Skip to contentSkip to conversation

Patch review

PATCH In review Steve Sakoman 4 participants 23 messages Nov 13, 2020 – Nov 16, 2020

lore
Share & tools
lore
#0 lore raw

cover Patch review

Please review this next set of patches for dunfell and have comments back by end of day on Tuesday.

Passed a-full on autobuilder:

https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/1588

The following changes since commit 4f395ad49ef9035954d0fda7b7df14dea18b49a0:

grub: clean up CVE patches (2020-11-08 16:51:24 -1000)

are available in the Git repository at:

git://git.openembedded.org/openembedded-core-contrib stable/dunfell-nut
http://cgit.openembedded.org/openembedded-core-contrib/log/?h=stable/dunfell-nut

Alex Kiernan (1):
openssh: Upgrade 8.2p1 -> 8.3p1

Alexander Kanavin (2):
clutter-gst-3.0: do not call out to host gstreamer plugin scanner
openssh: upgrade 8.3p1 -> 8.4p1 to fix CVE-2020-14145 and

    CVE-2020-15778

Denys Zagorui (1):
binutils: reproducibility: reuse debug-prefix-map for stabs

Diego Santa Cruz (1):
freetype: fix CVE-2020-15999, backport from 2.10.4

Gratian Crisan (1):
kernel-module-split.bbclass: identify kernel modconf files as
configuration files

Jose Quaresma (1):
gstreamer1.0: warn the user when something is wrong with GstBufferPool

Konrad Weihmann (3):
oeqa/core/context: expose results as variable
oeqa/core/context: initialize _run_end_time
testimage: print results for interrupted runs

Nathan Rossi (1):
diffstat: add nativesdk to BBCLASSEXTEND

Ricardo Salveti (1):
dosfstools: add mkfs.vfat to ALTERNATIVE

Richard Purdie (3):
ptest-runner: Fix license as it contains 'or later' clause
libdnf: Fix license as it contains 'or later' clause
alsa-utils: Fix license to GPLv2 only

Ross Burton (1):
syslinux: add link to upstream discussion in patch

Steve Sakoman (1):
openssh: whitelist CVE-2014-9278

akash hadke (1):
systemd: udev SECLABEL{selinux} crash fix

 meta/classes/kernel-module-split.bbclass      |  1 +
 meta/classes/testimage.bbclass                | 19 ++++---
 meta/lib/oeqa/core/context.py                 |  4 ++
 .../{openssh_8.2p1.bb => openssh_8.4p1.bb}    |  7 ++-
 ...temd-udev-seclabel-options-crash-fix.patch | 30 +++++++++++
 meta/recipes-core/systemd/systemd_244.3.bb    |  1 +
 .../binutils/binutils-2.34.inc                |  1 +
 ...oducibility-for-stabs-debugging-data.patch | 32 ++++++++++++
 .../diffstat/diffstat_1.63.bb                 |  2 +
 .../dosfstools/dosfstools_4.1.bb              |  6 ++-
 meta/recipes-devtools/libdnf/libdnf_0.28.1.bb |  2 +-
 ...nux-syslinux-support-ext2-3-4-device.patch |  2 +-
 .../clutter/clutter-gst-3.0.inc               |  2 +
 ...-sfnt-Fix-heap-buffer-overflow-59308.patch | 51 +++++++++++++++++++
 .../freetype/freetype_2.10.1.bb               |  1 +
 .../alsa/alsa-utils_1.2.1.bb                  |  3 +-
 ...size-in-reset-when-maxsize-is-larger.patch | 49 ++++++++++++++++++
 .../gstreamer/gstreamer1.0_1.16.3.bb          |  1 +
 .../ptest-runner/ptest-runner_2.3.2.bb        |  2 +-
 19 files changed, 201 insertions(+), 15 deletions(-)
 rename meta/recipes-connectivity/openssh/{openssh_8.2p1.bb => openssh_8.4p1.bb} (95%)
 create mode 100644 meta/recipes-core/systemd/systemd/systemd-udev-seclabel-options-crash-fix.patch
 create mode 100644 meta/recipes-devtools/binutils/binutils/0001-gas-improve-reproducibility-for-stabs-debugging-data.patch
 create mode 100644 meta/recipes-graphics/freetype/freetype/0001-sfnt-Fix-heap-buffer-overflow-59308.patch
 create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0/0005-bufferpool-only-resize-in-reset-when-maxsize-is-larger.patch
Signature · 1 line
2.17.1
#1 lore raw

PATCH 1/18 dosfstools: add mkfs.vfat to ALTERNATIVE

From: Ricardo Salveti <[email protected]>

The mkfs.vfat tool can also be provided by busybox via the CONFIG_MKFS_VFAT configuration (not enabled by default in OE but can be enabled on systems avoiding components based on GPLv3).

(cherry picked from commit 1227a29974671fd52014deaca7ac859a037cdeb5)
---
 meta/recipes-devtools/dosfstools/dosfstools_4.1.bb | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
meta/recipes-devtools/dosfstools/dosfstools_4.1.bb +5 -1
index 23b8836670..e4ab113391 100644
@@ -16,7 +16,7 @@ SRC_URI[sha256sum] = "e6b2aca70ccc3fe3687365009dd94a2e18e82b688ed4e260e04b741247
 
 UPSTREAM_CHECK_URI = "https://github.com/dosfstools/dosfstools/releases"
 
-inherit autotools pkgconfig
+inherit autotools pkgconfig update-alternatives
 
 EXTRA_OECONF = "--without-udev --enable-compat-symlinks"
 
@@ -26,3 +26,7 @@ BBCLASSEXTEND = "native"
 
 # Add codepage437 to avoid error from `dosfsck -l`
 RRECOMMENDS_${PN}_append_libc-glibc = " glibc-gconv-ibm437"
+
+ALTERNATIVE_PRIORITY = "100"
+ALTERNATIVE_${PN} = "mkfs.vfat"
+ALTERNATIVE_LINK_NAME[mkfs.vfat] = "${sbindir}/mkfs.vfat"
Signature · 1 line
2.17.1
#2 lore raw

PATCH 2/18 ptest-runner: Fix license as it contains 'or later' clause

In reply to Steve Sakoman

From: Richard Purdie <[email protected]>

The license headers are clear that the code is "or later", fix LICENSE to match.

