linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines
@ 2021-08-03 10:59 Thomas Huth
  2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
                   ` (4 more replies)
  0 siblings, 5 replies; 11+ messages in thread
From: Thomas Huth @ 2021-08-03 10:59 UTC (permalink / raw)
  To: linux-kernel, linux-fsdevel, linux-nfs
  Cc: J. Bruce Fields, Chuck Lever, Trond Myklebust, Anna Schumaker,
	Luis Chamberlain, Kees Cook, Iurii Zaikin, linux-s390, Jia He,
	Pan Xinhui

There is an endianess problem with /proc/sys/fs/nfs/nsm_use_hostnames
(which can e.g. be seen on an s390x host) :

 # modprobe lockd nsm_use_hostnames=1
 # cat /proc/sys/fs/nfs/nsm_use_hostnames
 16777216

The nsm_use_hostnames variable is declared as "bool" which is required
for the correct type for the module parameter. However, this does not
work correctly with the entry in the /proc filesystem since this
currently requires "int".

Jia He already provided patches for this problem a couple of years ago,
but apparently they felt through the cracks and never got merged. So
here's a rebased version to finally fix this issue.

Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1764075

Jia He (2):
  sysctl: introduce new proc handler proc_dobool
  lockd: change the proc_handler for nsm_use_hostnames

 fs/lockd/svc.c         |  2 +-
 include/linux/sysctl.h |  2 ++
 kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 45 insertions(+), 1 deletion(-)

-- 
2.27.0


^ permalink raw reply	[flat|nested] 11+ messages in thread

* [PATCH 1/2] sysctl: introduce new proc handler proc_dobool
  2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
@ 2021-08-03 10:59 ` Thomas Huth
  2021-08-18  9:38   ` Thomas Huth
  2021-09-16 22:03   ` Kees Cook
  2021-08-03 10:59 ` [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames Thomas Huth
                   ` (3 subsequent siblings)
  4 siblings, 2 replies; 11+ messages in thread
From: Thomas Huth @ 2021-08-03 10:59 UTC (permalink / raw)
  To: linux-kernel, linux-fsdevel, linux-nfs
  Cc: J. Bruce Fields, Chuck Lever, Trond Myklebust, Anna Schumaker,
	Luis Chamberlain, Kees Cook, Iurii Zaikin, linux-s390, Jia He,
	Pan Xinhui

From: Jia He <hejianet@gmail.com>

This is to let bool variable could be correctly displayed in
big/little endian sysctl procfs. sizeof(bool) is arch dependent,
proc_dobool should work in all arches.

Suggested-by: Pan Xinhui <xinhui@linux.vnet.ibm.com>
Signed-off-by: Jia He <hejianet@gmail.com>
[thuth: rebased the patch to the current kernel version]
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 include/linux/sysctl.h |  2 ++
 kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 44 insertions(+)

diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h
index d99ca99837de..1fa2b69c6fc3 100644
--- a/include/linux/sysctl.h
+++ b/include/linux/sysctl.h
@@ -48,6 +48,8 @@ typedef int proc_handler(struct ctl_table *ctl, int write, void *buffer,
 		size_t *lenp, loff_t *ppos);
 
 int proc_dostring(struct ctl_table *, int, void *, size_t *, loff_t *);
+int proc_dobool(struct ctl_table *table, int write, void *buffer,
+		size_t *lenp, loff_t *ppos);
 int proc_dointvec(struct ctl_table *, int, void *, size_t *, loff_t *);
 int proc_douintvec(struct ctl_table *, int, void *, size_t *, loff_t *);
 int proc_dointvec_minmax(struct ctl_table *, int, void *, size_t *, loff_t *);
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 272f4a272f8c..25e49b4d8049 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -536,6 +536,21 @@ static void proc_put_char(void **buf, size_t *size, char c)
 	}
 }
 
+static int do_proc_dobool_conv(bool *negp, unsigned long *lvalp,
+				int *valp,
+				int write, void *data)
+{
+	if (write) {
+		*(bool *)valp = *lvalp;
+	} else {
+		int val = *(bool *)valp;
+
+		*lvalp = (unsigned long)val;
+		*negp = false;
+	}
+	return 0;
+}
+
 static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
 				 int *valp,
 				 int write, void *data)
@@ -798,6 +813,26 @@ static int do_proc_douintvec(struct ctl_table *table, int write,
 				   buffer, lenp, ppos, conv, data);
 }
 
