public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Rebecca Cran" <rebecca@bsdio.com>
To: devel@edk2.groups.io, michael.d.kinney@intel.com,
	Rebecca Cran <rebecca@quicinc.com>,
	"Gao, Liming" <gaoliming@byosoft.com.cn>,
	"Liu, Zhiguang" <zhiguang.liu@intel.com>,
	Andrew Fish <afish@apple.com>
Subject: Re: [edk2-devel] [PATCH v4 1/1] MdePkg: Use ANSI colors to indicate debug message severity
Date: Mon, 20 Feb 2023 12:44:01 -0700	[thread overview]
Message-ID: <2c0c2488-aef2-61c0-0ebb-6f2e9ed7b641@bsdio.com> (raw)
In-Reply-To: <CO1PR11MB4929D0F28CD0588E76AF0BBDD2019@CO1PR11MB4929.namprd11.prod.outlook.com>

Hi Mike,

Sorry, I only just saw this message.
Your implementation looks better since as you say it doesn't require 
updates to the DebugLib instances.

Since you're more familiar with the gotchas around adding this, I'll 
leave it to you to decide if this can be committed or not.

-- 
Rebecca Cran

On 11/10/22 16:14, Michael D Kinney wrote:
> Hi Rebecca,
> 
> Here is an alternate way to implement this feature.  Just an
> experiment because it does not follow the recommended methods
> to add a feature that can be enabled/disabled.  I have tested
> this on EmulatorPkg and OvmfPkg and it works as expected.
> Does not require any updates to any DebugLib instances.
> 
> https://github.com/mdkinney/edk2/tree/POC_DebugLibAnsi
> https://github.com/mdkinney/edk2/commit/8961ee82e95495651eca53145ae76c4f0097c637
> 
> In general we do not like using #if to enable/disable features.
> However, this feature does increase the size of modules to enable
> ANSI color sequences.  This approach has zero size impact
> if MDEPKG_DEBUGANSI is not defined.
> 
> There are also challenges with this feature for platform FW
> builds that may integrate binaries from other sources if there
> is a mix of modules that enable or disable the feature.
> 
> Only the platform developer knows if the device that all
> DEBUG()/ASSERT() messages are routed supports ANSI color
> sequences or not.  This means adding the sequences needs
> to be in a platform component.  Or we need a mode where all
> generated messages are more like trace messages that include
> the message type and error level, and let the receiver of
> the trace messages reformat as needed.
> 
> Best regards,
> 
> Mike
> 
>> -----Original Message-----
>> From: Kinney, Michael D <michael.d.kinney@intel.com>
>> Sent: Thursday, November 10, 2022 11:12 AM
>> To: Rebecca Cran <rebecca@quicinc.com>; devel@edk2.groups.io; Gao, Liming <gaoliming@byosoft.com.cn>; Liu, Zhiguang
>> <zhiguang.liu@intel.com>; Andrew Fish <afish@apple.com>; Kinney, Michael D <michael.d.kinney@intel.com>
>> Subject: RE: [PATCH v4 1/1] MdePkg: Use ANSI colors to indicate debug message severity
>>
>> Hi Rebecca,
>>
>> In the edk2 repo, I see the following instances of the DebugPrintMarker() API
>> where you are currently adding he ANSI color sequences.  Should all of these
>> be updated?  I did not review the edk2-platforms repo.  And there may be
>> downstream custom DebugLib instances.  It would better if this feature could
>> be enabled for all existing DebugLib instances, but the only common location
>> is the DEBUG() macro definition in MdePkg/Include/Library/DebugLib.h and
>> adding code to that macro adds statements to the module calling the DebugLib
>> services, which can increase code size.
