Skip to contentSkip to conversation

ocfs2: Replace strlcpy with strscpy

PATCH In review Azeem Shaikh 3 participants 8 messages Aug 30, 2023 – Aug 31, 2023 OCFS2

lore
Share & tools
lore
#0 lore raw

cover ocfs2: Replace strlcpy with strscpy

Hi,

The main patch this series is targeting is [PATCH 2] which replaces strlcpy() call with strscpy(). However, while I was tinkering through the code I noticed that module_param_call is marked obsolete and module_param_cb is preferred instead. So I have included [PATCH 1] which does that.

A crucial thing in [PATCH 2] I would like to bring to reviewer's attention is that it changes behavior for the case where sizeof(@buffer) < DLMFS_CAPABILITIES. Currently, this is silently ignored but with the current change it returns -errno.

Thanks,
Azeem

Azeem Shaikh (2):
Replace module_param_call with module_param_cb
Replace strlcpy with strscpy

 fs/ocfs2/dlmfs/dlmfs.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

--
2.42.0.283.g2d96d420d3-goog

#1 lore raw

PATCH 1/2 ocfs2: Replace module_param_call with module_param_cb

module_param_call has been marked obsolete [1], so replacing its usage with module_param_cb instead.

[1] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L296

---
 fs/ocfs2/dlmfs/dlmfs.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/ocfs2/dlmfs/dlmfs.c +5 -2
index 81265123ce6c..33e529de93b2 100644
@@ -83,8 +83,11 @@ static int param_get_dlmfs_capabilities(char *buffer,
 	return strlcpy(buffer, DLMFS_CAPABILITIES,
 		       strlen(DLMFS_CAPABILITIES) + 1);
 }
-module_param_call(capabilities, param_set_dlmfs_capabilities,
-		  param_get_dlmfs_capabilities, NULL, 0444);
+static const struct kernel_param_ops dlmfs_capabilities_ops = {
+	.set = param_set_dlmfs_capabilities,
+	.get = param_get_dlmfs_capabilities,
+};
+module_param_cb(capabilities, &dlmfs_capabilities_ops, NULL, 0444);
 MODULE_PARM_DESC(capabilities, DLMFS_CAPABILITIES);
 
 

--
2.42.0.283.g2d96d420d3-goog

#2 lore raw

Re: [PATCH 1/2] ocfs2: Replace module_param_call with module_param_cb

Quoted 16 lines · Azeem Shaikh

module_param_call has been marked obsolete [1], so replacing its usage with module_param_cb instead.

[1] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L296

---
 fs/ocfs2/dlmfs/dlmfs.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/ocfs2/dlmfs/dlmfs.c +0 -0
index 81265123ce6c..33e529de93b2 100644
@@ -83,8 +83,11 @@ static int param_get_dlmfs_capabilities(char *buffer,
 	return strlcpy(buffer, DLMFS_CAPABILITIES,

fs/ocfs2/dlmfs/dlmfs.c static int param_get_dlmfs_capabilities(char *buffer, · line 91

Quoted 10 lines
fs/ocfs2/dlmfs/dlmfs.c +5 -2
 		       strlen(DLMFS_CAPABILITIES) + 1);
 }
-module_param_call(capabilities, param_set_dlmfs_capabilities,
-		  param_get_dlmfs_capabilities, NULL, 0444);
+static const struct kernel_param_ops dlmfs_capabilities_ops = {
+	.set = param_set_dlmfs_capabilities,
+	.get = param_get_dlmfs_capabilities,
+};
+module_param_cb(capabilities, &dlmfs_capabilities_ops, NULL, 0444);
 MODULE_PARM_DESC(capabilities, DLMFS_CAPABILITIES);

Oh, hm. Yeah, that's not good documentation. It was originally added in 9bbb9e5a33109b2832e2e63dcc7a132924ab374b, but that was doing some casting and other things during the conversion to an ops structure.

I cleaned up all the last of those back in 2017, and I should have dropped the "Obsolete" comment:
ece1996a21eeb344b49200e627c6660111009c10 b2f270e8747387335d80428c576118e7d87f69cc

The resulting patch you sent ends up literally open coding what it already does...

#define module_param_call(name, _set, _get, arg, perm)                  \
        static const struct kernel_param_ops __param_ops_##name =       \
                { .flags = 0, .set = _set, .get = _get };               \
        __module_param_call(MODULE_PARAM_PREFIX,                        \
                            name, &__param_ops_##name, arg, perm, -1, 0)

#define module_param_cb(name, ops, arg, perm) \
__module_param_call(MODULE_PARAM_PREFIX, name, ops, arg, perm, -1, 0)

Looking at usage, it's still common:

$ git grep '^module_param_call(' | wc -l
54
$ git grep '^module_param_cb(' | wc -l
93

And the users of module_param_cb() appear to be almost universally open-coding the result. Only a few initialize struct members that aren't .get and .set:

$ git grep -B6 ^module_param_cb

I'd say drop this patch and instead patch moduleparam.h to not say it's deprecated. :P

-Kees

Signature · 1 line
Kees Cook
#3 lore raw

PATCH 2/2 ocfs2: Replace strlcpy with strscpy

In reply to Azeem Shaikh

strlcpy() reads the entire source buffer first. This read may exceed the destination size limit. This is both inefficient and can lead to linear read overflows if a source string is not NUL-terminated [1]. In an effort to remove strlcpy() completely [2], replace strlcpy() here with strscpy().

