[edk2-devel] 回复: 回复: 回复: [edk2-stable202305] [PATCH] MdeModulePkg/Core/Pei: set AprioriCount=0 before walking through next FV

[gaoliming via groups.io]

Wendy:
  I agree this is bug fix in PeiCore to support more than one FV images those all have apriori list.

  If no other comments, I will merge it for this stable tag 202305. 

Thanks
Liming
> -----邮件原件-----
> 发件人: devel at edk2.groups.io <devel at edk2.groups.io> 代表 Wendy Liao
> via groups.io
> 发送时间: 2023年5月15日 13:18
> 收件人: gaoliming <gaoliming at byosoft.com.cn>; devel at edk2.groups.io
> 抄送: 'Leon Chen' <leon.chen at insyde.com>; 'Tim Lewis'
> <tim.lewis at insyde.com>
> 主题: Re: 回复: 回复: [edk2-devel] [PATCH] MdeModulePkg/Core/Pei: set
> AprioriCount=0 before walking through next FV
> 
> Hi Liming,
> 
> Merged https://github.com/tianocore/edk2/pull/4397
> Thank you.
> 
> Best Regards,
> 
> Wendy Liao
> Insyde Software Corp.
> Phone: +886-2-6608-3688 Ext.8731
> 
> gaoliming 於 2023/05/11 上午 10:53 寫道:
> > Wendy:
> >    I understand the problem now. Your fix is correct. Reviewed-by: Liming
> Gao <gaoliming at byosoft.com.cn>
> >
> >    This is a bug fix. I suggest to merge it for this stable tag 202305.
> >
> > Thanks
> > Liming
> >> -----邮件原件-----
> >> 发件人: devel at edk2.groups.io <devel at edk2.groups.io> 代表 Wendy
> Liao
> >> via groups.io
> >> 发送时间: 2023年5月11日 10:23
> >> 收件人: gaoliming <gaoliming at byosoft.com.cn>; devel at edk2.groups.io
> >> 抄送: 'Leon Chen' <leon.chen at insyde.com>; 'Tim Lewis'
> >> <tim.lewis at insyde.com>
> >> 主题: Re: 回复: [edk2-devel] [PATCH] MdeModulePkg/Core/Pei: set
> >> AprioriCount=0 before walking through next FV
> >>
> >> Hi gaoliming,
> >>
> >> DiscoverPeimsAndOrderWithApriori () will not reset Private->AprioriCount
> >> to Zero when CoreFileHandle->ScanFv = TRUE.
> >> DiscoverPeimsAndOrderWithApriori () {
> >> ...
> >>     if (CoreFileHandle->ScanFv) {
> >>       Private->CurrentFvFileHandles = CoreFileHandle->FvFileHandles;
> >>       return;
> >>     }
> >> ...
> >> }
> >> After go through all FV and the last FV has one or more Apriori Peim,
> >> Private->AprioriCount will not be reset to 0 anymore.
> >>
> >> Scan loop 1
> >> [FV1]
> >> Before DiscoverPeimsAndOrderWithApriori () : Private->AprioriCount = 0,
> >> CoreFileHandle->ScanFv = FALSE
> >> After DiscoverPeimsAndOrderWithApriori ()    : Private->AprioriCount =
> >> 2, CoreFileHandle->ScanFv = TRUE
> >> [FV2]
> >> Before DiscoverPeimsAndOrderWithApriori () : Private->AprioriCount = 2,
> >> CoreFileHandle->ScanFv = FALSE
> >> After DiscoverPeimsAndOrderWithApriori ()    : Private->AprioriCount =
> >> 0, CoreFileHandle->ScanFv = TRUE
> >> ...
> >> [FVn]
> >> Before DiscoverPeimsAndOrderWithApriori () : Private->AprioriCount = 0,
> >> CoreFileHandle->ScanFv = FALSE
> >> After DiscoverPeimsAndOrderWithApriori ()    : Private->AprioriCount =
> >> 1, CoreFileHandle->ScanFv = TRUE
> >>
> >> Scan loop 2
> >> [FV1]
> >> Before DiscoverPeimsAndOrderWithApriori () : Private->AprioriCount = 1,
> >> CoreFileHandle->ScanFv = TRUE
> >> After DiscoverPeimsAndOrderWithApriori ()    : Private->AprioriCount =
