From: "Liang, Kan" <kan.liang@intel.com>
To: Peter Zijlstra <peterz@infradead.org>
CC: "mingo@redhat.com" <mingo@redhat.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "acme@kernel.org" <acme@kernel.org>,
        "jolsa@redhat.com" <jolsa@redhat.com>,
        "tglx@linutronix.de" <tglx@linutronix.de>,
        "eranian@google.com" <eranian@google.com>,
        "ak@linux.intel.com" <ak@linux.intel.com>
Subject: RE: [PATCH V5] perf: Add PERF_SAMPLE_PHYS_ADDR
Thread-Topic: [PATCH V5] perf: Add PERF_SAMPLE_PHYS_ADDR
Thread-Index: AQHTF4UsqecasEDGFUSsTMnxyenm4KKQHSYAgACG6CA=
Date: Tue, 22 Aug 2017 17:15:35 +0000
Message-ID: <37D7C6CF3E00A74B8858931C1DB2F0775378A21B@SHSMSX103.ccr.corp.intel.com>
References: <1502993843-6837-1-git-send-email-kan.liang@intel.com>
 <20170822170736.GI32112@worktop.programming.kicks-ass.net>
In-Reply-To: <20170822170736.GI32112@worktop.programming.kicks-ass.net>
Accept-Language: zh-CN, en-US
Content-Language: en-US
dlp-product: dlpe-windows
dlp-version: 10.0.102.7
dlp-reaction: no-action
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 8BIT
MIME-Version: 1.0
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2108
Lines: 63

> On Thu, Aug 17, 2017 at 02:17:23PM -0400, kan.liang@intel.com wrote:
> > diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
> > index a322fed..f0e8d9c 100644
> > --- a/arch/x86/events/intel/ds.c
> > +++ b/arch/x86/events/intel/ds.c
> > @@ -1065,6 +1065,35 @@ static inline u64 intel_hsw_transaction(struct
> pebs_record_skl *pebs)
> >  	return txn;
> >  }
> >
> > +static u64 dla_to_phys(u64 dla)
> > +{
> > +	u64 phys_addr = 0;
> > +	struct page *p = NULL;
> > +
> > +	if (dla >= TASK_SIZE) {
> > +		/* If it's vmalloc()d memory, leave phys_addr as 0 */
> > +		if (virt_addr_valid(dla) &&
> > +		    !(dla >= VMALLOC_START && dla < VMALLOC_END))
> > +			phys_addr = (u64)virt_to_phys((void *)(uintptr_t)dla);
> > +	} else {
> > +		/*
> > +		 * Walking the pages tables for user address.
> > +		 * Interrupts are disabled, so it prevents any tear down
> > +		 * of the page tables.
> > +		 * Try IRQ-safe __get_user_pages_fast first.
> > +		 * If failed, leave phys_addr as 0.
> > +		 */
> > +		if ((current->mm != NULL) &&
> > +		    (__get_user_pages_fast(dla, 1, 0, &p) == 1))
> > +			phys_addr = page_to_phys(p) + dla % PAGE_SIZE;
> > +
> > +		if (p)
> > +			put_page(p);
> > +	}
> > +
> > +	return phys_addr;
> > +}
> 
> Is this in any way x86 specific? AFAICT this should work in generic code as
> long as data->addr is provided.

No, it's not X86 specific. I think it can be used as generic code to convert
virtual address to physical address.

Thanks,
Kan
> 
> >  static void setup_pebs_sample_data(struct perf_event *event,
> >  				   struct pt_regs *iregs, void *__pebs,
> >  				   struct perf_sample_data *data, @@ -
> 1179,6 +1208,9 @@ static
> > void setup_pebs_sample_data(struct perf_event *event,
> >  	    x86_pmu.intel_cap.pebs_format >= 1)
> >  		data->addr = pebs->dla;
> >
> > +	if ((sample_type & PERF_SAMPLE_PHYS_ADDR) && (data->addr != 0))
> > +		data->phys_addr = dla_to_phys(data->addr);
> > +
> >  	if (x86_pmu.intel_cap.pebs_format >= 2) {
> >  		/* Only set the TSX weight when no memory weight. */
> >  		if ((sample_type & PERF_SAMPLE_WEIGHT) && !fll)
>