>>
>> ArmPkg\Library\SemiHostingDebugLib\DebugLib.c:
>>     75  VOID
>>     76: DebugPrintMarker (
>>     77    IN  UINTN        ErrorLevel,
>>
>> IntelFsp2Pkg\Library\BaseFspDebugLibSerialPort\DebugLib.c:
>>     89  VOID
>>     90: DebugPrintMarker (
>>     91    IN  UINTN        ErrorLevel,
>>
>> MdeModulePkg\Library\PeiDxeDebugLibReportStatusCode\DebugLib.c:
>>     85  VOID
>>     86: DebugPrintMarker (
>>     87    IN  UINTN        ErrorLevel,
>>
>> MdePkg\Library\BaseDebugLibSerialPort\DebugLib.c:
>>     97  VOID
>>     98: DebugPrintMarker (
>>     99    IN  UINTN        ErrorLevel,
>>
>> MdePkg\Library\DxeRuntimeDebugLibSerialPort\DebugLib.c:
>>    156  VOID
>>    157: DebugPrintMarker (
>>    158    IN  UINTN        ErrorLevel,
>>
>> MdePkg\Library\UefiDebugLibConOut\DebugLib.c:
>>     79  VOID
>>     80: DebugPrintMarker (
>>     81    IN  UINTN        ErrorLevel,
>>
>> MdePkg\Library\UefiDebugLibDebugPortProtocol\DebugLib.c:
>>    137  VOID
>>    138: DebugPrintMarker (
>>    139    IN  UINTN        ErrorLevel,
>>
>> MdePkg\Library\UefiDebugLibStdErr\DebugLib.c:
>>     79  VOID
>>     80: DebugPrintMarker (
>>     81    IN  UINTN        ErrorLevel,
>>
>> OvmfPkg\Library\PlatformDebugLibIoPort\DebugLib.c:
>>     79  VOID
>>     80: DebugPrintMarker (
>>     81    IN  UINTN        ErrorLevel,
>>
>> Also, the ErrorLevel parameter is a bitmask.  It cannot be used in
>> a switch/case statement for only 1 bit being set.  To test for a
>> debug message of type ERROR or WARN, a check must be done for that
>> one bit being set in ErrorLevel.
>>
>> The DebugLib also provides support for ASSERT() macros.  Should
>> ASSERT() messages have the same color as an ERROR message?  Or
>> its own color?
>>
>> The logic below changes the color for the message based on message
>> type.  But is does not save/restore the current ANSI color setting.
>> It see it using the END_ESC_SEQ.  Does that put the color settings
>> back to the previous setting?
>>
>> Thanks,
>>
>> Mike
>>
>>> -----Original Message-----
>>> From: Rebecca Cran <rebecca@quicinc.com>
>>> Sent: Thursday, October 27, 2022 11:51 AM
>>> To: devel@edk2.groups.io; Kinney, Michael D <michael.d.kinney@intel.com>; Gao, Liming <gaoliming@byosoft.com.cn>; Liu,
>> Zhiguang
>>> <zhiguang.liu@intel.com>; Andrew Fish <afish@apple.com>
>>> Cc: Rebecca Cran <rebecca@quicinc.com>
>>> Subject: [PATCH v4 1/1] MdePkg: Use ANSI colors to indicate debug message severity
>>>
>>> There currently isn't a way to differentiate the different
>>> levels of DEBUG output: DEBUG_ERROR, DEBUG_WARN, DEBUG_INFO
>>> etc.
>>>
>>> To improve this, wrap DEBUG_ERROR and DEBUG_WARN level
>>> messages in ANSI color code escape sequences. DEBUG_ERROR
>>> messages will be displayed in red text, and DEBUG_WARN
>>> in yellow.
>>>
>>> Only enable this new functionality if the FeatureFlag
>>> gEfiMdePkgTokenSpaceGuid.PcdDebugAnsiSeqSupport
>>> is set to TRUE. By default it's FALSE.
>>>
>>> Signed-off-by: Rebecca Cran <rebecca@quicinc.com>
>>> ---
>>>   MdePkg/MdePkg.dec                                                |  6 ++
>>>   MdePkg/Library/BaseDebugLibSerialPort/BaseDebugLibSerialPort.inf |  2 +-
>>>   MdePkg/Library/UefiDebugLibConOut/UefiDebugLibConOut.inf         |  2 +-
>>>   MdePkg/Library/BaseDebugLibSerialPort/DebugLib.c                 | 64 ++++++++++++++++++++
>>>   MdePkg/Library/UefiDebugLibConOut/DebugLib.c                     | 64 ++++++++++++++++++++
>>>   5 files changed, 136 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/MdePkg/MdePkg.dec b/MdePkg/MdePkg.dec
>>> index 4c81cbd75ab2..8ddc46b62e7d 100644
>>> --- a/MdePkg/MdePkg.dec
>>> +++ b/MdePkg/MdePkg.dec
>>> @@ -4,6 +4,7 @@
>>>   # It also provides the definitions(including PPIs/PROTOCOLs/GUIDs) of
>>>   # EFI1.10/UEFI2.7/PI1.7 and some Industry Standards.
