From: "Zeng, Star" <star.zeng@intel.com>
To: "Ni, Ruiyu" <ruiyu.ni@intel.com>,
"'edk2-devel@lists.01.org'" <edk2-devel@lists.01.org>
Cc: "Chiu, Chasel" <chasel.chiu@intel.com>,
"Zeng, Star" <star.zeng@intel.com>
Subject: Re: [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io overhead when polling
Date: Wed, 2 May 2018 06:39:11 +0000 [thread overview]
Message-ID: <0C09AFA07DD0434D9E2A0C6AEB0483103BAE670E@shsmsx102.ccr.corp.intel.com> (raw)
In-Reply-To: <734D49CCEBEEF84792F5B80ED585239D5BC84D55@SHSMSX104.ccr.corp.intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com> if it is updated. :)
Thanks,
Star
-----Original Message-----
From: Ni, Ruiyu
Sent: Wednesday, May 2, 2018 2:37 PM
To: Zeng, Star <star.zeng@intel.com>; 'edk2-devel@lists.01.org' <edk2-devel@lists.01.org>
Cc: Chiu, Chasel <chasel.chiu@intel.com>
Subject: RE: [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io overhead when polling
Star,
You are correct.
">" is enough here. I will change it when committing the patch.
Thanks/Ray
> -----Original Message-----
> From: Ni, Ruiyu
> Sent: Wednesday, May 2, 2018 1:12 PM
> To: Zeng, Star <star.zeng@intel.com>; edk2-devel@lists.01.org
> Cc: Chiu, Chasel <chasel.chiu@intel.com>
> Subject: RE: [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io
> overhead when polling
>
> If Multiplicand * Multiplier + Remainder = MAX_UINT64, Even
> Multiplicand =
> MAX_UINT64 / Multiplier, Overflow still happens.
>
> So ">=" is used here.
>
>
>
> Thanks/Ray
>
> > -----Original Message-----
> > From: Zeng, Star
> > Sent: Wednesday, May 2, 2018 11:44 AM
> > To: Ni, Ruiyu <ruiyu.ni@intel.com>; edk2-devel@lists.01.org
> > Cc: Chiu, Chasel <chasel.chiu@intel.com>; Zeng, Star
> > <star.zeng@intel.com>
> > Subject: RE: [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io
> > overhead when polling
> >
> > Is it more accurate
> > if (Multiplicand >= DivU64x64Remainder (MAX_UINT64, Multiplier,
> > NULL)) {
> > ->
> > if (Multiplicand > DivU64x64Remainder (MAX_UINT64, Multiplier,
> > NULL)) {
> >
> >
> > Thanks,
> > Star
> > -----Original Message-----
> > From: Ni, Ruiyu
> > Sent: Thursday, April 26, 2018 10:24 AM
> > To: edk2-devel@lists.01.org
> > Cc: Zeng, Star <star.zeng@intel.com>; Chiu, Chasel
> > <chasel.chiu@intel.com>
> > Subject: [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io
> overhead
> > when polling
> >
> > RootBridgeIo.PollMem()/PollIo() originally don't count the IO/MMIO
> > access overhead when delaying.
> > The patch changes the implementation to count the access overhead so
> > that the actually delay equals to user required delay.
> >
> > Contributed-under: TianoCore Contribution Agreement 1.1
> > Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
> > Cc: Star Zeng <star.zeng@intel.com>
> > Cc: Chasel Chiu <chasel.chiu@intel.com>
> > ---
> > .../Bus/Pci/PciHostBridgeDxe/PciHostBridgeDxe.inf | 3 +-
> > .../Bus/Pci/PciHostBridgeDxe/PciRootBridge.h | 3 +-
> > .../Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c | 151
> +++++++++++++++-
> > -----
> > 3 files changed, 115 insertions(+), 42 deletions(-)
> >
> > diff --git
> > a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridgeDxe.inf
> > b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridgeDxe.inf
> > index 42bd8a23cb..2e56959a8f 100644
> > --- a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridgeDxe.inf
> > +++ b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciHostBridgeDxe.inf
> > @@ -1,7 +1,7 @@
> > ## @file
> > # Generic PCI Host Bridge driver.
