Fenghua,
On Tue, Apr 13 2021 at 23:40, Fenghua Yu wrote:
> On Mon, Apr 12, 2021 at 09:15:08AM +0200, Thomas Gleixner wrote:
>> Aside of that why are you trying to make this throttling in any way
>> accurate? It does not matter at all, really. Limit reached, put it to
>> sleep for some time and be
Hi, Thomas,
On Mon, Apr 12, 2021 at 09:15:08AM +0200, Thomas Gleixner wrote:
> On Sat, Apr 03 2021 at 01:04, Fenghua Yu wrote:
> > On Sat, Mar 20, 2021 at 01:42:52PM +0100, Thomas Gleixner wrote:
> >> On Fri, Mar 19 2021 at 22:19, Fenghua Yu wrote:
> >> And even with throttling the injection rate
On Sat, Apr 03 2021 at 01:04, Fenghua Yu wrote:
> On Sat, Mar 20, 2021 at 01:42:52PM +0100, Thomas Gleixner wrote:
>> On Fri, Mar 19 2021 at 22:19, Fenghua Yu wrote:
>> And even with throttling the injection rate further down to 25k per
>> second the impact on the workload is still significant in
Hi, Thomas,
On Sat, Mar 20, 2021 at 01:42:52PM +0100, Thomas Gleixner wrote:
> On Fri, Mar 19 2021 at 22:19, Fenghua Yu wrote:
> > On Fri, Mar 19, 2021 at 10:30:50PM +0100, Thomas Gleixner wrote:
> >> > +if (sscanf(arg, "ratelimit:%d", ) == 1 && ratelimit >
> >> > 0) {
> >> > +
Hi, Thomas,
On Sat, Mar 20, 2021 at 02:57:52PM +0100, Thomas Gleixner wrote:
> On Sat, Mar 20 2021 at 02:01, Thomas Gleixner wrote:
>
> > On Fri, Mar 19 2021 at 21:50, Tony Luck wrote:
> >>> What is the justifucation for making this rate limit per UID and not
> >>> per task, per process or
On Sat, Mar 20 2021 at 02:01, Thomas Gleixner wrote:
> On Fri, Mar 19 2021 at 21:50, Tony Luck wrote:
>>> What is the justifucation for making this rate limit per UID and not
>>> per task, per process or systemwide?
>>
>> The concern is that a malicious user is running a workload that loops
>>
On Fri, Mar 19 2021 at 22:19, Fenghua Yu wrote:
> On Fri, Mar 19, 2021 at 10:30:50PM +0100, Thomas Gleixner wrote:
>> > + if (sscanf(arg, "ratelimit:%d", ) == 1 && ratelimit > 0) {
>> > + bld_ratelimit = ratelimit;
>>
>> So any rate up to INTMAX/s is valid here, right?
>
> Yes. I don't
On Fri, Mar 19 2021 at 21:50, Tony Luck wrote:
>> What is the justifucation for making this rate limit per UID and not
>> per task, per process or systemwide?
>
> The concern is that a malicious user is running a workload that loops
> obtaining the buslock. This brings the whole system to its
Hi, Thomas,
On Fri, Mar 19, 2021 at 10:30:50PM +0100, Thomas Gleixner wrote:
> On Sat, Mar 13 2021 at 05:49, Fenghua Yu wrote:
> > Change Log:
> > v5:
> > Address all comments from Thomas:
> > - Merge patch 2 and patch 3 into one patch so all "split_lock_detect="
> > options are processed in
> What is the justifucation for making this rate limit per UID and not
> per task, per process or systemwide?
The concern is that a malicious user is running a workload that loops
obtaining the buslock. This brings the whole system to its knees.
Limiting per task doesn't help. The user can
On Sat, Mar 13 2021 at 05:49, Fenghua Yu wrote:
> Change Log:
> v5:
> Address all comments from Thomas:
> - Merge patch 2 and patch 3 into one patch so all "split_lock_detect="
> options are processed in one patch.
What? I certainly did not request that. I said:
"Why is this seperate and an
Bus locks degrade performance for the whole system, not just for the CPU
that requested the bus lock. Two CPU features "#AC for split lock" and
"#DB for bus lock" provide hooks so that the operating system may choose
one of several mitigation strategies.
#AC for split lock is already implemented.
12 matches
Mail list logo
