12 Apr
2024
12 Apr
'24
8:18 p.m.
On 4/12/24 16:10, Sean Anderson wrote:
On 4/12/24 15:46, Pavel Machek wrote:
Sean Anderson sean.anderson@linux.dev soc: fsl: qbman: Use raw spinlock for cgr_lock
As we don't have commit ef2a8d5478b9 ("net: dpaa: Adjust queue depth on rate change") in 4.19, we should not really need this.
Sorry, I missed this the first time around, but this bug will still occur (just much more rarely) without this commit.
--Sean
Plus, 10msec
under raw spinlock is quite evil, such kernel is not realtime any more. We should not be doing that.
Patches welcome :)
At some point I will try and revisit this.
--Sean