> > #
> > -# Copyright (c) 2009 - 2016, Intel Corporation. All rights
> > reserved.<BR>
> > +# Copyright (c) 2009 - 2018, Intel Corporation. All rights
> > +reserved.<BR>
> > #
> > # This program and the accompanying materials # are licensed and
> > made available under the terms and conditions of the BSD License @@
> > -43,6 +43,7 @@ [LibraryClasses]
> > PciSegmentLib
> > UefiLib
> > PciHostBridgeLib
> > + TimerLib
> >
> > [Protocols]
> > gEfiMetronomeArchProtocolGuid ## CONSUMES
> > diff --git a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridge.h
> > b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridge.h
> > index d3dfb57fc6..e2f651aee4 100644
> > --- a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridge.h
> > +++ b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridge.h
> > @@ -2,7 +2,7 @@
> >
> > The PCI Root Bridge header file.
> >
> > -Copyright (c) 1999 - 2017, Intel Corporation. All rights
> > reserved.<BR>
> > +Copyright (c) 1999 - 2018, Intel Corporation. All rights
> > +reserved.<BR>
> > This program and the accompanying materials are licensed and made
> > available under the terms and conditions of the BSD License which
> > accompanies this distribution. The full text of the license may be
> > found at @@ -36,6 +36,7 @@ WITHOUT WARRANTIES OR
> REPRESENTATIONS OF
> > ANY KIND, EITHER EXPRESS OR IMPLIED.
> > #include <Library/BaseLib.h>
> > #include <Library/PciSegmentLib.h>
> > #include <Library/UefiLib.h>
> > +#include <Library/TimerLib.h>
> > #include "PciHostResource.h"
> >
> >
> > diff --git a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c
> > b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c
> > index 5764c2f49f..459e962fe7 100644
> > --- a/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c
> > +++ b/MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c
> > @@ -2,7 +2,7 @@
> >
> > PCI Root Bridge Io Protocol code.
> >
> > -Copyright (c) 1999 - 2017, Intel Corporation. All rights
> > reserved.<BR>
> > +Copyright (c) 1999 - 2018, Intel Corporation. All rights
> > +reserved.<BR>
> > This program and the accompanying materials are licensed and made
> > available under the terms and conditions of the BSD License which
> > accompanies this distribution. The full text of the license may be
> > found at @@ -468,6 +468,85 @@
> RootBridgeIoGetMemTranslationByAddress (
> > return EFI_SUCCESS;
> > }
> >
> > +/**
> > + Return the result of (Multiplicand * Multiplier / Divisor).
> > +
> > + @param Multiplicand A 64-bit unsigned value.
> > + @param Multiplier A 64-bit unsigned value.
> > + @param Divisor A 32-bit unsigned value.
> > + @param Remainder A pointer to a 32-bit unsigned value. This
> parameter
> > is
> > + optional and may be NULL.
> > +
> > + @return Multiplicand * Multiplier / Divisor.
> > +**/
> > +UINT64
> > +MultThenDivU64x64x32 (
> > + IN UINT64 Multiplicand,
> > + IN UINT64 Multiplier,
> > + IN UINT32 Divisor,
> > + OUT UINT32 *Remainder OPTIONAL
> > + )
> > +{
> > + UINT64 Uint64;
> > + UINT32 LocalRemainder;
> > + UINT32 Uint32;
> > + if (Multiplicand >= DivU64x64Remainder (MAX_UINT64, Multiplier,
> > +NULL))
> > {
> > + //
> > + // Make sure Multiplicand is the bigger one.