> >> 1, CoreFileHandle->ScanFv = TRUE
> >>
> >> Best Regards,
> >>
> >> Wendy Liao
> >> Insyde Software Corp.
> >> Phone: +886-2-6608-3688 Ext.8731
> >>
> >> gaoliming 於 2023/05/11 上午 09:32 寫道:
> >>> Wendy:
> >>>     DiscoverPeimsAndOrderWithApriori () has the logic to reset
> >> Private->AprioriCount as zero.
> >>>     It will set the real AprioriCount for each FV when this FV first scans,
> >> then dispatch the peims in the apriori list.
> >>>     So, I don't think there is the issue here. Do you find the real
> >> functionality issue?
> >>> Thanks
> >>> Liming
> >>>> -----邮件原件-----
> >>>> 发件人: devel at edk2.groups.io <devel at edk2.groups.io> 代表 Wendy
> >> Liao
> >>>> via groups.io
> >>>> 发送时间: 2023年5月10日 14:23
> >>>> 收件人: devel at edk2.groups.io
> >>>> 抄送: Leon Chen <leon.chen at insyde.com>; Tim Lewis
> >>>> <tim.lewis at insyde.com>
> >>>> 主题: [edk2-devel] [PATCH] MdeModulePkg/Core/Pei: set
> AprioriCount=0
> >>>> before walking through next FV
> >>>>
> >>>>
> >>>> REF:https://bugzilla.tianocore.org/show_bug.cgi?id=4438
> >>>>
> >>>> The main dispatch loop in PeiDispatcher() goes through each FV and
> >>>> calls DiscoverPeimsAndOrderWithApriori() to search Apriori file to
> >>>> reorder all PEIMs then do the PEIM dispatched.
> >>>>
> >>>> DiscoverPeimsAndOrderWithApriori() calculates Apriori file count for
> >>>> every FV once and set Private->AprioriCount, but Private->AprioriCount
> >>>> doesn't be set to 0 before dispatch loop walking through the next FV.
> >>>>
> >>>> It causes the peim which sort on less than Private->AprioriCount and
> >>>> depex is not satisfied would be dispatched when dispatch loop go
> through
> >>>> to a scaned FV, even the peim is not set in APRIORI file.
> >>>>
> >>>> Cc: Leon Chen <leon.chen at insyde.com>
> >>>> Cc: Tim Lewis <tim.lewis at insyde.com>
> >>>> Reported-by: Esther Lee <esther.lee at insyde.com>
> >>>> Signed-off-by: Wendy Liao <wendy.liao at insyde.com>
> >>>>
> >>>> ---
> >>>>     MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c | 1 +
> >>>>     1 file changed, 1 insertion(+)
> >>>>
> >>>> diff --git a/MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c
> >>>> b/MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c
> >>>> index d8284f9f4f..5f32ebb560 100644
> >>>> --- a/MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c
> >>>> +++ b/MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c
> >>>> @@ -1630,6 +1630,7 @@ PeiDispatcher (
> >>>>           Private->CurrentFileHandle    = NULL;
> >>>>           Private->CurrentPeimCount     = 0;
> >>>>           Private->CurrentFvFileHandles = NULL;
> >>>> +      Private->AprioriCount         = 0;
> >>>>         }
> >>>>
> >>>>         //
> >>>> --
> >>>> 2.29.2.windows.2
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>
> >>
> >>
> >
> >
> 
> 
> 
> 





-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#104905): https://edk2.groups.io/g/devel/message/104905
Mute This Topic: https://groups.io/mt/98918118/1813853
Group Owner: devel+owner at edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [edk2-devel-archive at redhat.com]
-=-=-=-=-=-=-=-=-=-=-=-