Before adding another AP enumeration mode, clarify the documentation on
the current logic. No functional changes.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1515
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
---

Notes:
    v2:
    - new patch

 UefiCpuPkg/Library/MpInitLib/MpLib.c | 38 +++++++++++++++-----
 1 file changed, 30 insertions(+), 8 deletions(-)

diff --git a/UefiCpuPkg/Library/MpInitLib/MpLib.c b/UefiCpuPkg/Library/MpInitLib/MpLib.c
index d6f84c6f45c0..594a035d8b92 100644
--- a/UefiCpuPkg/Library/MpInitLib/MpLib.c
+++ b/UefiCpuPkg/Library/MpInitLib/MpLib.c
@@ -1045,14 +1045,36 @@ WakeUpAP (
     }
     if (CpuMpData->InitFlag == ApInitConfig) {
       //
-      // Here support two methods to collect AP count through adjust
-      // PcdCpuApInitTimeOutInMicroSeconds values.
-      //
-      // one way is set a value to just let the first AP to start the
-      // initialization, then through the later while loop to wait all Aps
-      // finsh the initialization.
-      // The other way is set a value to let all APs finished the initialzation.
-      // In this case, the later while loop is useless.
+      // The AP enumeration algorithm below is suitable for two use cases.
+      //
+      // (1) The check-in time for an individual AP is bounded, and APs run
+      //     through their initialization routines strongly concurrently. In
+      //     particular, the number of concurrently running APs
+      //     ("NumApsExecuting") is never expected to fall to zero
+      //     *temporarily* -- it is expected to fall to zero only when all
+      //     APs have checked-in.
+      //
+      //     In this case, the platform is supposed to set
+      //     PcdCpuApInitTimeOutInMicroSeconds to a low-ish value (just long
+      //     enough for one AP to start initialization). The timeout will be
+      //     reached soon, and remaining APs are collected by watching
+      //     NumApsExecuting fall to zero. If NumApsExecuting falls to zero
+      //     mid-process, while some APs have not completed initialization,
+      //     the behavior is undefined.
+      //
+      // (2) The check-in time for an individual AP is unbounded, and/or APs
+      //     may complete their initializations widely spread out. In
+      //     particular, some APs may finish initialization before some APs
+      //     even start.
+      //
+      //     In this case, the platform is supposed to set
+      //     PcdCpuApInitTimeOutInMicroSeconds to a high-ish value. The AP
+      //     enumeration will always take that long (except when the boot CPU
+      //     count happens to be maximal, that is,
+      //     PcdCpuMaxLogicalProcessorNumber). All APs are expected to
+      //     check-in before the timeout, and NumApsExecuting is assumed zero
+      //     at timeout. APs that miss the time-out may cause undefined
+      //     behavior.
       //
       TimedWaitForApFinish (
         CpuMpData,
-- 
2.19.1.3.g30247aa5d201



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#48723): https://edk2.groups.io/g/devel/message/48723
Mute This Topic: https://groups.io/mt/34473662/1787277
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub  [importer@patchew.org]
-=-=-=-=-=-=-=-=-=-=-=-

Reviewed-by: Ray Ni <ray.ni@intel.com>

> -----Original Message-----
> From: Laszlo Ersek <lersek@redhat.com>
> Sent: Thursday, October 10, 2019 7:30 PM
> To: edk2-devel-groups-io <devel@edk2.groups.io>
> Cc: Dong, Eric <eric.dong@intel.com>; Ni, Ray <ray.ni@intel.com>
> Subject: [PATCH v2 1/2] UefiCpuPkg/MpInitLib: expand comment on initial
> AP enumeration
> 
> Before adding another AP enumeration mode, clarify the documentation on
> the current logic. No functional changes.
> 
> Cc: Eric Dong <eric.dong@intel.com>
> Cc: Ray Ni <ray.ni@intel.com>
> Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1515
> Signed-off-by: Laszlo Ersek <lersek@redhat.com>
> ---
> 
> Notes:
>     v2:
>     - new patch
> 
>  UefiCpuPkg/Library/MpInitLib/MpLib.c | 38 +++++++++++++++-----
>  1 file changed, 30 insertions(+), 8 deletions(-)
> 
> diff --git a/UefiCpuPkg/Library/MpInitLib/MpLib.c
> b/UefiCpuPkg/Library/MpInitLib/MpLib.c
> index d6f84c6f45c0..594a035d8b92 100644
> --- a/UefiCpuPkg/Library/MpInitLib/MpLib.c
> +++ b/UefiCpuPkg/Library/MpInitLib/MpLib.c
> @@ -1045,14 +1045,36 @@ WakeUpAP (
>      }
>      if (CpuMpData->InitFlag == ApInitConfig) {
>        //
> -      // Here support two methods to collect AP count through adjust
> -      // PcdCpuApInitTimeOutInMicroSeconds values.
> -      //
> -      // one way is set a value to just let the first AP to start the
> -      // initialization, then through the later while loop to wait all Aps
> -      // finsh the initialization.
> -      // The other way is set a value to let all APs finished the initialzation.
> -      // In this case, the later while loop is useless.
> +      // The AP enumeration algorithm below is suitable for two use cases.
> +      //
> +      // (1) The check-in time for an individual AP is bounded, and APs run
> +      //     through their initialization routines strongly concurrently. In
> +      //     particular, the number of concurrently running APs
> +      //     ("NumApsExecuting") is never expected to fall to zero
> +      //     *temporarily* -- it is expected to fall to zero only when all
> +      //     APs have checked-in.
> +      //
> +      //     In this case, the platform is supposed to set
> +      //     PcdCpuApInitTimeOutInMicroSeconds to a low-ish value (just long
> +      //     enough for one AP to start initialization). The timeout will be
> +      //     reached soon, and remaining APs are collected by watching
> +      //     NumApsExecuting fall to zero. If NumApsExecuting falls to zero
> +      //     mid-process, while some APs have not completed initialization,
> +      //     the behavior is undefined.
> +      //
> +      // (2) The check-in time for an individual AP is unbounded, and/or APs
> +      //     may complete their initializations widely spread out. In
> +      //     particular, some APs may finish initialization before some APs
> +      //     even start.
> +      //
> +      //     In this case, the platform is supposed to set
> +      //     PcdCpuApInitTimeOutInMicroSeconds to a high-ish value. The AP
> +      //     enumeration will always take that long (except when the boot CPU
> +      //     count happens to be maximal, that is,
> +      //     PcdCpuMaxLogicalProcessorNumber). All APs are expected to
> +      //     check-in before the timeout, and NumApsExecuting is assumed zero
> +      //     at timeout. APs that miss the time-out may cause undefined
> +      //     behavior.
>        //
>        TimedWaitForApFinish (
>          CpuMpData,
> --
> 2.19.1.3.g30247aa5d201
> 


-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#48805): https://edk2.groups.io/g/devel/message/48805
Mute This Topic: https://groups.io/mt/34473662/1787277
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub  [importer@patchew.org]
-=-=-=-=-=-=-=-=-=-=-=-