Skip to contentSkip to conversation

libseccomp: fix seccomp_export_bpf_mem out-of-bounds read

PATCH In review Khem Raj 3 participants 5 messages Aug 23, 2025 – Sep 4, 2025

lore
Share & tools
lore

Khem Raj

author
#0 lore raw

PATCH 1/2 libseccomp: fix seccomp_export_bpf_mem out-of-bounds read

Fixes segfaults in ptests on musl
Failed ptests:
{'libseccomp': ['11-basic-basic_errors%%001-00001_11-basic-basic_errors_rc=139']}

---
 ...mp_export_bpf_mem-out-of-bounds-read.patch | 37 +++++++++++++++++++
 .../libseccomp/libseccomp_2.6.0.bb            |  1 +
 2 files changed, 38 insertions(+)
 create mode 100644 meta/recipes-support/libseccomp/files/0001-api-fix-seccomp_export_bpf_mem-out-of-bounds-read.patch
meta/recipes-support/libseccomp/files/0001-api-fix-seccomp_export_bpf_mem-out-of-bounds-read.patch new file +37 -0
new file mode 100644
index 00000000000..0d2910846c7
@@ -0,0 +1,37 @@
+From 83a56d74482ef6b6cc9311b14bf9ea23573f5d4e Mon Sep 17 00:00:00 2001
+From: Alyssa Ross <[email protected]>
+Date: Thu, 13 Feb 2025 12:05:17 +0100
+Subject: [PATCH] api: fix seccomp_export_bpf_mem out-of-bounds read
+
+*len is the length of the destination buffer, but program->blks is
+probably not anywhere near that long.  It's already been checked above
+that BPF_PGM_SIZE(program) is less than or equal to *len, so that's
+the correct value to use here to avoid either reading or writing too
+much.
+
+I noticed this because tests/11-basic-basic_errors started failing on
+musl after e797591 ("all: add seccomp_precompute() functionality").
+
+Upstream-Status: Backport [https://github.com/seccomp/libseccomp/commit/dd759e8c4f5685b526638fba9ec4fc24c37c9aec]
+Signed-off-by: Alyssa Ross <[email protected]>
+Acked-by: Tom Hromatka <[email protected]>
+Signed-off-by: Paul Moore <[email protected]>
+(imported from commit e8dbc6b555fb936bdfb8ab86f9a45fda96a8b7a2)
+Signed-off-by: Khem Raj <[email protected]>
+---
+ src/api.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/src/api.c b/src/api.c
+index adccef3..65a277a 100644
+--- a/src/api.c
++++ b/src/api.c
+@@ -786,7 +786,7 @@ API int seccomp_export_bpf_mem(const scmp_filter_ctx ctx, void *buf,
+ 		if (BPF_PGM_SIZE(program) > *len)
+ 			rc = _rc_filter(-ERANGE);
+ 		else
+-			memcpy(buf, program->blks, *len);
++			memcpy(buf, program->blks, BPF_PGM_SIZE(program));
+ 	}
+ 	*len = BPF_PGM_SIZE(program);
+ 
meta/recipes-support/libseccomp/libseccomp_2.6.0.bb +1 -0
index 97988afa27b..998f1b5548d 100644
@@ -10,6 +10,7 @@ DEPENDS += "gperf-native"
 SRCREV = "c7c0caed1d04292500ed4b9bb386566053eb9775"
 
