Linux 6.10 Preps For "When Things Go Seriously Wrong" On Bigger Servers

[phoronix]

Phoronix: Linux 6.10 Preps For "When Things Go Seriously Wrong" On Bigger Servers

While machine check exception (MCE) events tend to be uncommon, a change made by Intel engineers is accommodating the ability in the Linux kernel to store more machine check records for "when things go seriously wrong" on increasingly high core count servers...

https://www.phoronix.com/news/Linux-6.10-Larger-CPUs-MCE

[schmidtbag]

Considering 80 still isn't enough for many modern servers, why didn't they expand it further?

[Mahboi]

Considering that Sierra Forest is going to be something like 288 e-cores, I'm not surprised.
And to be fair AMD Zen 5 is also going to throw 192 Zen5 compact cores so that's not much better versus a 20 error buffer.

[sbin]

Considering 80 still isn't enough for many modern servers, why didn't they expand it further?

You misread, 80 was the old limit. Reading the commit diff, the new limit is determined by multiplying CPU count by 2.

[strtj]

Considering 80 still isn't enough for many modern servers, why didn't they expand it further?

Did you read the article? It's dynamic now, with a minimum size of 80 but more if the system warrants.

[yump]

I wonder why it's N*num_cpus instead of each CPU having its own buffer? Wouldn't want to compound the problem with a locking storm when 'everything goes seriously wrong'.