(cherry picked from commit 5f0b5cdfcb104ac50222a47652e090ad8770e49f)
---
 meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb +1 -1
index 187f22df04..3a0dbf84fd 100644
@@ -4,7 +4,7 @@ program which loops through all installed ptest test suites and \
 runs them in sequence."
 HOMEPAGE = "http://git.yoctoproject.org/cgit/cgit.cgi/ptest-runner2/about/"
 
-LICENSE = "GPLv2"
+LICENSE = "GPLv2+"
 LIC_FILES_CHKSUM = "file://LICENSE;md5=751419260aa954499f7abaabaa882bbe"
 
 SRCREV = "7015e9199ce748c0717addeebe7a8c47448bab03"
Signature · 1 line
2.17.1
#5 lore raw

PATCH 3/18 libdnf: Fix license as it contains 'or later' clause

In reply to Steve Sakoman

From: Richard Purdie <[email protected]>

The license headers are clear that the code is "or later", fix LICENSE to match.

(cherry picked from commit e565e0b908c71ad5106d1c6c73d269b819787e55)
---
 meta/recipes-devtools/libdnf/libdnf_0.28.1.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-devtools/libdnf/libdnf_0.28.1.bb +1 -1
index cc2ceb8816..43de06e7f9 100644
@@ -1,5 +1,5 @@
 SUMMARY = "Library providing simplified C and Python API to libsolv"
-LICENSE = "LGPLv2.1"
+LICENSE = "LGPLv2.1+"
 LIC_FILES_CHKSUM = "file://COPYING;md5=4fbd65380cdd255951079008b364516c"
 
 SRC_URI = "git://github.com/rpm-software-management/libdnf \
Signature · 1 line
2.17.1
#6 lore raw

PATCH 4/18 alsa-utils: Fix license to GPLv2 only

In reply to Steve Sakoman

From: Richard Purdie <[email protected]>

Parts of alsa-utils are v2 only, parts are v2 or later. The effect is the end result is GPLv2 and there seems little value in marking everything as being a mixture of both. Fix LICENSE to match reality.

(cherry picked from commit a9a17a991174b732597e21045763ea851f486a01)
---
 meta/recipes-multimedia/alsa/alsa-utils_1.2.1.bb | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
meta/recipes-multimedia/alsa/alsa-utils_1.2.1.bb +2 -1
index 9144af628a..1dc30f377b 100644
@@ -2,7 +2,8 @@ SUMMARY = "ALSA sound utilities"
 HOMEPAGE = "http://www.alsa-project.org"
 BUGTRACKER = "http://alsa-project.org/main/index.php/Bug_Tracking"
 SECTION = "console/utils"
-LICENSE = "GPLv2+"
+# Some parts are GPLv2+, some are only GPLv2 (e.g. axfer, alsactl) so result is GPLv2
+LICENSE = "GPLv2"
 LIC_FILES_CHKSUM = "file://COPYING;md5=59530bdf33659b29e73d4adb9f9f6552 \
                     file://alsactl/utils.c;beginline=3;endline=18;md5=96cc06a4cebe5eb7975688ffb0e65642"
 DEPENDS = "alsa-lib ncurses libsamplerate0"
Signature · 1 line
2.17.1
#7 lore raw

PATCH 5/18 kernel-module-split.bbclass: identify kernel modconf files as configuration files

In reply to Steve Sakoman

From: Gratian Crisan <[email protected]>

Currently the modconf fragments representing the configuration for kernel modules are written out to appropriate .conf files and added to the FILES variable. However they are not identified as 'configuration files' and installing a new version of a kernel module results in a conflict and a failed installed because the respective .conf file is already in place from a previous install.

Add the generated .conf files to the CONFFILES variable denoting their true nature.

(cherry picked from commit 1a70a92d1f1006be115429a4262259c9084f484d)
---
 meta/classes/kernel-module-split.bbclass | 1 +
 1 file changed, 1 insertion(+)
meta/classes/kernel-module-split.bbclass +1 -0
index 221022b7bc..c8ede26996 100644
@@ -120,6 +120,7 @@ python split_kernel_module_packages () {
         files = d.getVar('FILES_%s' % pkg)
         files = "%s /etc/modules-load.d/%s.conf /etc/modprobe.d/%s.conf" % (files, basename, basename)
         d.setVar('FILES_%s' % pkg, files)
+        d.setVar('CONFFILES_%s' % pkg, files)
 
         if "description" in vals:
             old_desc = d.getVar('DESCRIPTION_' + pkg) or ""
Signature · 1 line
2.17.1
#8 lore raw

PATCH 6/18 syslinux: add link to upstream discussion in patch

In reply to Steve Sakoman

From: Ross Burton <[email protected]>

(cherry picked from commit dfc2b114e9d62f0eee04129009a24a8edb2a8dd1)
---
 .../syslinux/0001-linux-syslinux-support-ext2-3-4-device.patch  | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-devtools/syslinux/syslinux/0001-linux-syslinux-support-ext2-3-4-device.patch +1 -1
index 3ab7875274..47a8dac10e 100644
@@ -10,7 +10,7 @@ Subject: [PATCH 1/9] linux/syslinux: support ext2/3/4 device
 * The ext2/3/4 support doesn't require root privileges since it doesn't need
   mount (but write permission is required).
 