+/**
+ * proc_dobool - read/write a bool
+ * @table: the sysctl table
+ * @write: %TRUE if this is a write to the sysctl file
+ * @buffer: the user buffer
+ * @lenp: the size of the user buffer
+ * @ppos: file position
+ *
+ * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
+ * values from/to the user buffer, treated as an ASCII string.
+ *
+ * Returns 0 on success.
+ */
+int proc_dobool(struct ctl_table *table, int write, void *buffer,
+		size_t *lenp, loff_t *ppos)
+{
+	return do_proc_dointvec(table, write, buffer, lenp, ppos,
+				do_proc_dobool_conv, NULL);
+}
+
 /**
  * proc_dointvec - read a vector of integers
  * @table: the sysctl table
@@ -1630,6 +1665,12 @@ int proc_dostring(struct ctl_table *table, int write,
 	return -ENOSYS;
 }
 
+int proc_dobool(struct ctl_table *table, int write,
+		void *buffer, size_t *lenp, loff_t *ppos)
+{
+	return -ENOSYS;
+}
+
 int proc_dointvec(struct ctl_table *table, int write,
 		  void *buffer, size_t *lenp, loff_t *ppos)
 {
@@ -3425,6 +3466,7 @@ int __init sysctl_init(void)
  * No sense putting this after each symbol definition, twice,
  * exception granted :-)
  */
+EXPORT_SYMBOL(proc_dobool);
 EXPORT_SYMBOL(proc_dointvec);
 EXPORT_SYMBOL(proc_douintvec);
 EXPORT_SYMBOL(proc_dointvec_jiffies);
-- 
2.27.0


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames
  2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
  2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
@ 2021-08-03 10:59 ` Thomas Huth
  2021-09-16 22:05   ` Kees Cook
  2021-08-03 14:08 ` [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines J. Bruce Fields
                   ` (2 subsequent siblings)
  4 siblings, 1 reply; 11+ messages in thread
From: Thomas Huth @ 2021-08-03 10:59 UTC (permalink / raw)
  To: linux-kernel, linux-fsdevel, linux-nfs
  Cc: J. Bruce Fields, Chuck Lever, Trond Myklebust, Anna Schumaker,
	Luis Chamberlain, Kees Cook, Iurii Zaikin, linux-s390, Jia He,
	Pan Xinhui

From: Jia He <hejianet@gmail.com>

nsm_use_hostnames is a module parameter and it will be exported to sysctl
procfs. This is to let user sometimes change it from userspace. But the
minimal unit for sysctl procfs read/write it sizeof(int).
In big endian system, the converting from/to  bool to/from int will cause
error for proc items.

This patch use a new proc_handler proc_dobool to fix it.

Signed-off-by: Jia He <hejianet@gmail.com>
Reviewed-by: Pan Xinhui <xinhui.pan@linux.vnet.ibm.com>
[thuth: Fix typo in commit message]
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 fs/lockd/svc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 2de048f80eb8..0ab9756ed235 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -584,7 +584,7 @@ static struct ctl_table nlm_sysctls[] = {
 		.data		= &nsm_use_hostnames,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
-		.proc_handler	= proc_dointvec,
+		.proc_handler	= proc_dobool,
 	},
 	{
 		.procname	= "nsm_local_state",
-- 
2.27.0


^ permalink raw reply related	[flat|nested] 11+ messages in thread

* Re: [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines
  2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
  2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
  2021-08-03 10:59 ` [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames Thomas Huth
@ 2021-08-03 14:08 ` J. Bruce Fields
  2021-08-03 14:24   ` Chuck Lever III
  2021-08-04  2:12 ` Jia He
  2021-08-09 17:50 ` Chuck Lever III
  4 siblings, 1 reply; 11+ messages in thread
From: J. Bruce Fields @ 2021-08-03 14:08 UTC (permalink / raw)
  To: Thomas Huth
  Cc: linux-kernel, linux-fsdevel, linux-nfs, Chuck Lever,
	Trond Myklebust, Anna Schumaker, Luis Chamberlain, Kees Cook,
	Iurii Zaikin, linux-s390, Jia He, Pan Xinhui

Looks good to me.  Could Chuck take it with nfsd stuff if somebody could
ACK the sysctl part?

