Skip to contentSkip to conversation

libxml2: fix CVE-2020-7595

PATCH In review Adrian Bunk 2 participants 5 messages Feb 21, 2020 – Feb 23, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 libxml2: fix CVE-2020-7595

From: Anuj Mittal <[email protected]>

---
 .../libxml/libxml2/CVE-2020-7595.patch        | 36 +++++++++++++++++++
 meta/recipes-core/libxml/libxml2_2.9.9.bb     |  1 +
 2 files changed, 37 insertions(+)
 create mode 100644 meta/recipes-core/libxml/libxml2/CVE-2020-7595.patch
meta/recipes-core/libxml/libxml2/CVE-2020-7595.patch new file +36 -0
new file mode 100644
index 0000000000..facfefd362
@@ -0,0 +1,36 @@
+From 0e1a49c8907645d2e155f0d89d4d9895ac5112b5 Mon Sep 17 00:00:00 2001
+From: Zhipeng Xie <[email protected]>
+Date: Thu, 12 Dec 2019 17:30:55 +0800
+Subject: [PATCH] Fix infinite loop in xmlStringLenDecodeEntities
+
+When ctxt->instate == XML_PARSER_EOF,xmlParseStringEntityRef
+return NULL which cause a infinite loop in xmlStringLenDecodeEntities
+
+Found with libFuzzer.
+
+Signed-off-by: Zhipeng Xie <[email protected]>
+
+Upstream-Status: Backport [https://gitlab.gnome.org/GNOME/libxml2/commit/0e1a49c89076]
+CVE: CVE-2020-7595
+Signed-off-by: Anuj Mittal <[email protected]> 
+---
+ parser.c | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/parser.c b/parser.c
+index d1c31963..a34bb6cd 100644
+--- a/parser.c
++++ b/parser.c
+@@ -2646,7 +2646,8 @@ xmlStringLenDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
+     else
+         c = 0;
+     while ((c != 0) && (c != end) && /* non input consuming loop */
+-	   (c != end2) && (c != end3)) {
++           (c != end2) && (c != end3) &&
++           (ctxt->instate != XML_PARSER_EOF)) {
+ 
+ 	if (c == 0) break;
+         if ((c == '&') && (str[1] == '#')) {
+-- 
+2.24.1
+
meta/recipes-core/libxml/libxml2_2.9.9.bb +1 -0
index c44a90b1c2..5797dd6bee 100644
@@ -21,6 +21,7 @@ SRC_URI = "http://www.xmlsoft.org/sources/libxml2-${PV}.tar.gz;name=libtar \
            file://0001-Make-ptest-run-the-python-tests-if-python-is-enabled.patch \
            file://fix-execution-of-ptests.patch \
            file://Fix-CVE-2019-19956.patch \
+           file://CVE-2020-7595.patch \
            "
 
 SRC_URI[libtar.md5sum] = "c04a5a0a042eaa157e8e8c9eabe76bd6"
Signature · 1 line
2.17.1
#1 lore raw

PATCH 2/2 libxml2: Fix CVE-2019-20388

From: Lee Chee Yang <[email protected]>

see:
https://gitlab.gnome.org/GNOME/libxml2/merge_requests/68

---
 .../libxml/libxml2/CVE-2019-20388.patch       | 37 +++++++++++++++++++
 meta/recipes-core/libxml/libxml2_2.9.9.bb     |  1 +
 2 files changed, 38 insertions(+)
 create mode 100644 meta/recipes-core/libxml/libxml2/CVE-2019-20388.patch
meta/recipes-core/libxml/libxml2/CVE-2019-20388.patch new file +37 -0
new file mode 100644
index 0000000000..4ee2d4fe62
@@ -0,0 +1,37 @@
+From 7ffcd44d7e6c46704f8af0321d9314cd26e0e18a Mon Sep 17 00:00:00 2001
+From: Zhipeng Xie <[email protected]>
+Date: Tue, 20 Aug 2019 16:33:06 +0800
+Subject: [PATCH] Fix memory leak in xmlSchemaValidateStream
+
+When ctxt->schema is NULL, xmlSchemaSAXPlug->xmlSchemaPreRun
+alloc a new schema for ctxt->schema and set vctxt->xsiAssemble
+to 1. Then xmlSchemaVStart->xmlSchemaPreRun initialize
+vctxt->xsiAssemble to 0 again which cause the alloced schema
+can not be freed anymore.
+
+Found with libFuzzer.
+
+Upstream-Status: Accepted [https://gitlab.gnome.org/GNOME/libxml2/commit/7ffcd44d7e6c46704f8af0321d9314cd26e0e18a]
+CVE: CVE-2019-20388
+
+Signed-off-by: Zhipeng Xie <[email protected]>
+Signed-off-by: Lee Chee Yang <[email protected]>
+---
+ xmlschemas.c | 1 -
+ 1 file changed, 1 deletion(-)
+
+diff --git a/xmlschemas.c b/xmlschemas.c
+index 301c8449..39d92182 100644
+--- a/xmlschemas.c
++++ b/xmlschemas.c
+@@ -28090,7 +28090,6 @@ xmlSchemaPreRun(xmlSchemaValidCtxtPtr vctxt) {
+     vctxt->nberrors = 0;
+     vctxt->depth = -1;
+     vctxt->skipDepth = -1;
+-    vctxt->xsiAssemble = 0;
+     vctxt->hasKeyrefs = 0;
+ #ifdef ENABLE_IDC_NODE_TABLES_TEST
+     vctxt->createIDCNodeTables = 1;
+-- 
+2.24.1
+
meta/recipes-core/libxml/libxml2_2.9.9.bb +1 -0
index 5797dd6bee..1d898ab020 100644
@@ -22,6 +22,7 @@ SRC_URI = "http://www.xmlsoft.org/sources/libxml2-${PV}.tar.gz;name=libtar \
            file://fix-execution-of-ptests.patch \
            file://Fix-CVE-2019-19956.patch \
            file://CVE-2020-7595.patch \
+           file://CVE-2019-20388.patch \
            "
 
 SRC_URI[libtar.md5sum] = "c04a5a0a042eaa157e8e8c9eabe76bd6"
Signature · 1 line
2.17.1
#2 lore raw
In reply to Adrian Bunk
Quoted 5 lines · Adrian Bunk

From: Anuj Mittal <[email protected]>

Is this a backport from Master? please include masters commit id.

