error::pass1(7stap) [centos man page]

ERROR::PASS1(7stap)													       ERROR::PASS1(7stap)

NAME

       error::pass1 - systemtap pass-1 errors

DESCRIPTION

       Errors  that occur during pass 1 (parsing) usually mean a basic syntax error of some sort occurred in the systemtap script.  There are sev-
       eral classes of problems possible:

       plain syntax error
	      The systemtap script parser detects a large variety of errors, such as missing operands, bad punctuation.  It  tries  to	list  what
	      kinds  of  tokens  it was expecting to see, and will show the region of the source code with the problem.  Please review the stap(1)
	      man page and/or the tutorial, to correct the script's syntax.

       grammar ambiguities
	      There is at least one known ambiguity in the systemtap grammar.  It relates to the optionality of ; (semicolon)  separators  between
	      statements, and the ++ and -- increment/decrement operators.  If the parser indicates an error, consider adding some explicit ; sep-
	      arators between nearby statements and try again.

       missing command line arguments
	      A systemtap script that uses the $N and @N constructs for substituting in command-line options may fail if not enough  options  were
	      given on the stap command line.

       compatibility changes
	      Some  versions  of systemtap have changed the language incompatibly, for example by adding the try/catch keywords for exception han-
	      dling.  In such cases, rerun systemtap with the --compatibility=VERSION option, substituting the last systemtap version  where  your
	      script was known to work.  You may also check the release-history NEWS file for compatibility changes.

GATHERING MORE INFORMATION

       Increasing the verbosity of pass-1 with an option such as --vp 1 can help pinpoint the problem.

SEE ALSO

       stap(1),
       error::reporting(7stap)

															       ERROR::PASS1(7stap)

ERROR::BUILDID(7stap)													     ERROR::BUILDID(7stap)

NAME

       error::buildid - build-id verification failures

DESCRIPTION

       Because	systemtap's script translation / execution stages may be executed at different times and places, it is sometimes necessary to ver-
       ify certain invariants.	One such invariant is that if a script was informed by translate-time analysis of  executables,  then  those  same
       executables  need  to  be  used	at run time.  This checking is done based upon the build-id, a binary hash that modern (post-2007) compil-
       ers/toolchains add as an NT_GNU_BUILD_ID ELF note to object files and executables.  Use the readelf -n command to examine the build-ids	of
       binaries, if you are interested.

       Only  scripts  are sensitive to executables' build-ids: generally those that perform deep analysis of the binaries or their debuginfo.  For
       example, scripts that place .function or .statement probes, or use stack  backtrace-related  tapset  functions  may  be	sensitive.   Other
       scripts	that  rely  only  on  process.mark  or kernel.trace probes do not require debuginfo.  See the DWARF DEBUGINFO section in the stap-
       probes(3stap) man page.

       During translation, systemtap saves a copy of the relevant files' build-ids within the compiled modules.  At run-time, the modules  compare
       the saved ones to the actual run-time build-ids in memory.  The error message indicates that they did not match, so the module will decline
       placing a probe that was computed based upon obsolete data.  This is important for safety, as placing  them  at	an  inappropriate  address
       could crash the programs.  However, this is not necessarily a fatal error, since probes unrelated to the mismatching binaries may operate.

       A build-id mismatch could be caused by a few different situations.  The main one is where the executable versions or architecture were dif-
       ferent between the systemtap translation and execution times/places.  For example, one may run a stap-server on a slightly  different  ver-
       sion of the OS distribution.  The kernel running on the workstation may be slightly different from the version being targeted - perhaps due
       to a pending kernel upgrade leaving different files on disk versus running in memory.  If your  OS  distribution  uses  separate  debuginfo
       packages, the split .IR .debug files may not exactly match the main binaries.

SEE ALSO

       http://fedoraproject.org/wiki/Releases/FeatureBuildId,
       stap(1),
       stapprobes(3stap),
       warning::debuginfo(7stap),
       error::reporting(7stap)

															     ERROR::BUILDID(7stap)