From: Thomas Gleixner <tglx@linutronix.de>
To: Alexei Lozovsky <me@ilammy.net>, Alexey Dobriyan <adobriyan@gmail.com>
Cc: Christoph Lameter <cl@linux.com>,
LKML <linux-kernel@vger.kernel.org>,
linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH 0/7] proc/stat: Maintain monotonicity of "intr" and "softirq"
Date: Tue, 14 Sep 2021 16:11:21 +0200 [thread overview]
Message-ID: <87y27zb62e.ffs@tglx> (raw)
In-Reply-To: <44F84890-521F-4BCA-9F48-B49D2C8A9E32@ilammy.net>
On Sun, Sep 12 2021 at 21:37, Alexei Lozovsky wrote:
> On Sun, Sep 12, 2021, at 18:30, Alexey Dobriyan wrote:
>> How about making everything "unsigned long" or even "u64" like NIC
>> drivers do?
>
> I see some possible hurdles ahead:
>
> - Not all architectures have atomic operations for 64-bit values
This is not about atomics.
> All those "unsigned int" counters are incremented with __this_cpu_inc()
> which tries to use atomics if possible. Though, I'm not quite sure
It does not use atomics. It's a CPU local increment.
> how this works for read side which does not seem to use atomic reads
> at all. I guess, just by the virtue of properly aligned 32-bit reads
> being atomic everywhere? If that's so, I think widening counters to
> 64 bits will come with an asterisk.
The stats are accumulated racy, i.e. the interrupt might be handled and
one of the per cpu counters or irq_desc->tot_count might be incremented
concurrently.
On 32bit systems a 32bit load (as long as the compiler does not emit
load tearing) is always consistent even when there is a concurrent
increment going on. It either gets the old or the new value.
A 64bit read on a 32bit system is always two loads which means that a
concurrent increment will make it possible to observe a half updated
value. And no, you can't play reread tricks here without adding barriers
on weakly ordered architectures.
> - We'll need to update all counters to be 64-bit.
>
> Like, *everyone*. Every field that gets summed up needs to be 64-bit
> (or else wrap-arounds will be incorrect). Basically every counter in
> every irq_cpustat_t will need to become twice as wide. If that's
> a fine price to pay for accurate, full-width counters...
The storage size should not be a problem.
> So right now I don't see why it shouldn't be doable in theory.
So much for the theory :)
Thanks,
tglx
next prev parent reply other threads:[~2021-09-14 14:11 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-10 8:53 /proc/stat interrupt counter wrap-around Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 0/7] proc/stat: Maintain monotonicity of "intr" and "softirq" Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 1/7] genirq: Use unsigned int for irqs_sum Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 2/7] powerpc/irq: arch_irq_stat_cpu() returns unsigned int Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 3/7] x86/irq: " Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 4/7] x86/irq: arch_irq_stat() " Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 5/7] proc/stat: Use unsigned int for "intr" sum Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 6/7] proc/stat: Use unsigned int for "softirq" sum Alexei Lozovsky
2021-09-11 3:48 ` [PATCH 7/7] docs: proc.rst: stat: Note the interrupt counter wrap-around Alexei Lozovsky
2021-09-11 3:59 ` Randy Dunlap
2021-09-12 9:30 ` [PATCH 0/7] proc/stat: Maintain monotonicity of "intr" and "softirq" Alexey Dobriyan
2021-09-12 12:37 ` Alexei Lozovsky
2021-09-14 14:11 ` Thomas Gleixner [this message]
2021-09-15 4:24 ` Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 00/12] " Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 01/12] genirq: Use READ_ONCE for IRQ counter reads Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 02/12] genirq: Use unsigned long for IRQ counters Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 03/12] powerpc/irq: Use READ_ONCE for IRQ counter reads Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 04/12] powerpc/irq: Use unsigned long for IRQ counters Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 05/12] powerpc/irq: Use unsigned long for IRQ counter sum Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 06/12] x86/irq: Use READ_ONCE for IRQ counter reads Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 07/12] x86/irq: Use unsigned long for IRQ counters Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 08/12] x86/irq: Use unsigned long for IRQ counters more Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 09/12] x86/irq: Use unsigned long for IRQ counter sum Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 10/12] proc/stat: Use unsigned long for "intr" sum Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 11/12] proc/stat: Use unsigned long for "softirq" sum Alexei Lozovsky
2021-09-15 17:58 ` [PATCH v2 12/12] docs: proc.rst: stat: Note the interrupt counter wrap-around Alexei Lozovsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87y27zb62e.ffs@tglx \
--to=tglx@linutronix.de \
--cc=adobriyan@gmail.com \
--cc=cl@linux.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=me@ilammy.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).