-Upstream-Status: Submitted
+Upstream-Status: Submitted [https://www.syslinux.org/archives/2015-January/023039.html]
 
 Signed-off-by: Robert Yang <[email protected]>
 Tested-by: Du Dolpher <[email protected]>
Signature · 1 line
2.17.1
#9 lore raw

PATCH 7/18 diffstat: add nativesdk to BBCLASSEXTEND

In reply to Steve Sakoman

From: Nathan Rossi <[email protected]>

The diffstat tool is part of HOSTTOOLS. To support hosts that do not have it installed with buildtools-tarball it must be enabled for nativesdk.

(cherry picked from commit 0ed002422bc46539f1d71ed19ee17358b6691bf0)
---
 meta/recipes-devtools/diffstat/diffstat_1.63.bb | 2 ++
 1 file changed, 2 insertions(+)
meta/recipes-devtools/diffstat/diffstat_1.63.bb +2 -0
index ee0a7e7bd6..61b2ea5dc2 100644
@@ -27,3 +27,5 @@ LDFLAGS += "${TOOLCHAIN_OPTIONS}"
 do_install_ptest() {
 	cp -r ${S}/testing ${D}${PTEST_PATH}
 }
+
+BBCLASSEXTEND = "nativesdk"
Signature · 1 line
2.17.1
#10 lore raw

PATCH 8/18 clutter-gst-3.0: do not call out to host gstreamer plugin scanner

In reply to Steve Sakoman

From: Alexander Kanavin <[email protected]>

This is host contamination and can also fail for all kinds of reasons when running under usermode qemu.

(cherry picked from commit fb60d0920b660dffb346b2212dc6f8ba2a0b9fde)
---
 meta/recipes-graphics/clutter/clutter-gst-3.0.inc | 2 ++
 1 file changed, 2 insertions(+)
meta/recipes-graphics/clutter/clutter-gst-3.0.inc +2 -0
index fc3eade886..7d9db1f38c 100644
@@ -6,6 +6,8 @@ inherit clutter features_check upstream-version-is-even gobject-introspection
 # depends on clutter-1.0 which depends on cogl-1.0
 REQUIRED_DISTRO_FEATURES ?= "opengl"
 
+export GST_PLUGIN_SCANNER_1_0="${S}/gst-plugin-scanner-dummy"
+
 SRC_URI += "file://0001-Install-example-binary-needed-for-core-image-clutter.patch"
 
 DEPENDS = "gstreamer1.0-plugins-base gstreamer1.0-plugins-bad clutter-1.0 libgudev"
Signature · 1 line
2.17.1
#11 lore raw

PATCH 9/18 gstreamer1.0: warn the user when something is wrong with GstBufferPool

In reply to Steve Sakoman

From: Jose Quaresma <[email protected]>

This is not a critical bug fix but it can be usefull in some BSP with exotic drivers like on nvidia tegra bsp.

---
 ...size-in-reset-when-maxsize-is-larger.patch | 49 +++++++++++++++++++
 .../gstreamer/gstreamer1.0_1.16.3.bb          |  1 +
 2 files changed, 50 insertions(+)
 create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0/0005-bufferpool-only-resize-in-reset-when-maxsize-is-larger.patch
meta/recipes-multimedia/gstreamer/gstreamer1.0/0005-bufferpool-only-resize-in-reset-when-maxsize-is-larger.patch new file +49 -0
new file mode 100644
index 0000000000..dacc399d28
@@ -0,0 +1,49 @@
+From a1b41b2b2493069365a8274c6a544e6799a5a8df Mon Sep 17 00:00:00 2001
+From: Matthew Waters <[email protected]>
+Date: Mon, 20 Jul 2020 17:08:32 +1000
+Subject: [PATCH] gst/bufferpool: only resize in reset when maxsize is larger
+
+Only resize the buffer if the maxsize is larger then the configued pool
+size.
+
+Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/570>
+
+Upstream-Status: Backport [https://gitlab.freedesktop.org/gstreamer/gstreamer/-/commit/a1b41b2b2493069365a8274c6a544e6799a5a8df]
+
+Signed-off-by: Jose Quaresma <[email protected]>
+---
+ gst/gstbufferpool.c | 18 +++++++++++++++---
+ 1 file changed, 15 insertions(+), 3 deletions(-)
+
+diff --git a/gst/gstbufferpool.c b/gst/gstbufferpool.c
+index 8ae868cf2c7..a8167d017d6 100644
+--- a/gst/gstbufferpool.c
++++ b/gst/gstbufferpool.c
+@@ -1223,9 +1223,21 @@ default_reset_buffer (GstBufferPool * pool, GstBuffer * buffer)
+ 
+   /* if the memory is intact reset the size to the full size */
+   if (!GST_BUFFER_FLAG_IS_SET (buffer, GST_BUFFER_FLAG_TAG_MEMORY)) {
+-    gsize offset;
+-    gst_buffer_get_sizes (buffer, &offset, NULL);
+-    gst_buffer_resize (buffer, -offset, pool->priv->size);
++    gsize offset, maxsize;
++    gst_buffer_get_sizes (buffer, &offset, &maxsize);
++    /* check if we can resize to at least the pool configured size.  If not,
++     * then this will fail internally in gst_buffer_resize().
++     * default_release_buffer() will drop the buffer from the pool if the
++     * sizes don't match */
++    if (maxsize >= pool->priv->size) {
++      gst_buffer_resize (buffer, -offset, pool->priv->size);
++    } else {
++      GST_WARNING_OBJECT (pool, "Buffer %p without the memory tag has "
++          "maxsize (%" G_GSIZE_FORMAT ") that is smaller than the "
++          "configured buffer pool size (%u). The buffer will be not be "
++          "reused. This is most likely a bug in this GstBufferPool subclass",
++          buffer, maxsize, pool->priv->size);
++    }
+   }
+ 
+   /* remove all metadata without the POOLED flag */
+-- 
+GitLab
+
meta/recipes-multimedia/gstreamer/gstreamer1.0_1.16.3.bb +1 -0
index 66ad3e3381..7afe56cd7b 100644
@@ -21,6 +21,7 @@ SRC_URI = " \
     file://0002-meson-build-gir-even-when-cross-compiling-if-introsp.patch \
     file://0003-meson-Add-valgrind-feature.patch \
     file://0004-meson-Add-option-for-installed-tests.patch \
+    file://0005-bufferpool-only-resize-in-reset-when-maxsize-is-larger.patch \
 "
 SRC_URI[md5sum] = "beecf6965a17fb17fa3b262fd36df70a"
 SRC_URI[sha256sum] = "692f037968e454e508b0f71d9674e2e26c78475021407fcf8193b1c7e59543c7"
Signature · 1 line
2.17.1
#12 lore raw