> > + //
> > + if (Multiplicand < Multiplier) {
> > + Uint64 = Multiplicand;
> > + Multiplicand = Multiplier;
> > + Multiplier = Uint64;
> > + }
> > + //
> > + // Because Multiplicand * Multiplier overflows,
> > + // Multiplicand * Multiplier / Divisor
> > + // = (2 * Multiplicand' + 1) * Multiplier / Divisor
> > + // = 2 * (Multiplicand' * Multiplier / Divisor) + Multiplier / Divisor
> > + //
> > + Uint64 = MultThenDivU64x64x32 (RShiftU64 (Multiplicand, 1),
> > + Multiplier,
> > Divisor, &LocalRemainder);
> > + Uint64 = LShiftU64 (Uint64, 1);
> > + Uint32 = 0;
> > + if ((Multiplicand & 0x1) == 1) {
> > + Uint64 += DivU64x32Remainder (Multiplier, Divisor, &Uint32);
> > + }
> > + return Uint64 + DivU64x32Remainder (Uint32 + LShiftU64
> > +(LocalRemainder, 1), Divisor, Remainder);
> > + } else {
> > + return DivU64x32Remainder (MultU64x64 (Multiplicand,
> > +Multiplier), Divisor, Remainder);
> > + }
> > +}
> > +
> > +/**
> > + Return the elapsed tick count from CurrentTick.
> > +
> > + @param CurrentTick On input, the previous tick count.
> > + On output, the current tick count.
> > + @param StartTick The value the performance counter starts with when
> it
> > + rolls over.
> > + @param EndTick The value that the performance counter ends with
> > before
> > + it rolls over.
> > +
> > + @return The elapsed tick count from CurrentTick.
> > +**/
> > +UINT64
> > +GetElapsedTick (
> > + UINT64 *CurrentTick,
> > + UINT64 StartTick,
> > + UINT64 EndTick
> > + )
> > +{
> > + UINT64 PreviousTick;
> > +
> > + PreviousTick = *CurrentTick;
> > + *CurrentTick = GetPerformanceCounter();
> > + if (StartTick < EndTick) {
> > + return *CurrentTick - PreviousTick;
> > + } else {
> > + return PreviousTick - *CurrentTick;
> > + }
> > +}
> > +
> > /**
> > Polls an address in memory mapped I/O space until an exit
> > condition is
> met,
> > or a timeout occurs.
> > @@ -517,6 +596,11 @@ RootBridgeIoPollMem (
> > EFI_STATUS Status;
> > UINT64 NumberOfTicks;
> > UINT32 Remainder;
> > + UINT64 StartTick;
> > + UINT64 EndTick;
> > + UINT64 CurrentTick;
> > + UINT64 ElapsedTick;
> > + UINT64 Frequency;
> >
> > if (Result == NULL) {
> > return EFI_INVALID_PARAMETER;
> > @@ -542,28 +626,18 @@ RootBridgeIoPollMem (
> > return EFI_SUCCESS;
> >
> > } else {
> > -
> > - //
> > - // Determine the proper # of metronome ticks to wait for polling the
> > - // location. The nuber of ticks is Roundup (Delay /
> > - // mMetronome->TickPeriod)+1
> > - // The "+1" to account for the possibility of the first tick being short
> > - // because we started in the middle of a tick.
> > //
> > - // BugBug: overriding mMetronome->TickPeriod with UINT32 until
> > Metronome
> > - // protocol definition is updated.
