From: albert.linde@gmail.com
To: akpm@linux-foundation.org, bp@alien8.de, mingo@redhat.com,
corbet@lwn.net, tglx@linutronix.de, arnd@arndb.de
Cc: akinobu.mita@gmail.com, hpa@zytor.com, viro@zeniv.linux.org.uk,
glider@google.com, andreyknvl@google.com, dvyukov@google.com,
elver@google.com, linux-doc@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
x86@kernel.org, albert.linde@gmail.com,
Albert van der Linde <alinde@google.com>
Subject: [PATCH v2 1/3] lib, include/linux: add usercopy failure capability
Date: Fri, 28 Aug 2020 14:13:42 +0000 [thread overview]
Message-ID: <20200828141344.2277088-2-alinde@google.com> (raw)
In-Reply-To: <20200828141344.2277088-1-alinde@google.com>
From: Albert van der Linde <alinde@google.com>
Add a failure injection capability to improve testing of fault-tolerance
in usages of user memory access functions.
Add CONFIG_FAULT_INJECTION_USERCOPY to enable faults in usercopy
functions. The should_fail_usercopy function is to be called by these
functions (copy_from_user, get_user, ...) in order to fail or not.
Signed-off-by: Albert van der Linde <alinde@google.com>
---
v2:
- adressed comments from Dmitry Vyukov
- removed failsize
- changed should_fail function to return bool
---
.../admin-guide/kernel-parameters.txt | 1 +
.../fault-injection/fault-injection.rst | 7 +++-
include/linux/fault-inject-usercopy.h | 22 +++++++++++
lib/Kconfig.debug | 7 ++++
lib/Makefile | 1 +
lib/fault-inject-usercopy.c | 39 +++++++++++++++++++
6 files changed, 76 insertions(+), 1 deletion(-)
create mode 100644 include/linux/fault-inject-usercopy.h
create mode 100644 lib/fault-inject-usercopy.c
diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index a1068742a6df..790e54988d4f 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -1332,6 +1332,7 @@
current integrity status.
failslab=
+ fail_usercopy=
fail_page_alloc=
fail_make_request=[KNL]
General fault injection mechanism.
diff --git a/Documentation/fault-injection/fault-injection.rst b/Documentation/fault-injection/fault-injection.rst
index f850ad018b70..31ecfe44e5b4 100644
--- a/Documentation/fault-injection/fault-injection.rst
+++ b/Documentation/fault-injection/fault-injection.rst
@@ -16,6 +16,10 @@ Available fault injection capabilities
injects page allocation failures. (alloc_pages(), get_free_pages(), ...)
+- fail_usercopy
+
+ injects failures in user memory access functions. (copy_from_user(), get_user(), ...)
+
- fail_futex
injects futex deadlock and uaddr fault errors.
@@ -177,6 +181,7 @@ use the boot option::
failslab=
fail_page_alloc=
+ fail_usercopy=
fail_make_request=
fail_futex=
mmc_core.fail_request=<interval>,<probability>,<space>,<times>
@@ -222,7 +227,7 @@ How to add new fault injection capability
- debugfs entries
- failslab, fail_page_alloc, and fail_make_request use this way.
+ failslab, fail_page_alloc, fail_usercopy, and fail_make_request use this way.
Helper functions:
fault_create_debugfs_attr(name, parent, attr);
diff --git a/include/linux/fault-inject-usercopy.h b/include/linux/fault-inject-usercopy.h
new file mode 100644
index 000000000000..56c3a693fdd9
--- /dev/null
+++ b/include/linux/fault-inject-usercopy.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LINUX_FAULT_INJECT_USERCOPY_H__
+#define __LINUX_FAULT_INJECT_USERCOPY_H__
+
+/*
+ * This header provides a wrapper for injecting failures to user space memory
+ * access functions.
+ */
+
+#include <linux/types.h>
+
+#ifdef CONFIG_FAULT_INJECTION_USERCOPY
+
+bool should_fail_usercopy(void);
+
+#else
+
+static inline bool should_fail_usercopy(void) { return false; }
+
+#endif /* CONFIG_FAULT_INJECTION_USERCOPY */
+
+#endif /* __LINUX_FAULT_INJECT_USERCOPY_H__ */
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index e068c3c7189a..2fc5049fba4e 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -1770,6 +1770,13 @@ config FAIL_PAGE_ALLOC
help
Provide fault-injection capability for alloc_pages().