PATCH 10/18 systemd: udev SECLABEL{selinux} crash fix

In reply to Steve Sakoman

From: akash hadke <[email protected]>

Adding SECLABEL{selinux}="some value" causes systemd-udev to crash.
So applied below available patch to fix the issue.

systemd-udev-seclabel-options-crash-fix.patch

---
 ...temd-udev-seclabel-options-crash-fix.patch | 30 +++++++++++++++++++
 meta/recipes-core/systemd/systemd_244.3.bb    |  1 +
 2 files changed, 31 insertions(+)
 create mode 100644 meta/recipes-core/systemd/systemd/systemd-udev-seclabel-options-crash-fix.patch
meta/recipes-core/systemd/systemd/systemd-udev-seclabel-options-crash-fix.patch new file +30 -0
new file mode 100644
index 0000000000..27b2b60fad
@@ -0,0 +1,30 @@
+From 0335d110afc08baf47d76b7011ce02510dfdd524 Mon Sep 17 00:00:00 2001
+From: Valery0xff <[email protected]>
+Date: Wed, 11 Mar 2020 02:20:36 +0200
+Subject: [PATCH] udev: fix SECLABEL{selinux} issue (#15064)
+
+Add SECLABEL{selinux}="some value" cause udevadm crash
+systemd-udevd[x]: Worker [x] terminated by signal 11 (SEGV)
+ 
+It happens since 25de7aa7b90 (Yu Watanabe 2019-04-25 01:21:11 +0200)
+when udev rules processing changed to token model. Yu forgot store
+attr to SECLABEL token so fix it.
+---
+ src/udev/udev-rules.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+Upstream-Status: Backport [https://github.com/systemd/systemd/commit/0335d110afc08baf47d76b7011ce02510dfdd524.patch]
+---
+diff --git a/src/udev/udev-rules.c b/src/udev/udev-rules.c
+index b9b350d1ef..b990f68e93 100644
+--- a/src/udev/udev-rules.c
++++ b/src/udev/udev-rules.c
+@@ -921,7 +921,7 @@ static int parse_token(UdevRules *rules, const char *key, char *attr, UdevRuleOp
+                         op = OP_ASSIGN;
+                 }
+ 
+-                r = rule_line_add_token(rule_line, TK_A_SECLABEL, op, value, NULL);
++                r = rule_line_add_token(rule_line, TK_A_SECLABEL, op, value, attr);
+         } else if (streq(key, "RUN")) {
+                 if (is_match || op == OP_REMOVE)
+                         return log_token_invalid_op(rules, key);
meta/recipes-core/systemd/systemd_244.3.bb +1 -0
index 850d64e8b0..64e3b18333 100644
@@ -21,6 +21,7 @@ SRC_URI += "file://touchscreen.rules \
            file://0001-binfmt-Don-t-install-dependency-links-at-install-tim.patch \
            file://0003-implment-systemd-sysv-install-for-OE.patch \
            file://CVE-2020-13776.patch \
+           file://systemd-udev-seclabel-options-crash-fix.patch \
            "
 
 # patches needed by musl
Signature · 1 line
2.17.1
#13 lore raw

PATCH 11/18 oeqa/core/context: expose results as variable

In reply to Steve Sakoman

From: Konrad Weihmann <[email protected]>

register an unittest handler for testresults and expose it as variable result.
With this even partial results from an interrupted test suite run can be made available

(cherry picked from commit a97ae47525157871b6c098ffc352293e365a4335)
---
 meta/lib/oeqa/core/context.py | 3 +++
 1 file changed, 3 insertions(+)
meta/lib/oeqa/core/context.py +3 -0
index 4705d608ac..4ce3089ad0 100644
@@ -31,6 +31,9 @@ class OETestContext(object):
         self._registry = {}
         self._registry['cases'] = collections.OrderedDict()
 
+        self.results = unittest.TestResult()
+        unittest.registerResult(self.results)
+
     def _read_modules_from_manifest(self, manifest):
         if not os.path.exists(manifest):
             raise OEQAMissingManifest("Manifest does not exist on %s" % manifest)
Signature · 1 line
2.17.1
#14 lore raw