--b.

On Tue, Aug 03, 2021 at 12:59:35PM +0200, Thomas Huth wrote:
> There is an endianess problem with /proc/sys/fs/nfs/nsm_use_hostnames
> (which can e.g. be seen on an s390x host) :
> 
>  # modprobe lockd nsm_use_hostnames=1
>  # cat /proc/sys/fs/nfs/nsm_use_hostnames
>  16777216
> 
> The nsm_use_hostnames variable is declared as "bool" which is required
> for the correct type for the module parameter. However, this does not
> work correctly with the entry in the /proc filesystem since this
> currently requires "int".
> 
> Jia He already provided patches for this problem a couple of years ago,
> but apparently they felt through the cracks and never got merged. So
> here's a rebased version to finally fix this issue.
> 
> Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1764075
> 
> Jia He (2):
>   sysctl: introduce new proc handler proc_dobool
>   lockd: change the proc_handler for nsm_use_hostnames
> 
>  fs/lockd/svc.c         |  2 +-
>  include/linux/sysctl.h |  2 ++
>  kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 45 insertions(+), 1 deletion(-)
> 
> -- 
> 2.27.0

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines
  2021-08-03 14:08 ` [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines J. Bruce Fields
@ 2021-08-03 14:24   ` Chuck Lever III
  0 siblings, 0 replies; 11+ messages in thread
From: Chuck Lever III @ 2021-08-03 14:24 UTC (permalink / raw)
  To: Bruce Fields
  Cc: Thomas Huth, LKML, linux-fsdevel, Linux NFS Mailing List,
	Trond Myklebust, Anna Schumaker, Luis Chamberlain, Kees Cook,
	Iurii Zaikin, linux-s390, Jia He, Pan Xinhui

Thanks for your review, Bruce. I'll watch for the additional Ack.

> On Aug 3, 2021, at 10:08 AM, J. Bruce Fields <bfields@fieldses.org> wrote:
> 
> Looks good to me.  Could Chuck take it with nfsd stuff if somebody could
> ACK the sysctl part?
> 
> --b.
> 
> On Tue, Aug 03, 2021 at 12:59:35PM +0200, Thomas Huth wrote:
>> There is an endianess problem with /proc/sys/fs/nfs/nsm_use_hostnames
>> (which can e.g. be seen on an s390x host) :
>> 
>> # modprobe lockd nsm_use_hostnames=1
>> # cat /proc/sys/fs/nfs/nsm_use_hostnames
>> 16777216
>> 
>> The nsm_use_hostnames variable is declared as "bool" which is required
>> for the correct type for the module parameter. However, this does not
>> work correctly with the entry in the /proc filesystem since this
>> currently requires "int".
>> 
>> Jia He already provided patches for this problem a couple of years ago,
>> but apparently they felt through the cracks and never got merged. So
>> here's a rebased version to finally fix this issue.
>> 
>> Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1764075
>> 
>> Jia He (2):
>>  sysctl: introduce new proc handler proc_dobool
>>  lockd: change the proc_handler for nsm_use_hostnames
>> 
>> fs/lockd/svc.c         |  2 +-
>> include/linux/sysctl.h |  2 ++
>> kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
>> 3 files changed, 45 insertions(+), 1 deletion(-)
>> 
>> -- 
>> 2.27.0

--
Chuck Lever




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines
  2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
                   ` (2 preceding siblings ...)
  2021-08-03 14:08 ` [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines J. Bruce Fields
@ 2021-08-04  2:12 ` Jia He
  2021-08-09 17:50 ` Chuck Lever III
  4 siblings, 0 replies; 11+ messages in thread
From: Jia He @ 2021-08-04  2:12 UTC (permalink / raw)
  To: Thomas Huth, linux-kernel, linux-fsdevel, linux-nfs
  Cc: J. Bruce Fields, Chuck Lever, Trond Myklebust, Anna Schumaker,
	Luis Chamberlain, Kees Cook, Iurii Zaikin, linux-s390,
	Pan Xinhui