> > + // NumberOfTicks = Frenquency * Delay /
> > EFI_TIMER_PERIOD_SECONDS(1)
> > //
> > - NumberOfTicks = DivU64x32Remainder (Delay, (UINT32) mMetronome-
> > >TickPeriod,
> > - &Remainder);
> > - if (Remainder != 0) {
> > - NumberOfTicks += 1;
> > + Frequency = GetPerformanceCounterProperties (&StartTick,
> &EndTick);
> > + NumberOfTicks = MultThenDivU64x64x32 (Frequency, Delay,
> > (UINT32)EFI_TIMER_PERIOD_SECONDS(1), &Remainder);
> > + if (Remainder >= (UINTN)EFI_TIMER_PERIOD_SECONDS(1) / 2) {
> > + NumberOfTicks++;
> > }
> > - NumberOfTicks += 1;
> > -
> > - while (NumberOfTicks != 0) {
> > -
> > - mMetronome->WaitForTick (mMetronome, 1);
> > -
> > + for ( ElapsedTick = 0, CurrentTick = GetPerformanceCounter()
> > + ; ElapsedTick <= NumberOfTicks
> > + ; ElapsedTick += GetElapsedTick (&CurrentTick, StartTick, EndTick)
> > + ) {
> > Status = This->Mem.Read (This, Width, Address, 1, Result);
> > if (EFI_ERROR (Status)) {
> > return Status;
> > @@ -572,8 +646,6 @@ RootBridgeIoPollMem (
> > if ((*Result & Mask) == Value) {
> > return EFI_SUCCESS;
> > }
> > -
> > - NumberOfTicks -= 1;
> > }
> > }
> > return EFI_TIMEOUT;
> > @@ -626,6 +698,11 @@ RootBridgeIoPollIo (
> > EFI_STATUS Status;
> > UINT64 NumberOfTicks;
> > UINT32 Remainder;
> > + UINT64 StartTick;
> > + UINT64 EndTick;
> > + UINT64 CurrentTick;
> > + UINT64 ElapsedTick;
> > + UINT64 Frequency;
> >
> > //
> > // No matter what, always do a single poll.
> > @@ -651,25 +728,18 @@ RootBridgeIoPollIo (
> > return EFI_SUCCESS;
> >
> > } else {
> > -
> > //
> > - // Determine the proper # of metronome ticks to wait for polling the
> > - // location. The number of ticks is Roundup (Delay /
> > - // mMetronome->TickPeriod)+1
> > - // The "+1" to account for the possibility of the first tick being short
> > - // because we started in the middle of a tick.
> > + // NumberOfTicks = Frenquency * Delay /
> > EFI_TIMER_PERIOD_SECONDS(1)
> > //
> > - NumberOfTicks = DivU64x32Remainder (Delay, (UINT32)mMetronome-
> > >TickPeriod,
> > - &Remainder);
> > - if (Remainder != 0) {
> > - NumberOfTicks += 1;
> > + Frequency = GetPerformanceCounterProperties (&StartTick,
> &EndTick);
> > + NumberOfTicks = MultThenDivU64x64x32 (Frequency, Delay,
> > (UINT32)EFI_TIMER_PERIOD_SECONDS(1), &Remainder);
> > + if (Remainder >= (UINTN)EFI_TIMER_PERIOD_SECONDS(1) / 2) {
> > + NumberOfTicks++;
> > }
> > - NumberOfTicks += 1;
> > -
> > - while (NumberOfTicks != 0) {
> > -
> > - mMetronome->WaitForTick (mMetronome, 1);
> > -
> > + for ( ElapsedTick = 0, CurrentTick = GetPerformanceCounter()
> > + ; ElapsedTick <= NumberOfTicks
> > + ; ElapsedTick += GetElapsedTick (&CurrentTick, StartTick, EndTick)
> > + ) {
> > Status = This->Io.Read (This, Width, Address, 1, Result);
> > if (EFI_ERROR (Status)) {
> > return Status;
> > @@ -678,13 +748,14 @@ RootBridgeIoPollIo (
> > if ((*Result & Mask) == Value) {
> > return EFI_SUCCESS;
> > }
> > -
> > - NumberOfTicks -= 1;
> > }
> > }
> > return EFI_TIMEOUT;
> > }
> >
> > +
> > +
> > +
> > /**
> > Enables a PCI driver to access PCI controller registers in the PCI root
> > bridge memory space.
> > --
> > 2.16.1.windows.1
next prev parent reply other threads:[~2018-05-02 6:39 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-26 2:23 [PATCH] MdeModulePkg/PciHostBridge: Count the (mm)io overhead when polling Ruiyu Ni
2018-05-02 3:44 ` Zeng, Star
2018-05-02 5:12 ` Ni, Ruiyu
2018-05-02 6:37 ` Ni, Ruiyu
2018-05-02 6:39 ` Zeng, Star [this message]
2018-05-02 7:57 ` Zeng, Star
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-list 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=0C09AFA07DD0434D9E2A0C6AEB0483103BAE670E@shsmsx102.ccr.corp.intel.com \
--to=devel@edk2.groups.io \
/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