PATCH 12/18 oeqa/core/context: initialize _run_end_time

In reply to Steve Sakoman

From: Konrad Weihmann <[email protected]>

with _run_start_time as value. For partial results of interrupted runs, this info might be otherwise missing for at least one testcase

(cherry picked from commit 1c5e8baf57fa2a33b9ef507b11d9ea9acaa77238)
---
 meta/lib/oeqa/core/context.py | 1 +
 1 file changed, 1 insertion(+)
meta/lib/oeqa/core/context.py +1 -0
index 4ce3089ad0..4e238f80a4 100644
@@ -85,6 +85,7 @@ class OETestContext(object):
         self.skipTests(skips)
 
         self._run_start_time = time.time()
+        self._run_end_time = self._run_start_time
         if not processes:
             self.runner.buffer = True
         result = self.runner.run(self.prepareSuite(self.suites, processes))
Signature · 1 line
2.17.1
#15 lore raw

PATCH 13/18 testimage: print results for interrupted runs

In reply to Steve Sakoman

From: Konrad Weihmann <[email protected]>

When a run is ended by overall timeout, print the already executed testcases, to provide some hints which testcase might made the test suite reach global timeout.
Nonetheless make the testrun exit with an error

(cherry picked from commit 2bcc643195a3b3c66d698fac8b7af037c08545ac)
---
 meta/classes/testimage.bbclass | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)
meta/classes/testimage.bbclass +11 -8
index 00f0c29836..c709384b91 100644
@@ -364,6 +364,7 @@ def testimage_main(d):
     package_extraction(d, tc.suites)
 
     results = None
+    complete = False
     orig_sigterm_handler = signal.signal(signal.SIGTERM, sigterm_exception)
     try:
         # We need to check if runqemu ends unexpectedly
@@ -375,6 +376,7 @@ def testimage_main(d):
         except ValueError:
             pass
         results = tc.runTests()
+        complete = True
     except (KeyboardInterrupt, BlockingIOError) as err:
         if isinstance(err, KeyboardInterrupt):
             bb.error('testimage interrupted, shutting down...')
@@ -382,20 +384,21 @@ def testimage_main(d):
             bb.error('runqemu failed, shutting down...')
         if results:
             results.stop()
-            results = None
+        results = tc.results
     finally:
         signal.signal(signal.SIGTERM, orig_sigterm_handler)
         tc.target.stop()
 
     # Show results (if we have them)
-    if not results:
+    if results:
+        configuration = get_testimage_configuration(d, 'runtime', machine)
+        results.logDetails(get_testimage_json_result_dir(d),
+                        configuration,
+                        get_testimage_result_id(configuration),
+                        dump_streams=d.getVar('TESTREPORT_FULLLOGS'))
+        results.logSummary(pn)
+    if not results or not complete:
         bb.fatal('%s - FAILED - tests were interrupted during execution' % pn, forcelog=True)
-    configuration = get_testimage_configuration(d, 'runtime', machine)
-    results.logDetails(get_testimage_json_result_dir(d),
-                       configuration,
-                       get_testimage_result_id(configuration),
-                       dump_streams=d.getVar('TESTREPORT_FULLLOGS'))
-    results.logSummary(pn)
     if not results.wasSuccessful():
         bb.fatal('%s - FAILED - check the task log and the ssh log' % pn, forcelog=True)
 
Signature · 1 line
2.17.1
#16 lore raw

PATCH 14/18 binutils: reproducibility: reuse debug-prefix-map for stabs

In reply to Steve Sakoman

From: Denys Zagorui <[email protected]>

powerpc 32bit Linux Kernel widely uses .stabs pseudo-op to produce debugging information in stabs format. Faced an issue that during Linux Kernel build with Yocto build system for 32bit powerpc platform resulting vmlinux contains absolute path in .stabstr section that cannot be remapped with -fdebug-prefix-map option.

Yocto uses scripts/mkmakefile Linux Kernel build approach that allows to store all generated files outside of kernel source tree. With this approach each compilier invocation is performed with an absolute path to a file that will be compiled and this absolute path is recorded in init stab. There is no way to remap this path.

Reuse remap_debug_filename api to make -fdebug-prefix-map flag aplicable for init stab.

---
 .../binutils/binutils-2.34.inc                |  1 +
 ...oducibility-for-stabs-debugging-data.patch | 32 +++++++++++++++++++
 2 files changed, 33 insertions(+)
 create mode 100644 meta/recipes-devtools/binutils/binutils/0001-gas-improve-reproducibility-for-stabs-debugging-data.patch
meta/recipes-devtools/binutils/binutils-2.34.inc +1 -0
index f65fdb7328..b5f5a1c69a 100644
@@ -43,5 +43,6 @@ SRC_URI = "\
      file://0016-Check-for-clang-before-checking-gcc-version.patch \
      file://0017-binutils-drop-redundant-program_name-definition-fno-.patch \
      file://CVE-2020-0551.patch \
+     file://0001-gas-improve-reproducibility-for-stabs-debugging-data.patch \
 "
 S  = "${WORKDIR}/git"
meta/recipes-devtools/binutils/binutils/0001-gas-improve-reproducibility-for-stabs-debugging-data.patch new file +32 -0
new file mode 100644
index 0000000000..98dab1b41d
@@ -0,0 +1,32 @@
+From 96882eb9263b6b1c953e33a6764a83f4a87a9a41 Mon Sep 17 00:00:00 2001
+From: Denys Zagorui <[email protected]>
+Date: Mon, 9 Nov 2020 15:39:10 +0000
+Subject: [PATCH] gas: improve reproducibility for stabs debugging data format
+
+	* config/obj-elf (obj_elf_init_stab_section): Improve
+	reproducibility for stabs debugging data format
+
+Upstream-Status: Backport [https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=0541201782c006c09d029d18a45c6e743cfea906]
+---
+ gas/config/obj-elf.c | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/gas/config/obj-elf.c b/gas/config/obj-elf.c
+index 7cf921c051..92b55e06c2 100644
+--- a/gas/config/obj-elf.c
++++ b/gas/config/obj-elf.c
+@@ -2186,12 +2186,13 @@ obj_elf_init_stab_section (segT seg)
+   p = frag_more (12);
+   /* Zero it out.  */
+   memset (p, 0, 12);
+-  file = as_where (NULL);
++  file = remap_debug_filename (as_where (NULL));
+   stabstr_name = concat (segment_name (seg), "str", (char *) NULL);
+   stroff = get_stab_string_offset (file, stabstr_name, TRUE);
+   know (stroff == 1 || (stroff == 0 && file[0] == '\0'));
+   md_number_to_chars (p, stroff, 4);
+   seg_info (seg)->stabu.p = p;
++  xfree ((char *) file);
+ }
+ 
+ #endif
Signature · 1 line
2.17.1
#17 lore raw

