From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Dongxiao Xu <dongxiao.xu@intel.com>, xen-devel@lists.xen.org
Cc: keir@xen.org, Ian.Campbell@citrix.com,
George.Dunlap@eu.citrix.com, stefano.stabellini@eu.citrix.com,
Ian.Jackson@eu.citrix.com, JBeulich@suse.com,
dgdegra@tycho.nsa.gov
Subject: Re: [PATCH v13 07/10] x86: enable QoS monitoring for each domain RMID
Date: Mon, 4 Aug 2014 10:44:04 +0100 [thread overview]
Message-ID: <53DF55E4.40704@citrix.com> (raw)
In-Reply-To: <1407118626-76843-8-git-send-email-dongxiao.xu@intel.com>
On 04/08/14 03:17, Dongxiao Xu wrote:
> If the QoS monitoring service is attached to a domain, its related RMID
> will be set to hardware for monitoring when the domain's vcpu is
> scheduled in. When the domain's vcpu is scheduled out, RMID 0
> (system reserved) will be set for monitoring.
>
> Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
> Acked-by: Jan Beulich <jbeulich@suse.com>
> ---
> xen/arch/x86/domain.c | 5 +++++
> xen/arch/x86/pqos.c | 26 ++++++++++++++++++++++++++
> xen/include/asm-x86/msr-index.h | 3 +++
> xen/include/asm-x86/pqos.h | 1 +
> 4 files changed, 35 insertions(+)
>
> diff --git a/xen/arch/x86/domain.c b/xen/arch/x86/domain.c
> index f8e0e33..bb9a4d2 100644
> --- a/xen/arch/x86/domain.c
> +++ b/xen/arch/x86/domain.c
> @@ -1398,6 +1398,8 @@ static void __context_switch(void)
> {
> memcpy(&p->arch.user_regs, stack_regs, CTXT_SWITCH_STACK_BYTES);
> vcpu_save_fpu(p);
> + if ( pqos_monitor_enabled() && pqosm->rmid_inuse )
> + pqos_monitor_assoc_rmid(0);
> p->arch.ctxt_switch_from(p);
> }
>
> @@ -1422,6 +1424,9 @@ static void __context_switch(void)
> }
> vcpu_restore_fpu_eager(n);
> n->arch.ctxt_switch_to(n);
> +
> + if ( pqos_monitor_enabled() && n->domain->arch.pqos_rmid > 0 )
> + pqos_monitor_assoc_rmid(n->domain->arch.pqos_rmid);
> }
>
> gdt = !is_pv_32on64_vcpu(n) ? per_cpu(gdt_table, cpu) :
> diff --git a/xen/arch/x86/pqos.c b/xen/arch/x86/pqos.c
> index 97b527f..46baaa5 100644
> --- a/xen/arch/x86/pqos.c
> +++ b/xen/arch/x86/pqos.c
> @@ -18,10 +18,16 @@
> #include <xen/sched.h>
> #include <asm/pqos.h>
>
> +struct pqr_assoc {
> + uint64_t val;
> + bool_t initialized;
> +};
> +
> struct pqos_monitor *__read_mostly pqosm = NULL;
> static bool_t __initdata opt_pqos = 0;
> static bool_t __initdata opt_pqos_monitor = 1;
> static unsigned int __initdata opt_rmid_max = 255;
> +static DEFINE_PER_CPU(struct pqr_assoc, pqr_assoc);
>
> static void __init parse_pqos_param(char *s)
> {
> @@ -157,6 +163,26 @@ void pqos_monitor_free_rmid(struct domain *d)
> pqosm->rmid_inuse--;
> }
>
> +void pqos_monitor_assoc_rmid(unsigned int rmid)
> +{
> + uint64_t val;
> + uint64_t new_val;
> +
> + if ( !this_cpu(pqr_assoc).initialized )
this_cpu() is specifically unoptimisable by the compiler. Pull
prq_assoc out once at the start by pointer.
