[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a03e5f5e-ac9c-f8b1-0260-d58ede569d67@amd.com>
Date: Thu, 26 Oct 2023 17:40:25 -0500
From: "Moger, Babu" <bmoger@....com>
To: Tony Luck <tony.luck@...el.com>, Fenghua Yu <fenghua.yu@...el.com>,
Reinette Chatre <reinette.chatre@...el.com>,
Peter Newman <peternewman@...gle.com>,
Jonathan Corbet <corbet@....net>,
Shuah Khan <skhan@...uxfoundation.org>, x86@...nel.org
Cc: Shaopeng Tan <tan.shaopeng@...itsu.com>,
James Morse <james.morse@....com>,
Jamie Iles <quic_jiles@...cinc.com>,
Babu Moger <babu.moger@....com>,
Randy Dunlap <rdunlap@...radead.org>,
linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
patches@...ts.linux.dev
Subject: Re: [PATCH v3] x86/resctrl: mba_MBps: Fall back to total b/w if local
b/w unavailable
Hi Tony,
On 10/26/2023 3:02 PM, Tony Luck wrote:
> On Intel the various resource director technology (RDT) features are all
> orthogonal and independently enumerated. Thus it is possible to have
> a system that provides "total" memory bandwidth measurements without
> providing "local" bandwidth measurements.
>
> If local bandwidth measurement is not available, do not give up on
> providing the "mba_MBps" feedback option completely, make the code fall
> back to using total bandwidth.
>
> Signed-off-by: Tony Luck <tony.luck@...el.com>
> ---
> Change since v2:
>
> Babu doesn't like the global variable. So here's a version without it.
>
> Note that my preference is still the v2 version. But as I tell newbies
> to Linux "Your job isn't to get YOUR patch upstream. You job is to get
> the problem fixed.". So taking my own advice I don't really mind
> whether v2 or v3 is applied.
Hmm. I like v3 better. Few minor comments below.
>
> arch/x86/kernel/cpu/resctrl/monitor.c | 43 ++++++++++++++++++--------
> arch/x86/kernel/cpu/resctrl/rdtgroup.c | 2 +-
> 2 files changed, 31 insertions(+), 14 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c
> index f136ac046851..29e86310677d 100644
> --- a/arch/x86/kernel/cpu/resctrl/monitor.c
> +++ b/arch/x86/kernel/cpu/resctrl/monitor.c
> @@ -418,6 +418,20 @@ static int __mon_event_count(u32 rmid, struct rmid_read *rr)
> return 0;
> }
>
> +/*
> + * For legacy compatibility use the local memory bandwidth to drive
> + * the mba_MBps feedback control loop. But on platforms that do not
> + * provide the local event fall back to use the total bandwidth event
> + * instead.
> + */
> +static enum resctrl_event_id pick_mba_mbps_event(void)
> +{
> + if (is_mbm_local_enabled())
> + return QOS_L3_MBM_LOCAL_EVENT_ID;
> +
> + return QOS_L3_MBM_TOTAL_EVENT_ID;
> +}
> +
> /*
> * mbm_bw_count() - Update bw count from values previously read by
> * __mon_event_count().
> @@ -431,9 +445,11 @@ static int __mon_event_count(u32 rmid, struct rmid_read *rr)
> */
> static void mbm_bw_count(u32 rmid, struct rmid_read *rr)
> {
> - struct mbm_state *m = &rr->d->mbm_local[rmid];
> + enum resctrl_event_id mba_mbps_evt_id = pick_mba_mbps_event();
How about evt_id instead of mba_mbps_evt_id? It seems pretty mouthful
for temp variable.
> u64 cur_bw, bytes, cur_bytes;
> + struct mbm_state *m;
>
> + m = get_mbm_state(rr->d, rmid, mba_mbps_evt_id);
> cur_bytes = rr->val;
> bytes = cur_bytes - m->prev_bw_bytes;
> m->prev_bw_bytes = cur_bytes;
> @@ -518,6 +534,7 @@ void mon_event_count(void *info)
> */
> static void update_mba_bw(struct rdtgroup *rgrp, struct rdt_domain *dom_mbm)
> {
> + enum resctrl_event_id mba_mbps_evt_id = pick_mba_mbps_event();
same comment as above.
> u32 closid, rmid, cur_msr_val, new_msr_val;
> struct mbm_state *pmbm_data, *cmbm_data;
> u32 cur_bw, delta_bw, user_bw;
> @@ -526,14 +543,14 @@ static void update_mba_bw(struct rdtgroup *rgrp, struct rdt_domain *dom_mbm)
> struct list_head *head;
> struct rdtgroup *entry;
>
> - if (!is_mbm_local_enabled())
> + if (!is_mbm_enabled())
> return;
>
> r_mba = &rdt_resources_all[RDT_RESOURCE_MBA].r_resctrl;
>
> closid = rgrp->closid;
> rmid = rgrp->mon.rmid;
> - pmbm_data = &dom_mbm->mbm_local[rmid];
> + pmbm_data = get_mbm_state(dom_mbm, rmid, mba_mbps_evt_id);
>
> dom_mba = get_domain_from_cpu(smp_processor_id(), r_mba);
> if (!dom_mba) {
> @@ -553,7 +570,7 @@ static void update_mba_bw(struct rdtgroup *rgrp, struct rdt_domain *dom_mbm)
> */
> head = &rgrp->mon.crdtgrp_list;
> list_for_each_entry(entry, head, mon.crdtgrp_list) {
> - cmbm_data = &dom_mbm->mbm_local[entry->mon.rmid];
> + cmbm_data = get_mbm_state(dom_mbm, entry->mon.rmid, mba_mbps_evt_id);
> cur_bw += cmbm_data->prev_bw;
> delta_bw += cmbm_data->delta_bw;
> }
> @@ -595,7 +612,7 @@ static void update_mba_bw(struct rdtgroup *rgrp, struct rdt_domain *dom_mbm)
> */
> pmbm_data->delta_comp = true;
> list_for_each_entry(entry, head, mon.crdtgrp_list) {
> - cmbm_data = &dom_mbm->mbm_local[entry->mon.rmid];
> + cmbm_data = get_mbm_state(dom_mbm, entry->mon.rmid, mba_mbps_evt_id);
> cmbm_data->delta_comp = true;
> }
> }
> @@ -621,15 +638,15 @@ static void mbm_update(struct rdt_resource *r, struct rdt_domain *d, int rmid)
> rr.evtid = QOS_L3_MBM_LOCAL_EVENT_ID;
> rr.val = 0;
> __mon_event_count(rmid, &rr);
> -
> - /*
> - * Call the MBA software controller only for the
> - * control groups and when user has enabled
> - * the software controller explicitly.
> - */
> - if (is_mba_sc(NULL))
> - mbm_bw_count(rmid, &rr);
> }
> +
> + /*
> + * Call the MBA software controller only for the
> + * control groups and when user has enabled
> + * the software controller explicitly.
> + */
> + if (is_mba_sc(NULL))
> + mbm_bw_count(rmid, &rr);
> }
>
> /*
> diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> index 69a1de92384a..0c4f8a1b8df0 100644
> --- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> +++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> @@ -2294,7 +2294,7 @@ static bool supports_mba_mbps(void)
> {
> struct rdt_resource *r = &rdt_resources_all[RDT_RESOURCE_MBA].r_resctrl;
>
> - return (is_mbm_local_enabled() &&
> + return (is_mbm_enabled() &&
> r->alloc_capable && is_mba_linear());
> }
>
Otherwise looks good to me.
Thanks
Babu
Powered by blists - more mailing lists