>>>   #
>>> +# Copyright (c) 2022, Qualcomm Innovation Center, Inc. All rights reserved.<BR>
>>>   # Copyright (c) 2007 - 2022, Intel Corporation. All rights reserved.<BR>
>>>   # Portions copyright (c) 2008 - 2009, Apple Inc. All rights reserved.<BR>
>>>   # (C) Copyright 2016 - 2021 Hewlett Packard Enterprise Development LP<BR>
>>> @@ -1977,6 +1978,11 @@ [PcdsFeatureFlag]
>>>     # @Prompt Validate ORDERED_COLLECTION structure
>>>     gEfiMdePkgTokenSpaceGuid.PcdValidateOrderedCollection|FALSE|BOOLEAN|0x0000002a
>>>
>>> +  ## Indicates if DEBUG output should use ANSI sequences.<BR><BR>
>>> +  #   TRUE  - Will use ANSI sequences in DEBUG output.<BR>
>>> +  #   FALSE - Will not use ANSI sequences in DEBUG output.<BR>
>>> +  gEfiMdePkgTokenSpaceGuid.PcdDebugAnsiSeqSupport|FALSE|BOOLEAN|0x0000002f
>>> +
>>>   [PcdsFixedAtBuild]
>>>     ## Status code value for indicating a watchdog timer has expired.
>>>     # EFI_COMPUTING_UNIT_HOST_PROCESSOR | EFI_CU_HP_EC_TIMER_EXPIRED
>>> diff --git a/MdePkg/Library/BaseDebugLibSerialPort/BaseDebugLibSerialPort.inf
>>> b/MdePkg/Library/BaseDebugLibSerialPort/BaseDebugLibSerialPort.inf
>>> index 7504faee67f0..8d6ed759e974 100644
>>> --- a/MdePkg/Library/BaseDebugLibSerialPort/BaseDebugLibSerialPort.inf
>>> +++ b/MdePkg/Library/BaseDebugLibSerialPort/BaseDebugLibSerialPort.inf
>>> @@ -41,4 +41,4 @@ [Pcd]
>>>     gEfiMdePkgTokenSpaceGuid.PcdDebugClearMemoryValue  ## SOMETIMES_CONSUMES
>>>     gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask      ## CONSUMES
>>>     gEfiMdePkgTokenSpaceGuid.PcdFixedDebugPrintErrorLevel ## CONSUMES
>>> -
>>> +  gEfiMdePkgTokenSpaceGuid.PcdDebugAnsiSeqSupport       ## CONSUMES
>>> diff --git a/MdePkg/Library/UefiDebugLibConOut/UefiDebugLibConOut.inf
>> b/MdePkg/Library/UefiDebugLibConOut/UefiDebugLibConOut.inf
>>> index 53bbc8ce3f65..694494ffc7a3 100644
>>> --- a/MdePkg/Library/UefiDebugLibConOut/UefiDebugLibConOut.inf
>>> +++ b/MdePkg/Library/UefiDebugLibConOut/UefiDebugLibConOut.inf
>>> @@ -50,4 +50,4 @@ [Pcd]
>>>     gEfiMdePkgTokenSpaceGuid.PcdDebugClearMemoryValue        ## SOMETIMES_CONSUMES
>>>     gEfiMdePkgTokenSpaceGuid.PcdDebugPropertyMask            ## CONSUMES
>>>     gEfiMdePkgTokenSpaceGuid.PcdFixedDebugPrintErrorLevel    ## CONSUMES
>>> -
>>> +  gEfiMdePkgTokenSpaceGuid.PcdDebugAnsiSeqSupport          ## CONSUMES
>>> diff --git a/MdePkg/Library/BaseDebugLibSerialPort/DebugLib.c b/MdePkg/Library/BaseDebugLibSerialPort/DebugLib.c
>>> index bd5686947712..df31bd1ffb9f 100644
>>> --- a/MdePkg/Library/BaseDebugLibSerialPort/DebugLib.c
>>> +++ b/MdePkg/Library/BaseDebugLibSerialPort/DebugLib.c
>>> @@ -26,6 +26,10 @@
>>>   //
>>>   #define MAX_DEBUG_MESSAGE_LENGTH  0x100
>>>
>>> +#define RED_ESC_SEQ     "\033[31m"
>>> +#define YELLOW_ESC_SEQ  "\033[33m"
>>> +#define END_ESC_SEQ     "\033[0m"
>>> +
>>>   //
>>>   // VA_LIST can not initialize to NULL for all compiler, so we use this to
>>>   // indicate a null VA_LIST
>>> @@ -77,6 +81,62 @@ DebugPrint (
>>>     VA_END (Marker);
>>>   }
>>>
>>> +/**
>>> +  Wraps a message with ANSI color escape codes.
