Added Olaf Kirch the originator of nsm_use_hostnames to the cc.
Jia He <[email protected]> writes:
> nsm_use_hostnames is a module paramter 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_dou8vec.
>
> Suggested-by: Pan Xinhui <[email protected]>
> Signed-off-by: Jia He <[email protected]>
> ---
> 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 fc4084e..7a4ad9d 100644
> --- a/fs/lockd/svc.c
> +++ b/fs/lockd/svc.c
> @@ -561,7 +561,7 @@ static struct ctl_table nlm_sysctls[] = {
> .data = &nsm_use_hostnames,
> .maxlen = sizeof(int),
> .mode = 0644,
> - .proc_handler = proc_dointvec,
> + .proc_handler = proc_dou8vec,
proc_dou8vec does not exist in my tree so I don't know what it does,
but if it's name is accurate this change is wrong. As the maxlen has
not changed so you are implying that it is a vector of u8 and
sizeof(int) long.
Further this is wrong if nsm_use_hostnames is a bool as this sysctl
can be assigned values that are out of bounds for a bool.
Furthermore this is wrong in that a bool is not necessarily a u8, the
size of bool is very architecture dependent. So for either
nsm_use_hostnames needs to become an int, a proc_dobool helper needs
to be added, or the sysctl needs to be removed entirely as module
parameters can be edited at runtime through sysfs.
But I completely agree that this decade old bug needs to be fixed.
Eric
> },
> {
> .procname = "nsm_local_state",
Hi Eric
On 12/12/16 3:32 PM, Eric W. Biederman wrote:
> Added Olaf Kirch the originator of nsm_use_hostnames to the cc.
>
> Jia He <[email protected]> writes:
>
>> nsm_use_hostnames is a module paramter 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_dou8vec.
>>
>> Suggested-by: Pan Xinhui <[email protected]>
>> Signed-off-by: Jia He <[email protected]>
>> ---
>> 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 fc4084e..7a4ad9d 100644
>> --- a/fs/lockd/svc.c
>> +++ b/fs/lockd/svc.c
>> @@ -561,7 +561,7 @@ static struct ctl_table nlm_sysctls[] = {
>> .data = &nsm_use_hostnames,
>> .maxlen = sizeof(int),
>> .mode = 0644,
>> - .proc_handler = proc_dointvec,
>> + .proc_handler = proc_dou8vec,
> proc_dou8vec does not exist in my tree so I don't know what it does,
> but if it's name is accurate this change is wrong. As the maxlen has
> not changed so you are implying that it is a vector of u8 and
> sizeof(int) long.
Please see my previous patch mail in the same thread.
The new proc handler looks like:
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -2112,6 +2112,30 @@ static int proc_put_char(void __user **buf, size_t *size,
char c)
return 0;
}
+
+static int do_proc_dou8vec_conv(bool *negp, unsigned long *lvalp,
+ int *valp,
+ int write, void *data)
+{
+ if (write) {
+ if (*negp)
+ *(u8 *)valp = -*lvalp;
+ else
+ *(u8 *)valp = *lvalp;
+ } else {
+ int val = *(u8 *)valp;
+
+ if (val < 0) {
+ *negp = true;
+ *lvalp = -(unsigned long)val;
+ } else {
+ *negp = false;
+ *lvalp = (unsigned long)val;
+ }
+ }
+ return 0;
+}
+
[...]
+int proc_dou8vec(struct ctl_table *table, int write,
+ void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+ return do_proc_dointvec(table, write, buffer, lenp, ppos,
+ do_proc_dou8vec_conv, NULL);
+}
+
I didn't change much logic in
do_proc_dou8vec_conv compared with do_proc_dointvec_cov.
Just let the variable int *valp be used as a (u8 *)
>
> Further this is wrong if nsm_use_hostnames is a bool as this sysctl
> can be assigned values that are out of bounds for a bool.
I thought even without this patch, the assignment is also _correct_.
The error is just the output of procfs and sysctl
>
> Furthermore this is wrong in that a bool is not necessarily a u8, the
> size of bool is very architecture dependent. So for either
> nsm_use_hostnames needs to become an int, a proc_dobool helper needs
> to be added, or the sysctl needs to be removed entirely as module
> parameters can be edited at runtime through sysfs.
Yes, ;) this is what I did in Patchv1: change nsm_use_hostnames from
bool to u32/int. But as suggested by Pan Xinhui, I thought Patchv3
is better.
e.g. you can insmod lockd.ko nsm_use_hostnames=y if nsm_use_hostnames
is still bool type
>
> But I completely agree that this decade old bug needs to be fixed.
indeed
> Eric
>
>
>> },
>> {
>> .procname = "nsm_local_state",