctly on multiple processors?

While looking into mca.c and mca_asm.S, they have single static stack area for MCA handling, but I could not find synchronization code between processors.

It seems that when MCA occurs on multiple processors, they may break stacks each other.

Best regards.
--
NOMURA, Jun'ichi <[email protected], [email protected]> HPC Operating System Group, 1st Computers Software Division, Computers Software Operations Unit, NEC Solutions.

_______________________________________________
Linux-IA64 mailing list
[email protected]
http://lists.linuxia64.org/lists/listinfo/linux-ia64