On 2021/8/3 18:59, Thomas Huth wrote:
> There is an endianess problem with /proc/sys/fs/nfs/nsm_use_hostnames
> (which can e.g. be seen on an s390x host) :
>
>   # modprobe lockd nsm_use_hostnames=1
>   # cat /proc/sys/fs/nfs/nsm_use_hostnames
>   16777216
>
> The nsm_use_hostnames variable is declared as "bool" which is required
> for the correct type for the module parameter. However, this does not
> work correctly with the entry in the /proc filesystem since this
> currently requires "int".
>
> Jia He already provided patches for this problem a couple of years ago,
> but apparently they felt through the cracks and never got merged. So
> here's a rebased version to finally fix this issue.
>
> Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1764075
>
> Jia He (2):
>    sysctl: introduce new proc handler proc_dobool
>    lockd: change the proc_handler for nsm_use_hostnames

Thanks for picking them up ;-)

---
Cheers,
Justin (Jia He)


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines
  2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
                   ` (3 preceding siblings ...)
  2021-08-04  2:12 ` Jia He
@ 2021-08-09 17:50 ` Chuck Lever III
  4 siblings, 0 replies; 11+ messages in thread
From: Chuck Lever III @ 2021-08-09 17:50 UTC (permalink / raw)
  To: Thomas Huth
  Cc: LKML, linux-fsdevel, Linux NFS Mailing List, Bruce Fields,
	Trond Myklebust, Anna Schumaker, Luis Chamberlain, Kees Cook,
	Iurii Zaikin, linux-s390, Jia He, Pan Xinhui



> On Aug 3, 2021, at 6:59 AM, Thomas Huth <thuth@redhat.com> wrote:
> 
> There is an endianess problem with /proc/sys/fs/nfs/nsm_use_hostnames
> (which can e.g. be seen on an s390x host) :
> 
> # modprobe lockd nsm_use_hostnames=1
> # cat /proc/sys/fs/nfs/nsm_use_hostnames
> 16777216
> 
> The nsm_use_hostnames variable is declared as "bool" which is required
> for the correct type for the module parameter. However, this does not
> work correctly with the entry in the /proc filesystem since this
> currently requires "int".
> 
> Jia He already provided patches for this problem a couple of years ago,
> but apparently they felt through the cracks and never got merged. So
> here's a rebased version to finally fix this issue.
> 
> Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1764075
> 
> Jia He (2):
>  sysctl: introduce new proc handler proc_dobool
>  lockd: change the proc_handler for nsm_use_hostnames
> 
> fs/lockd/svc.c         |  2 +-
> include/linux/sysctl.h |  2 ++
> kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 45 insertions(+), 1 deletion(-)
> 
> -- 
> 2.27.0

To get these patches in front of our zero-day apparatus,
I've applied them to the for-next topic branch here:

https://git.kernel.org/pub/scm/linux/kernel/git/cel/linux.git/log/?h=for-next

However I haven't seen an Ack for the kernel/sysctl.c change yet.


--
Chuck Lever




^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 1/2] sysctl: introduce new proc handler proc_dobool
  2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
@ 2021-08-18  9:38   ` Thomas Huth
  2021-09-16  5:16     ` Thomas Huth
  2021-09-16 22:03   ` Kees Cook
  1 sibling, 1 reply; 11+ messages in thread
From: Thomas Huth @ 2021-08-18  9:38 UTC (permalink / raw)
  To: linux-kernel, linux-fsdevel, Luis Chamberlain, Kees Cook, Iurii Zaikin
  Cc: J. Bruce Fields, Chuck Lever, linux-s390, Jia He