PATCH 15/18 openssh: Upgrade 8.2p1 -> 8.3p1

In reply to Steve Sakoman

From: Alex Kiernan <[email protected]>

(cherry picked from commit b8ec59302bd2fc1a78f4d828ba93b3ad64ab7f37)
---
 .../openssh/{openssh_8.2p1.bb => openssh_8.3p1.bb}             | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
 rename meta/recipes-connectivity/openssh/{openssh_8.2p1.bb => openssh_8.3p1.bb} (98%)
meta/recipes-connectivity/openssh/openssh_8.3p1.bb renamed from meta/recipes-connectivity/openssh/openssh_8.2p1.bb +1 -2
similarity index 98%
rename from meta/recipes-connectivity/openssh/openssh_8.2p1.bb
rename to meta/recipes-connectivity/openssh/openssh_8.3p1.bb
index 17138c0ad8..e007328704 100644
@@ -25,8 +25,7 @@ SRC_URI = "http://ftp.openbsd.org/pub/OpenBSD/OpenSSH/portable/openssh-${PV}.tar
            file://sshd_check_keys \
            file://add-test-support-for-busybox.patch \
            "
-SRC_URI[md5sum] = "3076e6413e8dbe56d33848c1054ac091"
-SRC_URI[sha256sum] = "43925151e6cf6cee1450190c0e9af4dc36b41c12737619edff8bcebdff64e671"
+SRC_URI[sha256sum] = "f2befbe0472fe7eb75d23340eb17531cb6b3aac24075e2066b41f814e12387b2"
 
 PAM_SRC_URI = "file://sshd"
 
Signature · 1 line
2.17.1
#18 lore raw

PATCH 16/18 openssh: upgrade 8.3p1 -> 8.4p1 to fix CVE-2020-14145 and CVE-2020-15778

In reply to Steve Sakoman

From: Alexander Kanavin <[email protected]>

CVE-2020-14145

The client side in OpenSSH 5.7 through 8.3 has an Observable Discrepancy leading to an information leak in the algorithm negotiation. This allows man-in-the-middle attackers to target initial connection attempts (where no host key for the server has been cached by the client).

CVE-2020-15778

scp in OpenSSH through 8.3p1 allows command injection in the scp.c toremote function, as demonstrated by backtick characters in the destination argument.

(cherry picked from commit fc394ade8a0033bc695d979e592e8e92a882c54d)
---
 .../openssh/{openssh_8.3p1.bb => openssh_8.4p1.bb}              | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
 rename meta/recipes-connectivity/openssh/{openssh_8.3p1.bb => openssh_8.4p1.bb} (98%)
meta/recipes-connectivity/openssh/openssh_8.4p1.bb renamed from meta/recipes-connectivity/openssh/openssh_8.3p1.bb +1 -1
similarity index 98%
rename from meta/recipes-connectivity/openssh/openssh_8.3p1.bb
rename to meta/recipes-connectivity/openssh/openssh_8.4p1.bb
index e007328704..720b238e71 100644
@@ -25,7 +25,7 @@ SRC_URI = "http://ftp.openbsd.org/pub/OpenBSD/OpenSSH/portable/openssh-${PV}.tar
            file://sshd_check_keys \
            file://add-test-support-for-busybox.patch \
            "
-SRC_URI[sha256sum] = "f2befbe0472fe7eb75d23340eb17531cb6b3aac24075e2066b41f814e12387b2"
+SRC_URI[sha256sum] = "5a01d22e407eb1c05ba8a8f7c654d388a13e9f226e4ed33bd38748dafa1d2b24"
 
 PAM_SRC_URI = "file://sshd"
 
Signature · 1 line
2.17.1
#21 lore raw

PATCH 17/18 openssh: whitelist CVE-2014-9278

In reply to Steve Sakoman

The OpenSSH server, as used in Fedora and Red Hat Enterprise Linux 7 and when running in a Kerberos environment, allows remote authenticated users to log in as another user when they are listed in the .k5users file of that user, which might bypass intended authentication requirements that would force a local login.

Whitelist the CVE since this issue is Redhat specific.