>>> +
>>> +  @param String     The string to wrap.
>>> +  @param StringLen  The size of the String buffer in characters.
>>> +  @param ErrorLevel The error level.
>>> +
>>> +  @retval RETURN_SUCCESS          The string was successfully updated.
>>> +  @retval RETURN_BUFFER_TOO_SMALL The buffer is too small.
>>> +
>>> +**/
>>> +STATIC
>>> +RETURN_STATUS
>>> +AsciiDebugGetColorString (
>>> +  IN OUT CHAR8  *String,
>>> +  IN UINTN      StringLen,
>>> +  IN UINTN      ErrorLevel
>>> +  )
>>> +{
>>> +  CHAR8  Buffer[MAX_DEBUG_MESSAGE_LENGTH];
>>> +  UINTN  ReqBufferLen;
>>> +
>>> +  ReqBufferLen = AsciiStrLen (String) +
>>> +                 AsciiStrLen (RED_ESC_SEQ) +
>>> +                 AsciiStrLen (END_ESC_SEQ) +
>>> +                 1;
>>> +
>>> +  if (StringLen < ReqBufferLen) {
>>> +    return RETURN_BUFFER_TOO_SMALL;
>>> +  }
>>> +
>>> +  ZeroMem (Buffer, sizeof (Buffer));
>>> +
>>> +  switch (ErrorLevel) {
>>> +    case DEBUG_WARN:
>>> +      AsciiStrCpyS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, YELLOW_ESC_SEQ);
>>> +      break;
>>> +    case DEBUG_ERROR:
>>> +      AsciiStrCpyS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, RED_ESC_SEQ);
>>> +      break;
>>> +  }
>>> +
>>> +  AsciiStrCatS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, String);
>>> +
>>> +  switch (ErrorLevel) {
>>> +    case DEBUG_WARN:
>>> +    case DEBUG_ERROR:
>>> +      AsciiStrCatS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, END_ESC_SEQ);
>>> +      break;
>>> +  }
>>> +
>>> +  AsciiStrCpyS (String, StringLen, Buffer);
>>> +
>>> +  return RETURN_SUCCESS;
>>> +}
>>> +
>>>   /**
>>>     Prints a debug message to the debug output device if the specified
>>>     error level is enabled base on Null-terminated format string and a
>>> @@ -125,6 +185,10 @@ DebugPrintMarker (
>>>       AsciiBSPrint (Buffer, sizeof (Buffer), Format, BaseListMarker);
>>>     }
>>>
>>> +  if (FeaturePcdGet (PcdDebugAnsiSeqSupport)) {
>>> +    AsciiDebugGetColorString (Buffer, MAX_DEBUG_MESSAGE_LENGTH, ErrorLevel);
>>> +  }
>>> +
>>>     //
>>>     // Send the print string to a Serial Port
>>>     //
>>> diff --git a/MdePkg/Library/UefiDebugLibConOut/DebugLib.c b/MdePkg/Library/UefiDebugLibConOut/DebugLib.c
>>> index 65c8dc2b4654..521298a7c8a7 100644
>>> --- a/MdePkg/Library/UefiDebugLibConOut/DebugLib.c
>>> +++ b/MdePkg/Library/UefiDebugLibConOut/DebugLib.c
>>> @@ -20,6 +20,10 @@
>>>   //
>>>   #define MAX_DEBUG_MESSAGE_LENGTH  0x100
>>>
>>> +#define RED_ESC_SEQ     L"\033[31m"
>>> +#define YELLOW_ESC_SEQ  L"\033[33m"
>>> +#define END_ESC_SEQ     L"\033[0m"
>>> +
>>>   //
>>>   // VA_LIST can not initialize to NULL for all compiler, so we use this to
>>>   // indicate a null VA_LIST
>>> @@ -59,6 +63,62 @@ DebugPrint (
>>>     VA_END (Marker);
>>>   }
>>>
>>> +/**
>>> +  Wraps a message with ANSI color escape codes.
