[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f4808265-f136-631e-0220-9019e64ee9d3@arm.com>
Date: Mon, 20 Mar 2023 17:10:40 +0000
From: James Morse <james.morse@....com>
To: Babu Moger <babu.moger@....com>, corbet@....net,
reinette.chatre@...el.com, tglx@...utronix.de, mingo@...hat.com,
bp@...en8.de
Cc: fenghua.yu@...el.com, dave.hansen@...ux.intel.com, x86@...nel.org,
hpa@...or.com, paulmck@...nel.org, akpm@...ux-foundation.org,
quic_neeraju@...cinc.com, rdunlap@...radead.org,
damien.lemoal@...nsource.wdc.com, songmuchun@...edance.com,
peterz@...radead.org, jpoimboe@...nel.org, pbonzini@...hat.com,
chang.seok.bae@...el.com, pawan.kumar.gupta@...ux.intel.com,
jmattson@...gle.com, daniel.sneddon@...ux.intel.com,
sandipan.das@....com, tony.luck@...el.com,
linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
bagasdotme@...il.com, eranian@...gle.com,
christophe.leroy@...roup.eu, jarkko@...nel.org,
adrian.hunter@...el.com, quic_jiles@...cinc.com,
peternewman@...gle.com
Subject: Re: [PATCH v3 5/7] x86/resctrl: Display the RMID and COSID for
resctrl groups
Hi Babu,
On 02/03/2023 20:24, Babu Moger wrote:
> When a user creates a control or monitor group, the CLOSID or RMID
> are not visible to the user. These are architecturally defined entities.
On x86. Any other architecture is going to have a hard time supporting this.
> There is no harm in displaying these in resctrl groups. Sometimes it
> can help to debug the issues.
By comparing it with what? Unless user-space can see into the hardware, resctrl is the
only gateway to this stuff. What difference does the allocated value here make?
Could you elaborate on what issues this can help debug?
> Add CLOSID and RMID to the control/monitor groups display in resctrl
> interface.
>
> $cat /sys/fs/resctrl/clos1/closid
> 1
> $cat /sys/fs/resctrl/mon_groups/mon1/rmid
> 3
Er. Please don't expose this to user-space!
MPAM has no equivalent value to RMID, so whatever this is for, can't work on MPAM.
Where I have needed this value for MPAM is to pass the closid/rmid to another kernel
interface. Because the user-space interface needs to be architecture agnostic, I proposed
it as a u64 called 'id' that each architecture can encode/decode as appropriate. [0]
To prevent user-space trying to base anything on the raw closid/rmid values, I went as far
as obfuscating them with a random value picked at boot, to ensure scripts always read the
current value when passing the control/monitor group.
I'm curious what the raw value is useful for.
Thanks,
James
[0]
https://git.kernel.org/pub/scm/linux/kernel/git/morse/linux.git/commit/?h=mpam/snapshot/v6.2&id=d568cf2ba58b7c4970ce41a8d4d6224e285a177e
Powered by blists - more mailing lists