On 03/08/2021 12.59, Thomas Huth wrote:
> From: Jia He <hejianet@gmail.com>
> 
> This is to let bool variable could be correctly displayed in
> big/little endian sysctl procfs. sizeof(bool) is arch dependent,
> proc_dobool should work in all arches.
> 
> Suggested-by: Pan Xinhui <xinhui@linux.vnet.ibm.com>
> Signed-off-by: Jia He <hejianet@gmail.com>
> [thuth: rebased the patch to the current kernel version]
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>   include/linux/sysctl.h |  2 ++
>   kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
>   2 files changed, 44 insertions(+)
> 
> diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h
> index d99ca99837de..1fa2b69c6fc3 100644
> --- a/include/linux/sysctl.h
> +++ b/include/linux/sysctl.h
> @@ -48,6 +48,8 @@ typedef int proc_handler(struct ctl_table *ctl, int write, void *buffer,
>   		size_t *lenp, loff_t *ppos);
>   
>   int proc_dostring(struct ctl_table *, int, void *, size_t *, loff_t *);
> +int proc_dobool(struct ctl_table *table, int write, void *buffer,
> +		size_t *lenp, loff_t *ppos);
>   int proc_dointvec(struct ctl_table *, int, void *, size_t *, loff_t *);
>   int proc_douintvec(struct ctl_table *, int, void *, size_t *, loff_t *);
>   int proc_dointvec_minmax(struct ctl_table *, int, void *, size_t *, loff_t *);
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index 272f4a272f8c..25e49b4d8049 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -536,6 +536,21 @@ static void proc_put_char(void **buf, size_t *size, char c)
>   	}
>   }
>   
> +static int do_proc_dobool_conv(bool *negp, unsigned long *lvalp,
> +				int *valp,
> +				int write, void *data)
> +{
> +	if (write) {
> +		*(bool *)valp = *lvalp;
> +	} else {
> +		int val = *(bool *)valp;
> +
> +		*lvalp = (unsigned long)val;
> +		*negp = false;
> +	}
> +	return 0;
> +}
> +
>   static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
>   				 int *valp,
>   				 int write, void *data)
> @@ -798,6 +813,26 @@ static int do_proc_douintvec(struct ctl_table *table, int write,
>   				   buffer, lenp, ppos, conv, data);
>   }
>   
> +/**
> + * proc_dobool - read/write a bool
> + * @table: the sysctl table
> + * @write: %TRUE if this is a write to the sysctl file
> + * @buffer: the user buffer
> + * @lenp: the size of the user buffer
> + * @ppos: file position
> + *
> + * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
> + * values from/to the user buffer, treated as an ASCII string.
> + *
> + * Returns 0 on success.
> + */
> +int proc_dobool(struct ctl_table *table, int write, void *buffer,
> +		size_t *lenp, loff_t *ppos)
> +{
> +	return do_proc_dointvec(table, write, buffer, lenp, ppos,
> +				do_proc_dobool_conv, NULL);
> +}
> +
>   /**
>    * proc_dointvec - read a vector of integers
>    * @table: the sysctl table
> @@ -1630,6 +1665,12 @@ int proc_dostring(struct ctl_table *table, int write,
>   	return -ENOSYS;
>   }
>   
> +int proc_dobool(struct ctl_table *table, int write,
> +		void *buffer, size_t *lenp, loff_t *ppos)
> +{
> +	return -ENOSYS;
> +}
> +
>   int proc_dointvec(struct ctl_table *table, int write,
>   		  void *buffer, size_t *lenp, loff_t *ppos)
>   {
> @@ -3425,6 +3466,7 @@ int __init sysctl_init(void)
>    * No sense putting this after each symbol definition, twice,
>    * exception granted :-)
>    */
> +EXPORT_SYMBOL(proc_dobool);
>   EXPORT_SYMBOL(proc_dointvec);
>   EXPORT_SYMBOL(proc_douintvec);
>   EXPORT_SYMBOL(proc_dointvec_jiffies);
> 

Friendly ping!

Luis, Kees, Iurii, could you please have a look and provide an Ack if this 
looks ok to you?

  Thanks,
   Thomas


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 1/2] sysctl: introduce new proc handler proc_dobool
  2021-08-18  9:38   ` Thomas Huth
@ 2021-09-16  5:16     ` Thomas Huth
  0 siblings, 0 replies; 11+ messages in thread
From: Thomas Huth @ 2021-09-16  5:16 UTC (permalink / raw)
  To: linux-kernel, linux-fsdevel, Luis Chamberlain, Kees Cook, Iurii Zaikin
  Cc: J. Bruce Fields, Chuck Lever, linux-s390, Jia He

