> On 7 Jun 2021, at 14:50, Jason Gunthorpe <jgg@xxxxxxxxxx> wrote:
>
> On Mon, Jun 07, 2021 at 02:39:45PM +0200, Greg KH wrote:
>> On Mon, Jun 07, 2021 at 09:14:11AM -0300, Jason Gunthorpe wrote:
>>> On Mon, Jun 07, 2021 at 12:25:03PM +0200, Greg KH wrote:
>>>> On Mon, Jun 07, 2021 at 11:17:35AM +0300, Leon Romanovsky wrote:
>>>>> From: Jason Gunthorpe <jgg@xxxxxxxxxx>
>>>>>
>>>>> This code is trying to attach a list of counters grouped into 4 groups to
>>>>> the ib_port sysfs. Instead of creating a bunch of kobjects simply express
>>>>> everything naturally as an ib_port_attribute and add a single
>>>>> attribute_groups list.
>>>>>
>>>>> Remove all the naked kobject manipulations.
>>>>
>>>> Much nicer.
>>>>
>>>> But why do you need your counters to be atomic in the first place? What
>>>> are they counting that requires this?
>>>
>>> The write side of the counter is being updated from concurrent kernel
>>> threads without locking, so this is an atomic because the write side
>>> needs atomic_add().
>>
>> So the atomic write forces a lock :(
>
> Of course, but a single atomic is cheaper than the double atomic in a
> full spinlock.
>
>>> Making them a naked u64 will cause significant corruption on the write
>>> side, and packet counters that are not accurate after quiescence are
>>> not very useful things.
>>
>> How "accurate" do these have to be?
>
> They have to be accurate. They are networking packet counters. What is
> the point of burning CPU cycles keeping track of inaccurate data?
Consider a CPU with a 32-bit wide datapath to memory, which reads and writes the most significant 4-byte word first:
Memory CPU1 CPU2
MSW LSW MSW LSW MSW LSW
0x0 0xffffffff
0x0 0xffffffff 0x0
0x0 0xffffffff 0x0 0xffffffff
0x0 0xffffffff 0x1 0x0 cpu1 has incremented its register
0x1 0xffffffff 0x1 0x0 cpu1 has written msw
0x1 0xffffffff 0x1 0x0 0x1 cpu2 has read msw
0x1 0xffffffff 0x1 0x0 0x1 0xffffffff
0x1 0x0 0x1 0x0 0x2 0x0
0x2 0x0 0x1 0x0 0x2 0x0
0x2 0x0 0x1 0x0 0x2 0x0
I would say that 0x200000000 vs. 0x100000001 is more than inaccurate!
>
>> And have you all tried them?
>
> I've used them over the years. This stuff is something like 15 years
> old now.
We are using them.
Thxs, Håkon
>> I'm pushing back here as I see a lot of atomics used for debugging
>> statistics for no good reason all over the place. Especially when
>> userspace just does not care.
>
> If userspace doesn't care then just delete the counter entirely.
>
> Reporting a wrong/misleading debugging counter data sounds just
> horrible to me.
>
> What good is any debug result you get from the counter if it has to be
> questioned because the counter is allowed to be wrong?
>
> ""The sender says it sent 7 packets, but the receivers debug counter
> reports only 6! I guess my bug is a lost packet in the network.""
>
> Jason