 SRC_URI = "git://github.com/seccomp/libseccomp.git;branch=release-2.6;protocol=https \
+           file://0001-api-fix-seccomp_export_bpf_mem-out-of-bounds-read.patch \
            file://run-ptest \
            "
 
author
#1 lore raw

PATCH 2/2 lttng-tools: fix unix socket high FD count test

Fixes ptest failure on musl

'ERROR: Exited from signal Killed (9)\n' 'DURATION: 663\n'}

---
 .../lttng/lttng-tools/test.patch              | 22 +++++++++++++++++++
 .../lttng/lttng-tools_2.14.0.bb               |  1 +
 2 files changed, 23 insertions(+)
 create mode 100644 meta/recipes-kernel/lttng/lttng-tools/test.patch
meta/recipes-kernel/lttng/lttng-tools/test.patch new file +22 -0
new file mode 100644
index 00000000000..99d99369637
@@ -0,0 +1,22 @@
+tests: fix unix socket high FD count test off-by-one
+
+The HIGH_FD_COUNT macro in test_unix_socket.cpp was set equal to
+LTTCOMM_MAX_SEND_FDS, which could cause the test to exceed the
+maximum allowed number of file descriptors. Adjust it to
+LTTCOMM_MAX_SEND_FDS - 1 to avoid hitting the limit and ensure the
+test remains valid.
+
+Upstream-Status: Inappropriate [Musl-specific]
+Signed-off-by: Khem Raj <[email protected]>
+
+--- a/tests/unit/test_unix_socket.cpp
++++ b/tests/unit/test_unix_socket.cpp
+@@ -25,7 +25,7 @@
+ #include <tap/tap.h>
+ #include <unistd.h>
+ 
+-#define HIGH_FD_COUNT		LTTCOMM_MAX_SEND_FDS
++#define HIGH_FD_COUNT		LTTCOMM_MAX_SEND_FDS - 1
+ #define MESSAGE_COUNT		4
+ #define LARGE_PAYLOAD_SIZE	(4 * 1024)
+ #define LARGE_PAYLOAD_RECV_SIZE 100
meta/recipes-kernel/lttng/lttng-tools_2.14.0.bb +1 -0
index 49575ef57cd..6f3fba1d735 100644
@@ -54,6 +54,7 @@ SRC_URI = "https://lttng.org/files/lttng-tools/lttng-tools-${PV}.tar.bz2 \
            file://disable-tests2.patch \
            file://0001-liblttng-ctl-drop-index-allocator-symbols-from-versi.patch \
            file://libc++.patch \
+           file://test.patch \
            "
 
 SRC_URI[sha256sum] = "d8c39c26cec13b7bd82551cd52a22efc358b888e36ebcf9c1b60ef1c3a3c2fd3"
#2 lore raw

Re: [OE-core] [PATCH 2/2] lttng-tools: fix unix socket high FD count test

meta/recipes-kernel/lttng/lttng-tools/test.patch line 9

Quoted 7 lines · Khem Raj via lists.openembedded.org
meta/recipes-kernel/lttng/lttng-tools/test.patch new file +7 -0
+The HIGH_FD_COUNT macro in test_unix_socket.cpp was set equal to
+LTTCOMM_MAX_SEND_FDS, which could cause the test to exceed the
+maximum allowed number of file descriptors. Adjust it to
+LTTCOMM_MAX_SEND_FDS - 1 to avoid hitting the limit and ensure the
+test remains valid.
+
+Upstream-Status: Inappropriate [Musl-specific]

This needs an upstream ticket.

Alex

#3 lore raw

Re: [OE-core] [PATCH 2/2] lttng-tools: fix unix socket high FD count test

In reply to Khem Raj
Quoted 1 line · Khem Raj via lists.openembedded.org

+tests: fix unix socket high FD count test off-by-one

Is this an off-by-one in the test suite (in which case it should be upstreamed), in musl, or in glibc?

Someone has a bug and we should endeavour to fix it at source.

Ross

author
#4 lore raw

Re: [OE-core] [PATCH 2/2] lttng-tools: fix unix socket high FD count test

Quoted 6 lines · Ross Burton
Quoted 1 line · Khem Raj via lists.openembedded.org

+tests: fix unix socket high FD count test off-by-one

Is this an off-by-one in the test suite (in which case it should be upstreamed), in musl, or in glibc?

I saw it on musl and this machine was running a lot of parallel tests, but it's not happening all the time. It's fine to drop this patch

Quoted 3 lines

Someone has a bug and we should endeavour to fix it at source.

Ross