回复: [edk2-devel] 回复: [edk2-devel] FDF parser performance degrades rapidly on non-trivially sized inputs

["gaoliming via groups.io" <gaoliming=byosoft.com.cn@groups.io>]

[-- Attachment #1: Type: text/plain, Size: 5652 bytes --]

Thanks. I will try it. 

 

发件人: via groups.io <davidr=ghs.com@groups.io> 
发送时间: 2024年12月6日 8:55
收件人: gaoliming <gaoliming@byosoft.com.cn>; devel@edk2.groups.io
主题: Re: [edk2-devel] 回复: [edk2-devel] FDF parser performance degrades rapidly on non-trivially sized inputs

 

Here is the code I was testing with. You can probably reduce the changes down to just _SkipWhiteSpace(), but I haven't tried that yet.

 

diff --git a/BaseTools/Source/Python/GenFds/FdfParser.py b/BaseTools/Source/Python/GenFds/FdfParser.py
index feb4c72779..8c57720116 100644
--- a/BaseTools/Source/Python/GenFds/FdfParser.py
+++ b/BaseTools/Source/Python/GenFds/FdfParser.py
@@ -12,6 +12,7 @@
 #
 from __future__ import print_function
 from __future__ import absolute_import
+from io import StringIO
 from re import compile, DOTALL
 from string import hexdigits
 from uuid import UUID
@@ -253,7 +254,7 @@ class FdfParser:
         self.CurrentFdName = None
         self.CurrentFvName = None
         self._Token = ""
-        self._SkippedChars = ""
+        self._SkippedChars = StringIO()
         GlobalData.gFdfParser = self
 
         # Used to section info
@@ -276,7 +277,7 @@ class FdfParser:
     def _SkipWhiteSpace(self):
         while not self._EndOfFile():
             if self._CurrentChar() in {TAB_PRINTCHAR_NUL, T_CHAR_CR, TAB_LINE_BREAK, TAB_SPACE_SPLIT, T_CHAR_TAB}:
-                self._SkippedChars += str(self._CurrentChar())
+                self._SkippedChars.write(str(self._CurrentChar()))
                 self._GetOneChar()
             else:
                 return
@@ -696,7 +697,7 @@ class FdfParser:
                     Header = self._Token
                     if not self._Token.endswith(TAB_SECTION_END):
                         self._SkipToToken(TAB_SECTION_END)
-                        Header += self._SkippedChars
+                        Header += self._SkippedChars.getvalue()
                     if Header.find('$(') != -1:
                         raise Warning("macro cannot be used in section header", self.FileName, self.CurrentLineNumber)
                     self._SectionHeaderParser(Header)
@@ -1226,7 +1227,7 @@ class FdfParser:
             raise Warning(QuoteToUse, self.FileName, self.CurrentLineNumber)
         if currentLineNumber != self.CurrentLineNumber:
             raise Warning(QuoteToUse, self.FileName, self.CurrentLineNumber)
-        self._Token = self._SkippedChars.rstrip(QuoteToUse)
+        self._Token = self._SkippedChars.getvalue().rstrip(QuoteToUse)
         return True
 
     ## _SkipToToken() method
@@ -1243,7 +1244,7 @@ class FdfParser:
     def _SkipToToken(self, String, IgnoreCase = False):
         StartPos = self.GetFileBufferPos()
 
-        self._SkippedChars = ""
+        self._SkippedChars = StringIO()
         while not self._EndOfFile():
             index = -1
             if IgnoreCase:
@@ -1252,13 +1253,13 @@ class FdfParser:
                 index = self._CurrentLine()[self.CurrentOffsetWithinLine: ].find(String)
             if index == 0:
                 self.CurrentOffsetWithinLine += len(String)
-                self._SkippedChars += String
+                self._SkippedChars.write(String)
                 return True
-            self._SkippedChars += str(self._CurrentChar())
+            self._SkippedChars.write(str(self._CurrentChar()))
             self._GetOneChar()
 
         self.SetFileBufferPos(StartPos)
-        self._SkippedChars = ""
+        self._SkippedChars = StringIO()
         return False
 
     ## GetFileBufferPos() method
@@ -2890,7 +2891,7 @@ class FdfParser:
             if not self._SkipToToken(T_CHAR_BRACE_R):
                 raise Warning.Expected("Depex expression ending '}'", self.FileName, self.CurrentLineNumber)
 
-            DepexSectionObj.Expression = self._SkippedChars.rstrip(T_CHAR_BRACE_R)
+            DepexSectionObj.Expression = self._SkippedChars.getvalue().rstrip(T_CHAR_BRACE_R)
             Obj.SectionList.append(DepexSectionObj)
 
         elif self._IsKeyword("SUBTYPE_GUID"):
@@ -3525,7 +3526,7 @@ class FdfParser:
         if not self._SkipToToken(TAB_SPLIT):
             raise Warning.Expected("'.'", self.FileName, self.CurrentLineNumber)
 
-        Arch = self._SkippedChars.rstrip(TAB_SPLIT)
+        Arch = self._SkippedChars.getvalue().rstrip(TAB_SPLIT)
 
         ModuleType = self._GetModuleType()
 

 

 

Also, if you would like to profile the build process you can apply this patch.

 

diff --git a/BaseTools/Source/Python/build/build.py b/BaseTools/Source/Python/build/build.py
index 51fb1f433e..396729efd9 100755
--- a/BaseTools/Source/Python/build/build.py
+++ b/BaseTools/Source/Python/build/build.py
@@ -2778,12 +2778,18 @@ def Main():
     Log_Agent.join()
     return ReturnCode
 
+import cProfile
+
 if __name__ == '__main__':
     try:
         mp.set_start_method('spawn')
     except:
         pass
-    r = Main()
+
+    with cProfile.Profile() as pr:
+        r = Main()
+        pr.print_stats('tottime')
+
     ## 0-127 is a safe return range, and 1 is a standard default error
     if r < 0 or r > 127: r = 1
     sys.exit(r)



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#120885): https://edk2.groups.io/g/devel/message/120885
Mute This Topic: https://groups.io/mt/109951333/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-



[-- Attachment #2: Type: text/html, Size: 11880 bytes --]