Conversation

Vlastimil Babka

STOP DOING LOCKS

- CPUS WERE NOT SUPPOSED TO BE WAITING IN VAIN

- YEARS OF SPINNING yet NO REAL-USE FOUND for mutual exclusion and critical SECTIONS

- Wanted to protect from races anyway for a laugh? We had a tool for that. It was called "IRQ DISABLE".

- "Yes please let me SPIN optimistically for a while. Please let me SLEEP now instead." - Statements dreamed up by the utterly Deranged

LOOK at what the Kernel Developers have been demanding your Respect for all this time, with all the local_locks and rt-mutexes we built for them
(This is REAL Code, done by REAL Developers)

- pcp_spin_trylock_irqsave() - ?????
- if (USE_LOCKLESS_FAST_PATH()) lockdep_assert_irqs_disabled();???????
- lockdep_assert_held(this_cpu_ptr(&s->cpu_slab->lock)) ?????????????????

"Hello I would like this_cpu_try_cmpxchg128(&apples) please"

They have played us for absolute fools.
3
25
45

@vbabka From the earliest days of SMP mainboards the kernel was an agent of Big Lock, even calling it the Big Kernel Lock.

But that's nothing compared to what user space is doing. I demand an investigation into the Lock Files. I want to know what's inside them, but most are suspiciously empty, a clear sign of a cover-up.

0
3
1
@wagi yeah it's an obnoxious micromanagement. Just trust everyone to know their lifetime.
0
0
4

@vbabka consider doing some merch with this

0
0
1