All of lore.kernel.org
 help / color / mirror / Atom feed
From: Alexander Bulekov <alxndr@bu.edu>
To: qemu-devel@nongnu.org
Cc: bonzini@redhat.com, peter.maydell@linaro.org,
	Darren Kenny <darren.kenny@oracle.com>,
	Alexander Bulekov <alxndr@bu.edu>
Subject: [PULL v3 4/8] fuzz: add an instrumentation filter
Date: Wed,  1 Sep 2021 07:50:00 -0400	[thread overview]
Message-ID: <20210901115004.34768-5-alxndr@bu.edu> (raw)
In-Reply-To: <20210901115004.34768-1-alxndr@bu.edu>

By default, -fsanitize=fuzzer instruments all code with coverage
information. However, this means that libfuzzer will track coverage over
hundreds of source files that are unrelated to virtual-devices. This
means that libfuzzer will optimize inputs for coverage observed in timer
code, memory APIs etc. This slows down the fuzzer and stores many inputs
that are not relevant to the actual virtual-devices.

With this change, clang versions that support the
"-fsanitize-coverage-allowlist" will only instrument a subset of the
compiled code, that is directly related to virtual-devices.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
---
 configure                                     | 28 +++++++++++++++----
 .../oss-fuzz/instrumentation-filter-template  | 15 ++++++++++
 2 files changed, 37 insertions(+), 6 deletions(-)
 create mode 100644 scripts/oss-fuzz/instrumentation-filter-template

diff --git a/configure b/configure
index 9a79a004d7..dcdbe3f068 100755
--- a/configure
+++ b/configure
@@ -4198,13 +4198,21 @@ fi
 
 ##########################################
 # checks for fuzzer
-if test "$fuzzing" = "yes" && test -z "${LIB_FUZZING_ENGINE+xxx}"; then
+if test "$fuzzing" = "yes" ; then
   write_c_fuzzer_skeleton
-  if compile_prog "$CPU_CFLAGS -Werror -fsanitize=fuzzer" ""; then
-    have_fuzzer=yes
-  else
-    error_exit "Your compiler doesn't support -fsanitize=fuzzer"
-    exit 1
+  if test -z "${LIB_FUZZING_ENGINE+xxx}"; then
+    if compile_prog "$CPU_CFLAGS -Werror -fsanitize=fuzzer" ""; then
+      have_fuzzer=yes
+    else
+      error_exit "Your compiler doesn't support -fsanitize=fuzzer"
+      exit 1
+    fi
+  fi
+
+  have_clang_coverage_filter=no
+  echo > $TMPTXT
+  if compile_prog "$CPU_CFLAGS -Werror -fsanitize=fuzzer -fsanitize-coverage-allowlist=$TMPTXT" ""; then
+    have_clang_coverage_filter=yes
   fi
 fi
 
@@ -4884,6 +4892,14 @@ if test "$fuzzing" = "yes" ; then
   else
     FUZZ_EXE_LDFLAGS="$LIB_FUZZING_ENGINE"
   fi
+
+  # Specify a filter to only instrument code that is directly related to
+  # virtual-devices.
+  if test "$have_clang_coverage_filter" = "yes" ; then
+    cp "$source_path/scripts/oss-fuzz/instrumentation-filter-template" \
+       instrumentation-filter
+    QEMU_CFLAGS="$QEMU_CFLAGS -fsanitize-coverage-allowlist=instrumentation-filter"
+  fi
 fi
 
 if test "$plugins" = "yes" ; then
diff --git a/scripts/oss-fuzz/instrumentation-filter-template b/scripts/oss-fuzz/instrumentation-filter-template
new file mode 100644
index 0000000000..76d2b6139a
--- /dev/null
+++ b/scripts/oss-fuzz/instrumentation-filter-template
@@ -0,0 +1,15 @@
+# Code that we actually want the fuzzer to target
+# See: https://clang.llvm.org/docs/SanitizerCoverage.html#disabling-instrumentation-without-source-modification
+#
+src:*/hw/*
+src:*/include/hw/*
+src:*/slirp/*
+src:*/net/*
+
+# We don't care about coverage over fuzzer-specific code, however we should
+# instrument the fuzzer entry-point so libFuzzer always sees at least some
+# coverage - otherwise it will exit after the first input
+src:*/tests/qtest/fuzz/fuzz.c
+
+# Enable instrumentation for all functions in those files
+fun:*
-- 
2.30.2



  parent reply	other threads:[~2021-09-01 11:56 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-01 11:49 [PULL v3 0/8] Fuzzing Patches for 2021-09-01 Alexander Bulekov
2021-09-01 11:49 ` [PULL v3 1/8] fuzz: fix sparse memory access in the DMA callback Alexander Bulekov
2021-09-01 11:49 ` [PULL v3 2/8] fuzz: adjust timeout to allow for longer inputs Alexander Bulekov
2021-09-01 11:49 ` [PULL v3 3/8] fuzz: make object-name matching case-insensitive Alexander Bulekov
2021-09-01 11:50 ` Alexander Bulekov [this message]
2021-09-01 11:50 ` [PULL v3 5/8] fuzz: use ITIMER_REAL for timeouts Alexander Bulekov
2021-09-01 11:50 ` [PULL v3 6/8] fuzz: unblock SIGALRM so the timeout works Alexander Bulekov
2021-09-01 11:50 ` [PULL v3 7/8] MAINTAINERS: Add myself as a reviewer for Device Fuzzing Alexander Bulekov
2021-09-01 11:50 ` [PULL v3 8/8] MAINTAINERS: add fuzzing reviewer Alexander Bulekov
2021-09-03  7:26 ` [PULL v3 0/8] Fuzzing Patches for 2021-09-01 Peter Maydell

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-all 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=20210901115004.34768-5-alxndr@bu.edu \
    --to=alxndr@bu.edu \
    --cc=bonzini@redhat.com \
    --cc=darren.kenny@oracle.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    --subject='Re: [PULL v3 4/8] fuzz: add an instrumentation filter' \
    /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

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.