On 18/08/2021 11.38, Thomas Huth wrote:
> On 03/08/2021 12.59, Thomas Huth wrote:
>> From: Jia He <hejianet@gmail.com>
>>
>> This is to let bool variable could be correctly displayed in
>> big/little endian sysctl procfs. sizeof(bool) is arch dependent,
>> proc_dobool should work in all arches.
>>
>> Suggested-by: Pan Xinhui <xinhui@linux.vnet.ibm.com>
>> Signed-off-by: Jia He <hejianet@gmail.com>
>> [thuth: rebased the patch to the current kernel version]
>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>> ---
>>   include/linux/sysctl.h |  2 ++
>>   kernel/sysctl.c        | 42 ++++++++++++++++++++++++++++++++++++++++++
>>   2 files changed, 44 insertions(+)
>>
>> diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h
>> index d99ca99837de..1fa2b69c6fc3 100644
>> --- a/include/linux/sysctl.h
>> +++ b/include/linux/sysctl.h
>> @@ -48,6 +48,8 @@ typedef int proc_handler(struct ctl_table *ctl, int 
>> write, void *buffer,
>>           size_t *lenp, loff_t *ppos);
>>   int proc_dostring(struct ctl_table *, int, void *, size_t *, loff_t *);
>> +int proc_dobool(struct ctl_table *table, int write, void *buffer,
>> +        size_t *lenp, loff_t *ppos);
>>   int proc_dointvec(struct ctl_table *, int, void *, size_t *, loff_t *);
>>   int proc_douintvec(struct ctl_table *, int, void *, size_t *, loff_t *);
>>   int proc_dointvec_minmax(struct ctl_table *, int, void *, size_t *, 
>> loff_t *);
>> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
>> index 272f4a272f8c..25e49b4d8049 100644
>> --- a/kernel/sysctl.c
>> +++ b/kernel/sysctl.c
>> @@ -536,6 +536,21 @@ static void proc_put_char(void **buf, size_t *size, 
>> char c)
>>       }
>>   }
>> +static int do_proc_dobool_conv(bool *negp, unsigned long *lvalp,
>> +                int *valp,
>> +                int write, void *data)
>> +{
>> +    if (write) {
>> +        *(bool *)valp = *lvalp;
>> +    } else {
>> +        int val = *(bool *)valp;
>> +
>> +        *lvalp = (unsigned long)val;
>> +        *negp = false;
>> +    }
>> +    return 0;
>> +}
>> +
>>   static int do_proc_dointvec_conv(bool *negp, unsigned long *lvalp,
>>                    int *valp,
>>                    int write, void *data)
>> @@ -798,6 +813,26 @@ static int do_proc_douintvec(struct ctl_table *table, 
>> int write,
>>                      buffer, lenp, ppos, conv, data);
>>   }
>> +/**
>> + * proc_dobool - read/write a bool
>> + * @table: the sysctl table
>> + * @write: %TRUE if this is a write to the sysctl file
>> + * @buffer: the user buffer
>> + * @lenp: the size of the user buffer
>> + * @ppos: file position
>> + *
>> + * Reads/writes up to table->maxlen/sizeof(unsigned int) integer
>> + * values from/to the user buffer, treated as an ASCII string.
>> + *
>> + * Returns 0 on success.
>> + */
>> +int proc_dobool(struct ctl_table *table, int write, void *buffer,
>> +        size_t *lenp, loff_t *ppos)
>> +{
>> +    return do_proc_dointvec(table, write, buffer, lenp, ppos,
>> +                do_proc_dobool_conv, NULL);
>> +}
>> +
>>   /**
>>    * proc_dointvec - read a vector of integers
>>    * @table: the sysctl table
>> @@ -1630,6 +1665,12 @@ int proc_dostring(struct ctl_table *table, int write,
>>       return -ENOSYS;
>>   }
>> +int proc_dobool(struct ctl_table *table, int write,
>> +        void *buffer, size_t *lenp, loff_t *ppos)
>> +{
>> +    return -ENOSYS;
>> +}
>> +
>>   int proc_dointvec(struct ctl_table *table, int write,
>>             void *buffer, size_t *lenp, loff_t *ppos)
>>   {
>> @@ -3425,6 +3466,7 @@ int __init sysctl_init(void)
>>    * No sense putting this after each symbol definition, twice,
>>    * exception granted :-)
>>    */
>> +EXPORT_SYMBOL(proc_dobool);
>>   EXPORT_SYMBOL(proc_dointvec);
>>   EXPORT_SYMBOL(proc_douintvec);
>>   EXPORT_SYMBOL(proc_dointvec_jiffies);
>>
> 
> Friendly ping!
> 
> Luis, Kees, Iurii, could you please have a look and provide an Ack if this 
> looks ok to you?

Ping again!

Could anybody please provide an Ack?