Since Master and Zeus are at different versions, was there anything you had to do to make this apply to Zeus?

I have adopted the kernel stable note of [Fixup for XXX context] so the folks know things had to be tweaked from the original commit.

- Armin

Quoted 60 lines
---
 .../libxml/libxml2/CVE-2020-7595.patch        | 36 +++++++++++++++++++
 meta/recipes-core/libxml/libxml2_2.9.9.bb     |  1 +
 2 files changed, 37 insertions(+)
 create mode 100644 meta/recipes-core/libxml/libxml2/CVE-2020-7595.patch
meta/recipes-core/libxml/libxml2/CVE-2020-7595.patch new file +36 -0
new file mode 100644
index 0000000000..facfefd362
@@ -0,0 +1,36 @@
+From 0e1a49c8907645d2e155f0d89d4d9895ac5112b5 Mon Sep 17 00:00:00 2001
+From: Zhipeng Xie <[email protected]>
+Date: Thu, 12 Dec 2019 17:30:55 +0800
+Subject: [PATCH] Fix infinite loop in xmlStringLenDecodeEntities
+
+When ctxt->instate == XML_PARSER_EOF,xmlParseStringEntityRef
+return NULL which cause a infinite loop in xmlStringLenDecodeEntities
+
+Found with libFuzzer.
+
+Signed-off-by: Zhipeng Xie <[email protected]>
+
+Upstream-Status: Backport [https://gitlab.gnome.org/GNOME/libxml2/commit/0e1a49c89076]
+CVE: CVE-2020-7595
+Signed-off-by: Anuj Mittal <[email protected]> 
+---
+ parser.c | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/parser.c b/parser.c
+index d1c31963..a34bb6cd 100644
+--- a/parser.c
++++ b/parser.c
+@@ -2646,7 +2646,8 @@ xmlStringLenDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
+     else
+         c = 0;
+     while ((c != 0) && (c != end) && /* non input consuming loop */
+-	   (c != end2) && (c != end3)) {
++           (c != end2) && (c != end3) &&
++           (ctxt->instate != XML_PARSER_EOF)) {
+ 
+ 	if (c == 0) break;
+         if ((c == '&') && (str[1] == '#')) {
+-- 
+2.24.1
+
meta/recipes-core/libxml/libxml2_2.9.9.bb +1 -0
index c44a90b1c2..5797dd6bee 100644
@@ -21,6 +21,7 @@ SRC_URI = "http://www.xmlsoft.org/sources/libxml2-${PV}.tar.gz;name=libtar \
            file://0001-Make-ptest-run-the-python-tests-if-python-is-enabled.patch \
            file://fix-execution-of-ptests.patch \
            file://Fix-CVE-2019-19956.patch \
+           file://CVE-2020-7595.patch \
            "
 
 SRC_URI[libtar.md5sum] = "c04a5a0a042eaa157e8e8c9eabe76bd6"
#3 lore raw
Quoted 8 lines · On Sat, Feb 22, 2020 at 10:50:13AM -0800, akuster808 wrote:
Quoted 5 lines · Adrian Bunk

From: Anuj Mittal <[email protected]>

Is this a backport from Master?

Yes.

Quoted 1 line

please include masters commit id.

openembedded-core commit id or poky commit id? They are always different, and I am working on top of poky.

Quoted 2 lines

Since Master and Zeus are at different versions, was there anything you had to do to make this apply to Zeus?

There is always the trivial change to the bb file.

Quoted 2 lines

I have adopted the kernel stable note of [Fixup for XXX context] so the folks know things had to be tweaked from the original commit.

I am already doing the same for non-trivial changes, e.g.
http://lists.openembedded.org/pipermail/openembedded-core/2020-February/293312.html

Quoted 1 line

- Armin

cu
Adrian

#4 lore raw
Quoted 4 lines · Adrian Bunk
Quoted 3 lines · On Sat, Feb 22, 2020 at 10:50:13AM -0800, akuster808 wrote:
Quoted 2 lines · Adrian Bunk

From: Anuj Mittal <[email protected]>

Quoted 8 lines
Quoted 4 lines
Quoted 3 lines

Is this a backport from Master?

Yes.

Quoted 1 line

please include masters commit id.

openembedded-core commit id or poky commit id?

Openembedded-core  ID

- armin

Quoted 14 lines

They are always different, and I am working on top of poky.

Quoted 2 lines

Since Master and Zeus are at different versions, was there anything you had to do to make this apply to Zeus?

There is always the trivial change to the bb file.

Quoted 2 lines

I have adopted the kernel stable note of [Fixup for XXX context] so the folks know things had to be tweaked from the original commit.

I am already doing the same for non-trivial changes, e.g.
http://lists.openembedded.org/pipermail/openembedded-core/2020-February/293312.html

Quoted 1 line

- Armin

cu
Adrian