struct prq_assoc *pqr = &this_cpu(pqr_assoc);
> + {
> + rdmsrl(MSR_IA32_PQR_ASSOC, this_cpu(pqr_assoc).val);
> + this_cpu(pqr_assoc).initialized = 1;
> + }
> + val = this_cpu(pqr_assoc).val;
> +
> + new_val = (val & ~pqosm->rmid_mask) | (rmid & pqosm->rmid_mask);
> + if ( val != new_val )
> + {
> + wrmsrl(MSR_IA32_PQR_ASSOC, new_val);
> + this_cpu(pqr_assoc).val = new_val;
> + }
> +}
> +
> /*
> * Local variables:
> * mode: C
> diff --git a/xen/include/asm-x86/msr-index.h b/xen/include/asm-x86/msr-index.h
> index 70a8201..f894dfa 100644
> --- a/xen/include/asm-x86/msr-index.h
> +++ b/xen/include/asm-x86/msr-index.h
> @@ -489,4 +489,7 @@
> /* Geode defined MSRs */
> #define MSR_GEODE_BUSCONT_CONF0 0x00001900
>
> +/* Platform QoS MSRs */
> +#define MSR_IA32_PQR_ASSOC 0x00000c8f
> +
Tabs please, like the rest of the file.
~Andrew
> #endif /* __ASM_MSR_INDEX_H */
> diff --git a/xen/include/asm-x86/pqos.h b/xen/include/asm-x86/pqos.h
> index b45c991..48978ea 100644
> --- a/xen/include/asm-x86/pqos.h
> +++ b/xen/include/asm-x86/pqos.h
> @@ -50,6 +50,7 @@ static inline bool_t pqos_monitor_enabled(void)
> void init_platform_qos(void);
> int pqos_monitor_alloc_rmid(struct domain *d);
> void pqos_monitor_free_rmid(struct domain *d);
> +void pqos_monitor_assoc_rmid(unsigned int rmid);
>
> #endif /* __ASM_PQOS_H__ */
>
next prev parent reply other threads:[~2014-08-04 9:44 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-08-04 2:16 [PATCH v13 00/10] enable Cache QoS Monitoring (CQM) feature Dongxiao Xu
2014-08-04 2:16 ` [PATCH v13 01/10] x86: add generic resource (e.g. MSR) access hypercall Dongxiao Xu
2014-08-04 11:07 ` Jan Beulich
2014-08-04 2:16 ` [PATCH v13 02/10] xsm: add resource operation related xsm policy Dongxiao Xu
2014-08-04 2:16 ` [PATCH v13 03/10] tools: provide interface for generic resource access Dongxiao Xu
2014-08-04 2:17 ` [PATCH v13 04/10] x86: detect and initialize Platform QoS Monitoring feature Dongxiao Xu
2014-08-04 11:19 ` Jan Beulich
2014-08-04 2:17 ` [PATCH v13 05/10] x86: dynamically attach/detach QoS monitoring service for a guest Dongxiao Xu
2014-08-04 9:38 ` Andrew Cooper
2014-08-04 11:23 ` Jan Beulich
2014-08-04 2:17 ` [PATCH v13 06/10] x86: collect global QoS monitoring information Dongxiao Xu
2014-08-04 11:31 ` Jan Beulich
2014-08-04 11:34 ` Jan Beulich
2014-08-04 2:17 ` [PATCH v13 07/10] x86: enable QoS monitoring for each domain RMID Dongxiao Xu
2014-08-04 9:44 ` Andrew Cooper [this message]
2014-08-04 11:33 ` Jan Beulich
2014-08-04 2:17 ` [PATCH v13 08/10] x86: add QoS monitoring related MSRs in allowed list Dongxiao Xu
2014-08-04 2:17 ` [PATCH v13 09/10] xsm: add platform QoS related xsm policies Dongxiao Xu
2014-08-04 2:17 ` [PATCH v13 10/10] tools: CMDs and APIs for Platform QoS Monitoring Dongxiao Xu
2014-08-04 10:29 ` [PATCH v13 00/10] enable Cache QoS Monitoring (CQM) feature Jan Beulich
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=53DF55E4.40704@citrix.com \
--to=andrew.cooper3@citrix.com \
--cc=George.Dunlap@eu.citrix.com \
--cc=Ian.Campbell@citrix.com \
--cc=Ian.Jackson@eu.citrix.com \
--cc=JBeulich@suse.com \
--cc=dgdegra@tycho.nsa.gov \
--cc=dongxiao.xu@intel.com \
--cc=keir@xen.org \
--cc=stefano.stabellini@eu.citrix.com \
--cc=xen-devel@lists.xen.org \
/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).