Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1562325

Re: [PATCH 2/7] ARM: Add Broadcom Brahma-B15 readahead cache support

From Florian Fainelli <florian.fainelli@broadcom.com>
Newsgroups linux.kernel
Subject Re: [PATCH 2/7] ARM: Add Broadcom Brahma-B15 readahead cache support
Date 2017-01-19 01:30 +0100
Message-ID <t18M1-2Gh-3@gated-at.bofh.it> (permalink)
References <t15br-o6-7@gated-at.bofh.it> <t15bs-o6-27@gated-at.bofh.it> <t17mW-1Gi-33@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


On 01/18/2017 02:56 PM, Russell King - ARM Linux wrote:
> On Wed, Jan 18, 2017 at 12:29:21PM -0800, Florian Fainelli wrote:
>> The readahead cache only intercepts reads, not writes, as such, some
>> data can remain stale in any of its buffers, such that we need to flush
>> it, which is an operation that needs to happen in a particular order:
>>
>> - disable the readahead cache
>> - flush it
>> - call the appropriate cache-v7.S function
>> - re-enable
> 
> I really do hope that the above explanation is wrong, because if that's
> really how it's implemented, it's going to cause coherency problems.
> 
> It's got to at least monitor writes, otherwise how do you guarantee
> that the CPU doesn't see stale data?  IOW:

Yes, it does monitor writes, the explanation given here was wrong. Thanks!

> 
> Consider this at the L2 memory-side interface (iow, downstream of the
> point-of-coherency):
> 
> 	CPU1		CPU2		Read-ahead buffer
> 			read cache line C
> 					reads cache line C and C+1
> 	writes cache line C+1
> 			read cache line C+1
> 
> What ensures that CPU2 sees the written out cache line from CPU1?
-- 
Florian

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 0/7] ARM: Broadcom Brahma-B15 readahead cache support Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 1/7] ARM: v7: allow setting different cache functions Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 8/8] ARM: smp: Remove CPU: shutdown notice Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 7/7] MAINTAINERS: Update brcmstb entries to cover B15 code Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 2/7] ARM: Add Broadcom Brahma-B15 readahead cache support Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
    Re: [PATCH 2/7] ARM: Add Broadcom Brahma-B15 readahead cache support Russell King - ARM Linux <linux@armlinux.org.uk> - 2017-01-19 00:00 +0100
      Re: [PATCH 2/7] ARM: Add Broadcom Brahma-B15 readahead cache support Florian Fainelli <florian.fainelli@broadcom.com> - 2017-01-19 01:30 +0100
  [PATCH 4/7] ARM: B15: Add CPU hotplug awareness Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 3/7] ARM: Hook B15 readahead cache functions based on processor Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 5/7] ARM: B15: Add suspend/resume hooks Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100
  [PATCH 6/7] ARM: B15: Register reboot notifier for KEXEC Florian Fainelli <f.fainelli@gmail.com> - 2017-01-18 21:40 +0100

csiph-web