(cherry picked from commit 309132e50d23b1e3f15ef8db1a101166b35f7ca4)
---
 meta/recipes-connectivity/openssh/openssh_8.4p1.bb | 4 ++++
 1 file changed, 4 insertions(+)
meta/recipes-connectivity/openssh/openssh_8.4p1.bb +4 -0
index 720b238e71..676a8a6533 100644
@@ -27,6 +27,10 @@ SRC_URI = "http://ftp.openbsd.org/pub/OpenBSD/OpenSSH/portable/openssh-${PV}.tar
            "
 SRC_URI[sha256sum] = "5a01d22e407eb1c05ba8a8f7c654d388a13e9f226e4ed33bd38748dafa1d2b24"
 
+# This CVE is specific to OpenSSH server, as used in Fedora and Red Hat Enterprise Linux 7
+# and when running in a Kerberos environment. As such it is not relevant to OpenEmbedded
+CVE_CHECK_WHITELIST += "CVE-2014-9278"
+
 PAM_SRC_URI = "file://sshd"
 
 inherit manpages useradd update-rc.d update-alternatives systemd
Signature · 1 line
2.17.1
#22 lore raw

PATCH 18/18 freetype: fix CVE-2020-15999, backport from 2.10.4

In reply to Steve Sakoman

From: Diego Santa Cruz <[email protected]>

---
 ...-sfnt-Fix-heap-buffer-overflow-59308.patch | 51 +++++++++++++++++++
 .../freetype/freetype_2.10.1.bb               |  1 +
 2 files changed, 52 insertions(+)
 create mode 100644 meta/recipes-graphics/freetype/freetype/0001-sfnt-Fix-heap-buffer-overflow-59308.patch
meta/recipes-graphics/freetype/freetype/0001-sfnt-Fix-heap-buffer-overflow-59308.patch new file +51 -0
new file mode 100644
index 0000000000..fa8a29b798
@@ -0,0 +1,51 @@
+From a3bab162b2ae616074c8877a04556932998aeacd Mon Sep 17 00:00:00 2001
+From: Werner Lemberg <[email protected]>
+Date: Mon, 19 Oct 2020 23:45:28 +0200
+Subject: [PATCH] [sfnt] Fix heap buffer overflow (#59308).
+
+This is CVE-2020-15999.
+
+* src/sfnt/pngshim.c (Load_SBit_Png): Test bitmap size earlier.
+
+Upstream-Status: Backport [https://git.savannah.gnu.org/cgit/freetype/freetype2.git/commit/?id=a3bab162b2ae616074c8877a04556932998aeacd]
+
+Signed-off-by: Diego Santa Cruz <[email protected]>
+---
+ src/sfnt/pngshim.c | 14 +++++++-------
+ 1 file changed, 7 insertions(+), 7 deletions(-)
+
+diff --git a/src/sfnt/pngshim.c b/src/sfnt/pngshim.c
+index 2e64e5846..f55016122 100644
+--- a/src/sfnt/pngshim.c
++++ b/src/sfnt/pngshim.c
+@@ -332,6 +332,13 @@
+ 
+     if ( populate_map_and_metrics )
+     {
++      /* reject too large bitmaps similarly to the rasterizer */
++      if ( imgHeight > 0x7FFF || imgWidth > 0x7FFF )
++      {
++        error = FT_THROW( Array_Too_Large );
++        goto DestroyExit;
++      }
++
+       metrics->width  = (FT_UShort)imgWidth;
+       metrics->height = (FT_UShort)imgHeight;
+ 
+@@ -340,13 +347,6 @@
+       map->pixel_mode = FT_PIXEL_MODE_BGRA;
+       map->pitch      = (int)( map->width * 4 );
+       map->num_grays  = 256;
+-
+-      /* reject too large bitmaps similarly to the rasterizer */
+-      if ( map->rows > 0x7FFF || map->width > 0x7FFF )
+-      {
+-        error = FT_THROW( Array_Too_Large );
+-        goto DestroyExit;
+-      }
+     }
+ 
+     /* convert palette/gray image to rgb */
+-- 
+2.18.4
+
meta/recipes-graphics/freetype/freetype_2.10.1.bb +1 -0
index d1c093054b..2d444bbf19 100644
@@ -14,6 +14,7 @@ LIC_FILES_CHKSUM = "file://docs/LICENSE.TXT;md5=4af6221506f202774ef74f64932878a1
 
 SRC_URI = "${SAVANNAH_NONGNU_MIRROR}/${BPN}/${BP}.tar.xz \
            file://use-right-libtool.patch \
+           file://0001-sfnt-Fix-heap-buffer-overflow-59308.patch \
           "
 SRC_URI[md5sum] = "bd42e75127f8431923679480efb5ba8f"
 SRC_URI[sha256sum] = "16dbfa488a21fe827dc27eaf708f42f7aa3bb997d745d31a19781628c36ba26f"
Signature · 1 line
2.17.1
#19 lore raw

Re: [OE-core][dunfell 16/18] openssh: upgrade 8.3p1 -> 8.4p1 to fix CVE-2020-14145 and CVE-2020-15778

In reply to Steve Sakoman

Hi Steve,

Quoted 9 lines · Steve Sakoman

From: Alexander Kanavin <[email protected]>

CVE-2020-14145

The client side in OpenSSH 5.7 through 8.3 has an Observable Discrepancy leading to an information leak in the algorithm negotiation. This allows man-in-the-middle attackers to target initial connection attempts (where no host key for the server has been cached by the client).