+config FAULT_INJECTION_USERCOPY
+ bool "Fault injection capability for usercopy functions"
+ depends on FAULT_INJECTION
+ help
+ Provides fault-injection capability to inject failures
+ in usercopy functions (copy_from_user(), get_user(), ...).
+
config FAIL_MAKE_REQUEST
bool "Fault-injection capability for disk IO"
depends on FAULT_INJECTION && BLOCK
diff --git a/lib/Makefile b/lib/Makefile
index a4a4c6864f51..18daad2bc606 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -207,6 +207,7 @@ obj-$(CONFIG_AUDIT_COMPAT_GENERIC) += compat_audit.o
obj-$(CONFIG_IOMMU_HELPER) += iommu-helper.o
obj-$(CONFIG_FAULT_INJECTION) += fault-inject.o
+obj-$(CONFIG_FAULT_INJECTION_USERCOPY) += fault-inject-usercopy.o
obj-$(CONFIG_NOTIFIER_ERROR_INJECTION) += notifier-error-inject.o
obj-$(CONFIG_PM_NOTIFIER_ERROR_INJECT) += pm-notifier-error-inject.o
obj-$(CONFIG_NETDEV_NOTIFIER_ERROR_INJECT) += netdev-notifier-error-inject.o
diff --git a/lib/fault-inject-usercopy.c b/lib/fault-inject-usercopy.c
new file mode 100644
index 000000000000..77558b6c29ca
--- /dev/null
+++ b/lib/fault-inject-usercopy.c
@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/fault-inject.h>
+#include <linux/fault-inject-usercopy.h>
+
+static struct {
+ struct fault_attr attr;
+} fail_usercopy = {
+ .attr = FAULT_ATTR_INITIALIZER,
+};
+
+static int __init setup_fail_usercopy(char *str)
+{
+ return setup_fault_attr(&fail_usercopy.attr, str);
+}
+__setup("fail_usercopy=", setup_fail_usercopy);
+
+#ifdef CONFIG_FAULT_INJECTION_DEBUG_FS
+
+static int __init fail_usercopy_debugfs(void)
+{
+ struct dentry *dir;
+
+ dir = fault_create_debugfs_attr("fail_usercopy", NULL,
+ &fail_usercopy.attr);
+ if (IS_ERR(dir))
+ return PTR_ERR(dir);
+
+ return 0;
+}
+
+late_initcall(fail_usercopy_debugfs);
+
+#endif /* CONFIG_FAULT_INJECTION_DEBUG_FS */
+
+bool should_fail_usercopy(void)
+{
+ return should_fail(&fail_usercopy.attr, 1);
+}
+EXPORT_SYMBOL_GPL(should_fail_usercopy);
--
2.28.0.402.g5ffc5be6b7-goog
next prev parent reply other threads:[~2020-08-28 14:15 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-28 14:13 [PATCH v2 0/3] add fault injection to user memory access albert.linde
2020-08-28 14:13 ` albert.linde [this message]
2020-08-28 14:13 ` [PATCH v2 2/3] lib, uaccess: add failure injection to usercopy functions albert.linde
2020-08-31 16:21 ` peterz
2020-08-28 14:13 ` [PATCH v2 3/3] x86: add failure injection to get/put/clear_user albert.linde
2020-08-31 14:29 ` [PATCH v2 0/3] add fault injection to user memory access Akinobu Mita
2020-08-31 15:49 ` Alexander Potapenko
2020-08-31 16:27 ` Akinobu Mita
2020-08-31 18:31 ` Alexander Potapenko
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=20200828141344.2277088-2-alinde@google.com \
--to=albert.linde@gmail.com \
--cc=akinobu.mita@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=alinde@google.com \
--cc=andreyknvl@google.com \
--cc=arnd@arndb.de \
--cc=bp@alien8.de \
--cc=corbet@lwn.net \
--cc=dvyukov@google.com \
--cc=elver@google.com \
--cc=glider@google.com \
--cc=hpa@zytor.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=tglx@linutronix.de \
--cc=viro@zeniv.linux.org.uk \
--cc=x86@kernel.org \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).