Direct replacement is assumed to be safe here since
it's ok for kernel_param_ops.get() to return -errno [3]. This changes the behavior such that instead of silently ignoring the case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
[2] https://github.com/KSPP/linux/issues/89
[3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52

---
 fs/ocfs2/dlmfs/dlmfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ocfs2/dlmfs/dlmfs.c +1 -1
index 33e529de93b2..b001eccdd2f3 100644
@@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
 static int param_get_dlmfs_capabilities(char *buffer,
 					const struct kernel_param *kp)
 {
-	return strlcpy(buffer, DLMFS_CAPABILITIES,
+	return strscpy(buffer, DLMFS_CAPABILITIES,
 		       strlen(DLMFS_CAPABILITIES) + 1);
 }
 static const struct kernel_param_ops dlmfs_capabilities_ops = {

--
2.42.0.283.g2d96d420d3-goog

#4 lore raw
Quoted 11 lines · Azeem Shaikh

strlcpy() reads the entire source buffer first. This read may exceed the destination size limit. This is both inefficient and can lead to linear read overflows if a source string is not NUL-terminated [1]. In an effort to remove strlcpy() completely [2], replace strlcpy() here with strscpy().

Direct replacement is assumed to be safe here since
it's ok for kernel_param_ops.get() to return -errno [3]. This changes the behavior such that instead of silently ignoring the case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.

Not super familiar with the semantics of kernel_param_ops.get() but do note that strscpy can only return -E2BIG and not -ERRNO specifically. Is this still OK?

Quoted 26 lines

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
[2] https://github.com/KSPP/linux/issues/89
[3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52

---
 fs/ocfs2/dlmfs/dlmfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ocfs2/dlmfs/dlmfs.c +1 -2
index 33e529de93b2..b001eccdd2f3 100644
@@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
 static int param_get_dlmfs_capabilities(char *buffer,
 					const struct kernel_param *kp)
 {
-	return strlcpy(buffer, DLMFS_CAPABILITIES,
+	return strscpy(buffer, DLMFS_CAPABILITIES,
 		       strlen(DLMFS_CAPABILITIES) + 1);
 }
 static const struct kernel_param_ops dlmfs_capabilities_ops = {
--

2.42.0.283.g2d96d420d3-goog

#5 lore raw
In reply to Azeem Shaikh
Quoted 26 lines · Azeem Shaikh

strlcpy() reads the entire source buffer first. This read may exceed the destination size limit. This is both inefficient and can lead to linear read overflows if a source string is not NUL-terminated [1]. In an effort to remove strlcpy() completely [2], replace strlcpy() here with strscpy().

Direct replacement is assumed to be safe here since
it's ok for kernel_param_ops.get() to return -errno [3]. This changes the behavior such that instead of silently ignoring the case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
[2] https://github.com/KSPP/linux/issues/89
[3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52

---
 fs/ocfs2/dlmfs/dlmfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ocfs2/dlmfs/dlmfs.c +0 -0
index 33e529de93b2..b001eccdd2f3 100644

fs/ocfs2/dlmfs/dlmfs.c static int param_set_dlmfs_capabilities(const char *val, · line 85

Quoted 7 lines
fs/ocfs2/dlmfs/dlmfs.c +1 -1
@@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
 static int param_get_dlmfs_capabilities(char *buffer,
 					const struct kernel_param *kp)
 {
-	return strlcpy(buffer, DLMFS_CAPABILITIES,
+	return strscpy(buffer, DLMFS_CAPABILITIES,
 		       strlen(DLMFS_CAPABILITIES) + 1);
 }

This is another case of "accidentally correct".

param->get() is hooked here, in the sysfs "show" callback:

static ssize_t param_attr_show(struct module_attribute *mattr,
struct module_kobject *mk, char *buf)

{
        int count;
        struct param_attribute *attribute = to_param_attr(mattr);

        if (!attribute->param->ops->get)
                return -EPERM;

        kernel_param_lock(mk->mod);
        count = attribute->param->ops->get(buf, attribute->param);
        kernel_param_unlock(mk->mod);
        return count;
}

Meaning ultimately this will show up here, if I'm reading names right:
/sys/module/ocfs/parameters/dlmfs_capabilities

Anyway, the "count" being returned would be quite bad if DLMFS_CAPABILITIES were dynamic and larger than PAGE_SIZE (the size of the sysfs buffer).

For this case, I would say replace strlcpy with sysfs_emit:

	return sysfs_emit(buffer, DLMFS_CAPABILITIES);

(Also, ew, existing code doesn't include a trailing "\n". Oh well.)

-Kees

Signature · 1 line
Kees Cook
#6 lore raw
Quoted 59 lines · Kees Cook
Quoted 26 lines · Azeem Shaikh

strlcpy() reads the entire source buffer first. This read may exceed the destination size limit. This is both inefficient and can lead to linear read overflows if a source string is not NUL-terminated [1]. In an effort to remove strlcpy() completely [2], replace strlcpy() here with strscpy().