I am not sure if this CVE should be considered fixed. Please see Section 3.1:

https://www.fzi.de/fileadmin/user_upload/2020-06-26-FSA-2020-2.pdf

Also, this isn't a bug fix release and has potentially incompatible changes that may affect existing configurations as per the release notes:

https://www.openssh.com/txt/release-8.4

Thanks,

Anuj

#20 lore raw

Re: [OE-core][dunfell 16/18] openssh: upgrade 8.3p1 -> 8.4p1 to fix CVE-2020-14145 and CVE-2020-15778

Quoted 14 lines · Anuj Mittal

Hi Steve,

Quoted 9 lines · Steve Sakoman

From: Alexander Kanavin <[email protected]>

CVE-2020-14145

The client side in OpenSSH 5.7 through 8.3 has an Observable Discrepancy leading to an information leak in the algorithm negotiation. This allows man-in-the-middle attackers to target initial connection attempts (where no host key for the server has been cached by the client).

Quoted 4 lines

I am not sure if this CVE should be considered fixed. Please see Section 3.1:

https://www.fzi.de/fileadmin/user_upload/2020-06-26-FSA-2020-2.pdf

I hadn't seen this. I'll drop the version upgrade from the pull request.

Hopefully someone can submit a patch set that does fix these issues.

Steve

Quoted 13 lines

Also, this isn't a bug fix release and has potentially incompatible changes that may affect existing configurations as per the release notes:

https://www.openssh.com/txt/release-8.4

Thanks,

Anuj

#3 lore raw

Re: [OE-core][dunfell 02/18] ptest-runner: Fix license as it contains 'or later' clause

In reply to Steve Sakoman

Hi Steve

There is a nasty problem with ptest-runner 2.3.2: It prints an internal warning for each test case, which got fixed by
commit: d3e9e5ebf36d1d1d7ee4bc38a64f6ec9e8fc819b
"Backport patch to fix inappropriate ioctl error"

You might want to consider updating ptest-runner to the latest version by picking 4 commits instead of just one.

- ptest-runner: Bump to 2.4.0 792d1ebba98b097d4d68c7261bf6af0c9d186eab
- ptest-runner: fix upstream version check 5df6dc44370c446c699007659782323814a4d9ce
- ptest-runner: Backport patch to fix inappropriate ioctl error d3e9e5ebf36d1d1d7ee4bc38a64f6ec9e8fc819b
- ptest-runner: Fix license as it contains 'or later' clause 744a463831016abdac3433db8ca9d1b656881915

We did this a few weeks ago on our local dunfell branch. So far it seems to be the right approach for us.

Thank you and regards,
Adrian

Quoted 19 lines · Steve Sakoman

From: Richard Purdie <[email protected]>

The license headers are clear that the code is "or later", fix LICENSE to match.

(cherry picked from commit 5f0b5cdfcb104ac50222a47652e090ad8770e49f)
---
 meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb +0 -0
index 187f22df04..3a0dbf84fd 100644

meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb program which loops through all installed ptest test suites and \ · line 10

Quoted 8 lines
meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb +1 -1
@@ -4,7 +4,7 @@ program which loops through all installed ptest test suites and \
 runs them in sequence."
 HOMEPAGE = "http://git.yoctoproject.org/cgit/cgit.cgi/ptest-runner2/about/"
 
-LICENSE = "GPLv2"
+LICENSE = "GPLv2+"
 LIC_FILES_CHKSUM = "file://LICENSE;md5=751419260aa954499f7abaabaa882bbe"
 
 SRCREV = "7015e9199ce748c0717addeebe7a8c47448bab03"

We have been doing this without any problems for a few weeks now.

#4 lore raw

Re: [OE-core][dunfell 02/18] ptest-runner: Fix license as it contains 'or later' clause

Quoted 20 lines · Adrian Freihofer

Hi Steve

There is a nasty problem with ptest-runner 2.3.2: It prints an internal warning for each test case, which got fixed by
commit: d3e9e5ebf36d1d1d7ee4bc38a64f6ec9e8fc819b
"Backport patch to fix inappropriate ioctl error"

You might want to consider updating ptest-runner to the latest version by picking 4 commits instead of just one.

- ptest-runner: Bump to 2.4.0 792d1ebba98b097d4d68c7261bf6af0c9d186eab
- ptest-runner: fix upstream version check 5df6dc44370c446c699007659782323814a4d9ce
- ptest-runner: Backport patch to fix inappropriate ioctl error d3e9e5ebf36d1d1d7ee4bc38a64f6ec9e8fc819b
- ptest-runner: Fix license as it contains 'or later' clause 744a463831016abdac3433db8ca9d1b656881915

Quoted 2 lines

We did this a few weeks ago on our local dunfell branch. So far it seems to be the right approach for us.

This looks reasonable to me. I'll add these to my next autobuilder test run and if all goes well will submit to the list for review/comments.

Steve

Quoted 37 lines

Thank you and regards,
Adrian

Quoted 19 lines · Steve Sakoman

From: Richard Purdie <[email protected]>

The license headers are clear that the code is "or later", fix LICENSE to match.

(cherry picked from commit 5f0b5cdfcb104ac50222a47652e090ad8770e49f)
---
 meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb +0 -0
index 187f22df04..3a0dbf84fd 100644

meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb program which loops through all installed ptest test suites and \ · line 10

Quoted 8 lines
meta/recipes-support/ptest-runner/ptest-runner_2.3.2.bb +1 -1
@@ -4,7 +4,7 @@ program which loops through all installed ptest test suites and \
 runs them in sequence."
 HOMEPAGE = "http://git.yoctoproject.org/cgit/cgit.cgi/ptest-runner2/about/"
 
-LICENSE = "GPLv2"
+LICENSE = "GPLv2+"
 LIC_FILES_CHKSUM = "file://LICENSE;md5=751419260aa954499f7abaabaa882bbe"
 
 SRCREV = "7015e9199ce748c0717addeebe7a8c47448bab03"

We have been doing this without any problems for a few weeks now.