2003-01-13 11:08:20

by Louis Zhuang

[permalink] [raw]
Subject: concerns about sysfs_ops

Dear Mochel,
I found you removed off/count params in new sysfs_ops functions. That's
good change for show functions, but for store function, if you don't
give the len of data, the chained xxx_store function has to assume
or/and guess how long the data is. This might be a potential issue. So I
suggest you add 'size_t count' in store function of sysfs_ops.

include/linux/sysfs.h: 1.21 1.22 louis 03/01/13 18:39:41 (modified,
needs delta)

@@ -18,7 +18,7 @@

struct sysfs_ops {
ssize_t (*show)(struct kobject *, struct attribute *,char *);
- ssize_t (*store)(struct kobject *,struct attribute *,const char *);
+ ssize_t (*store)(struct kobject *,struct attribute *,const char *,
size_t count);
};

extern int


Yours truly,
Louis Zhuang
---------------
Fault Injection Test Harness Project
BK tree: http://fault-injection.bkbits.net/linux-2.5
Home Page: http://sf.net/projects/fault-injection


2003-01-13 15:55:00

by Patrick Mochel

[permalink] [raw]
Subject: Re: concerns about sysfs_ops


On 13 Jan 2003, Louis Zhuang wrote:

> Dear Mochel,
> I found you removed off/count params in new sysfs_ops functions. That's
> good change for show functions, but for store function, if you don't
> give the len of data, the chained xxx_store function has to assume
> or/and guess how long the data is. This might be a potential issue. So I
> suggest you add 'size_t count' in store function of sysfs_ops.
>
> include/linux/sysfs.h: 1.21 1.22 louis 03/01/13 18:39:41 (modified,
> needs delta)
>
> @@ -18,7 +18,7 @@
>
> struct sysfs_ops {
> ssize_t (*show)(struct kobject *, struct attribute *,char *);
> - ssize_t (*store)(struct kobject *,struct attribute *,const char *);
> + ssize_t (*store)(struct kobject *,struct attribute *,const char *,
> size_t count);
> };

I agree, and Linus pointed out that obvious flaw in the interface when I
sent it to him on Friday. I'll fixing that, and should up in his tree
later today..

-pat

2003-01-14 01:41:15

by Louis Zhuang

[permalink] [raw]
Subject: Re: concerns about sysfs_ops

> I agree, and Linus pointed out that obvious flaw in the interface when I
> sent it to him on Friday. I'll fixing that, and should up in his tree
> later today..
>
> -pat
And more, could you possibly export 'kset_find_obj'? I thinks it is
useful for module too.
Yours truly,
Louis Zhuang
---------------
Fault Injection Test Harness Project
BK tree: http://fault-injection.bkbits.net/linux-2.5
Home Page: http://sf.net/projects/fault-injection