Direct replacement is assumed to be safe here since
it's ok for kernel_param_ops.get() to return -errno [3]. This changes the behavior such that instead of silently ignoring the case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
[2] https://github.com/KSPP/linux/issues/89
[3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52

---
 fs/ocfs2/dlmfs/dlmfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ocfs2/dlmfs/dlmfs.c +0 -0
index 33e529de93b2..b001eccdd2f3 100644

fs/ocfs2/dlmfs/dlmfs.c static int param_set_dlmfs_capabilities(const char *val, · line 85

Quoted 7 lines
fs/ocfs2/dlmfs/dlmfs.c +1 -1
@@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
 static int param_get_dlmfs_capabilities(char *buffer,
                                      const struct kernel_param *kp)
 {
-     return strlcpy(buffer, DLMFS_CAPABILITIES,
+     return strscpy(buffer, DLMFS_CAPABILITIES,
                     strlen(DLMFS_CAPABILITIES) + 1);
 }

This is another case of "accidentally correct".

param->get() is hooked here, in the sysfs "show" callback:

static ssize_t param_attr_show(struct module_attribute *mattr,
struct module_kobject *mk, char *buf)

{
        int count;
        struct param_attribute *attribute = to_param_attr(mattr);

        if (!attribute->param->ops->get)
                return -EPERM;

        kernel_param_lock(mk->mod);
        count = attribute->param->ops->get(buf, attribute->param);
        kernel_param_unlock(mk->mod);
        return count;
}

Meaning ultimately this will show up here, if I'm reading names right:
/sys/module/ocfs/parameters/dlmfs_capabilities

Quoted 8 lines

Anyway, the "count" being returned would be quite bad if DLMFS_CAPABILITIES were dynamic and larger than PAGE_SIZE (the size of the sysfs buffer).

For this case, I would say replace strlcpy with sysfs_emit:

        return sysfs_emit(buffer, DLMFS_CAPABILITIES);

Thanks, sending out a v2 for this. Out of curiosity - why sysfs_emit? Is it because DLMFS_CAPABILITIES is a hard-coded string?

#7 lore raw
Quoted 64 lines · Azeem Shaikh
Quoted 63 lines · Kees Cook
Quoted 26 lines · Azeem Shaikh

strlcpy() reads the entire source buffer first. This read may exceed the destination size limit. This is both inefficient and can lead to linear read overflows if a source string is not NUL-terminated [1]. In an effort to remove strlcpy() completely [2], replace strlcpy() here with strscpy().

Direct replacement is assumed to be safe here since
it's ok for kernel_param_ops.get() to return -errno [3]. This changes the behavior such that instead of silently ignoring the case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
[2] https://github.com/KSPP/linux/issues/89
[3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52

---
 fs/ocfs2/dlmfs/dlmfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ocfs2/dlmfs/dlmfs.c +0 -0
index 33e529de93b2..b001eccdd2f3 100644

fs/ocfs2/dlmfs/dlmfs.c static int param_set_dlmfs_capabilities(const char *val, · line 85

Quoted 7 lines
fs/ocfs2/dlmfs/dlmfs.c +1 -1
@@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
 static int param_get_dlmfs_capabilities(char *buffer,
                                      const struct kernel_param *kp)
 {
-     return strlcpy(buffer, DLMFS_CAPABILITIES,
+     return strscpy(buffer, DLMFS_CAPABILITIES,
                     strlen(DLMFS_CAPABILITIES) + 1);
 }

This is another case of "accidentally correct".

param->get() is hooked here, in the sysfs "show" callback:

static ssize_t param_attr_show(struct module_attribute *mattr,
struct module_kobject *mk, char *buf)

{
        int count;
        struct param_attribute *attribute = to_param_attr(mattr);

        if (!attribute->param->ops->get)
                return -EPERM;

        kernel_param_lock(mk->mod);
        count = attribute->param->ops->get(buf, attribute->param);
        kernel_param_unlock(mk->mod);
        return count;
}

Meaning ultimately this will show up here, if I'm reading names right:
/sys/module/ocfs/parameters/dlmfs_capabilities

Anyway, the "count" being returned would be quite bad if DLMFS_CAPABILITIES were dynamic and larger than PAGE_SIZE (the size of the sysfs buffer).

Quoted 7 lines
Quoted 4 lines

For this case, I would say replace strlcpy with sysfs_emit:

        return sysfs_emit(buffer, DLMFS_CAPABILITIES);

Thanks, sending out a v2 for this. Out of curiosity - why sysfs_emit? Is it because DLMFS_CAPABILITIES is a hard-coded string?

It's basically that sysfs_emit() Does The Right Thing for these
callbacks: it tracks the offset of the buffer and checks for buffer overflow. (All the sysfs callback work on a page-aligned page-sized buffer, and the API lacks any length info -- it's "understood" to be page-sized.) So, since this is doing a string copy into what is ultimately a sysfs buffer, it's best to use sysfs_emit().

Signature · 1 line
Kees Cook