* 答复: [SPAM] [????] Re: [PATCH] x86/math64: handle #DE in mul_u64_u64_div_u64()
@ 2025-07-22 2:09 Li,Rongqing
0 siblings, 0 replies; only message in thread
From: Li,Rongqing @ 2025-07-22 2:09 UTC (permalink / raw)
To: David Laight, Oleg Nesterov
Cc: Ingo Molnar, Peter Zijlstra, Thomas Gleixner, Borislav Petkov,
Dave Hansen, H. Peter Anvin, Steven Rostedt, linux-kernel, x86
> Even for the scheduler that recently trapped here, the 64bit sum was about the
> wrap and the very small value returned then almost certainly wouldn't have
> been handled correctly.
> (So that code needs fixing anyway to avoid the overflow.)
There are currently two bugs. The one Oleg is fixing is a crash bug in the x86-64 mul_u64_u64_div_u64() function caused by potential overflow. The mul_u64_u64_div_u64() function is used not only in cputime.c ,but also widely used within drivers.
The other one is a thread group statistics time overflow bug. If the overflow of mul_u64_u64_div_u64() can be fixed, it at least doesn't crash the system anymore. However, regarding the accuracy of the timing statistics, this bug might be more harder to fix.
-Li
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2025-07-22 2:10 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-07-22 2:09 答复: [SPAM] [????] Re: [PATCH] x86/math64: handle #DE in mul_u64_u64_div_u64() Li,Rongqing
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®