>>> +
>>> +  @param String     The string to wrap.
>>> +  @param StringLen  The size of the String buffer in Unicode characters.
>>> +  @param ErrorLevel The error level.
>>> +
>>> +  @retval RETURN_SUCCESS          The string was successfully updated.
>>> +  @retval RETURN_BUFFER_TOO_SMALL The buffer is too small.
>>> +
>>> +**/
>>> +STATIC
>>> +RETURN_STATUS
>>> +UnicodeDebugGetColorString (
>>> +  IN OUT CHAR16  *String,
>>> +  IN UINTN       StringLen,
>>> +  IN UINTN       ErrorLevel
>>> +  )
>>> +{
>>> +  CHAR16  Buffer[MAX_DEBUG_MESSAGE_LENGTH];
>>> +  UINTN   ReqBufferLen;
>>> +
>>> +  ReqBufferLen = StrLen (String) +
>>> +                 StrLen (RED_ESC_SEQ) +
>>> +                 StrLen (END_ESC_SEQ) +
>>> +                 1;
>>> +
>>> +  if (StringLen < ReqBufferLen) {
>>> +    return RETURN_BUFFER_TOO_SMALL;
>>> +  }
>>> +
>>> +  ZeroMem (Buffer, sizeof (Buffer));
>>> +
>>> +  switch (ErrorLevel) {
>>> +    case DEBUG_WARN:
>>> +      StrCpyS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, YELLOW_ESC_SEQ);
>>> +      break;
>>> +    case DEBUG_ERROR:
>>> +      StrCpyS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, RED_ESC_SEQ);
>>> +      break;
>>> +  }
>>> +
>>> +  StrCatS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, String);
>>> +
>>> +  switch (ErrorLevel) {
>>> +    case DEBUG_WARN:
>>> +    case DEBUG_ERROR:
>>> +      StrCatS (Buffer, MAX_DEBUG_MESSAGE_LENGTH, END_ESC_SEQ);
>>> +      break;
>>> +  }
>>> +
>>> +  StrCpyS (String, StringLen, Buffer);
>>> +
>>> +  return RETURN_SUCCESS;
>>> +}
>>> +
>>>   /**
>>>     Prints a debug message to the debug output device if the specified
>>>     error level is enabled base on Null-terminated format string and a
>>> @@ -108,6 +168,10 @@ DebugPrintMarker (
>>>         UnicodeBSPrintAsciiFormat (Buffer, sizeof (Buffer), Format, BaseListMarker);
>>>       }
>>>
>>> +    if (FeaturePcdGet (PcdDebugAnsiSeqSupport)) {
>>> +      UnicodeDebugGetColorString (Buffer, MAX_DEBUG_MESSAGE_LENGTH, ErrorLevel);
>>> +    }
>>> +
>>>       //
>>>       // Send the print string to the Console Output device
>>>       //
>>> --
>>> 2.30.2
> 
> 
> 
> 
> 
> 


  reply	other threads:[~2023-02-20 19:44 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-27 18:51 [PATCH v4 1/1] MdePkg: Use ANSI colors to indicate debug message severity Rebecca Cran
2022-11-10 19:11 ` Michael D Kinney
2022-11-10 23:14   ` Michael D Kinney
2023-02-20 19:44     ` Rebecca Cran [this message]
2023-02-20 19:14   ` Rebecca Cran
     [not found] <172201BE20D562E0.7329@groups.io>
2022-11-08 11:34 ` [edk2-devel] " Rebecca Cran

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=2c0c2488-aef2-61c0-0ebb-6f2e9ed7b641@bsdio.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