Thanks,
  Thomas


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 1/2] sysctl: introduce new proc handler proc_dobool
  2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
  2021-08-18  9:38   ` Thomas Huth
@ 2021-09-16 22:03   ` Kees Cook
  1 sibling, 0 replies; 11+ messages in thread
From: Kees Cook @ 2021-09-16 22:03 UTC (permalink / raw)
  To: Thomas Huth
  Cc: linux-kernel, linux-fsdevel, linux-nfs, J. Bruce Fields,
	Chuck Lever, Trond Myklebust, Anna Schumaker, Luis Chamberlain,
	Iurii Zaikin, linux-s390, Jia He, Pan Xinhui

On Tue, Aug 03, 2021 at 12:59:36PM +0200, Thomas Huth wrote:
> From: Jia He <hejianet@gmail.com>
> 
> This is to let bool variable could be correctly displayed in
> big/little endian sysctl procfs. sizeof(bool) is arch dependent,
> proc_dobool should work in all arches.
> 
> Suggested-by: Pan Xinhui <xinhui@linux.vnet.ibm.com>
> Signed-off-by: Jia He <hejianet@gmail.com>

Hi! I apologize for the delay. Yes, this looks good to me; thanks!

Acked-by: Kees Cook <keescook@chromium.org>

-- 
Kees Cook

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames
  2021-08-03 10:59 ` [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames Thomas Huth
@ 2021-09-16 22:05   ` Kees Cook
  0 siblings, 0 replies; 11+ messages in thread
From: Kees Cook @ 2021-09-16 22:05 UTC (permalink / raw)
  To: Thomas Huth
  Cc: linux-kernel, linux-fsdevel, linux-nfs, J. Bruce Fields,
	Chuck Lever, Trond Myklebust, Anna Schumaker, Luis Chamberlain,
	Iurii Zaikin, linux-s390, Jia He, Pan Xinhui

On Tue, Aug 03, 2021 at 12:59:37PM +0200, Thomas Huth wrote:
> From: Jia He <hejianet@gmail.com>
> 
> nsm_use_hostnames is a module parameter and it will be exported to sysctl
> procfs. This is to let user sometimes change it from userspace. But the
> minimal unit for sysctl procfs read/write it sizeof(int).
> In big endian system, the converting from/to  bool to/from int will cause
> error for proc items.
> 
> This patch use a new proc_handler proc_dobool to fix it.
> 
> Signed-off-by: Jia He <hejianet@gmail.com>
> Reviewed-by: Pan Xinhui <xinhui.pan@linux.vnet.ibm.com>
> [thuth: Fix typo in commit message]
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  fs/lockd/svc.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
> index 2de048f80eb8..0ab9756ed235 100644
> --- a/fs/lockd/svc.c
> +++ b/fs/lockd/svc.c
> @@ -584,7 +584,7 @@ static struct ctl_table nlm_sysctls[] = {
>  		.data		= &nsm_use_hostnames,
>  		.maxlen		= sizeof(int),

For robustness, maybe this should be:

		.maxlen		= sizeof(nsm_use_hostnames),

>  		.mode		= 0644,
> -		.proc_handler	= proc_dointvec,
> +		.proc_handler	= proc_dobool,
>  	},
>  	{
>  		.procname	= "nsm_local_state",
> -- 
> 2.27.0
> 

Reviewed-by: Kees Cook <keescook@chromium.org>

Also, I wonder what other BE-corrupted bools are out there?

-- 
Kees Cook

^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2021-09-16 22:05 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-03 10:59 [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines Thomas Huth
2021-08-03 10:59 ` [PATCH 1/2] sysctl: introduce new proc handler proc_dobool Thomas Huth
2021-08-18  9:38   ` Thomas Huth
2021-09-16  5:16     ` Thomas Huth
2021-09-16 22:03   ` Kees Cook
2021-08-03 10:59 ` [PATCH 2/2] lockd: change the proc_handler for nsm_use_hostnames Thomas Huth
2021-09-16 22:05   ` Kees Cook
2021-08-03 14:08 ` [PATCH 0/2] Fix /proc/sys/fs/nfs/nsm_use_hostnames on big endian machines J. Bruce Fields
2021-08-03 14:24   ` Chuck Lever III
2021-08-04  2:12 ` Jia He
2021-08-09 17:50 ` Chuck Lever III

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).