Skip to contentSkip to conversation

sys32_ioctl -> compat_ioctl -- generic

PATCH Pavel Machek 4 participants 7 messages Mar 3, 2003 – Mar 10, 2003

lore
Share & tools
lore
#0 lore raw

Hi!

This is generic part of sys32_ioctl -> compat_ioctl. Please apply,

Pavel

include/linux/ioctl32.h +5 -0
@@ -19,5 +19,10 @@
 
 extern int unregister_ioctl32_conversion(unsigned int cmd);
 
+struct ioctl_trans {
+	unsigned long cmd;
+	int (*handler)(unsigned int, unsigned int, unsigned long, struct file * filp);
+	struct ioctl_trans *next;
+};
 
 #endif
kernel/compat.c +222 -0

231 lines not on this page · raw message

Signature · 2 lines
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
#1 lore raw

Hi Pavel,

Quoted 2 lines · On Tue, 4 Mar 2003 00:21:22 +0100 Pavel Machek <[email protected]> wrote:

This is generic part of sys32_ioctl -> compat_ioctl. Please apply,

Thanks for this - you saved me a headache :-)

Some comments:

kernel/compat.c

Quoted 2 lines
kernel/compat.c +0 -0

All this really belongs in fs/compat.c ...

One thing that Linus (and I) wanted from the compatability layer is to try to keep all 32 bit assumptions out of the generic code - I understand that this my not be possible, but we would like to try.

So maybe you could start by changing ioctl32 to compat_ioctl everywhere - I know that this is just cosmetic, but it gives the better impression of what the code is about ...

Signature · 2 lines
Cheers,
Stephen Rothwell                    [email protected]
#2 lore raw

Hi!

Quoted 3 lines
Quoted 1 line

This is generic part of sys32_ioctl -> compat_ioctl. Please apply,

Thanks for this - you saved me a headache :-)

Well, I fear total sum of headache stayed constant ;-).

Anyway, producing these patches seems "easy" compared to actually merging them. You seem to have some experience with this, would you be willing to help?

Quoted 7 lines

Some comments:

kernel/compat.c

Quoted 2 lines
kernel/compat.c +0 -0

All this really belongs in fs/compat.c ...

Quoted 7 lines

One thing that Linus (and I) wanted from the compatability layer is to try to keep all 32 bit assumptions out of the generic code - I understand that this my not be possible, but we would like to try.

So maybe you could start by changing ioctl32 to compat_ioctl everywhere - I know that this is just cosmetic, but it gives the better impression of what the code is about ...

I thought about that, but I fear resulting diff would be too big (and would bitrot extremely quickly). I thought I'd try to merge simple (& small) stuff first, to see how bit it will be.....

Pavel

Signature · 2 lines
Horseback riding is like software...
...vgf orggre jura vgf serr.
#3 lore raw
In reply to Pavel Machek
Quoted 2 lines · On Tue, 4 Mar 2003 00:21:22 +0100 Pavel Machek <[email protected]> wrote:

+asmlinkage long compat_ioctl(unsigned int fd, unsigned int cmd, unsigned long arg)

For consistancy, this should be called compat_sys_ioctl.

kernel/compat.c line 423

Quoted 8 lines
kernel/compat.c +6 -0
+{
+	struct file * filp;
 
+	filp = fget(fd);
 
+			/* find the name of the device. */
+			if (path) {
+				struct file *f = fget(fd); 

Is it really necessary to do another fget(fd) ?

Also, if you are adding this much code, you should add a copyright notice to the top of the file ...

Signature · 2 lines
Cheers,
Stephen Rothwell                    [email protected]
#4 lore raw

Hi!

Quoted 3 lines
Quoted 1 line

+asmlinkage long compat_ioctl(unsigned int fd, unsigned int cmd, unsigned long arg)

For consistancy, this should be called compat_sys_ioctl.

Done. (And moved whole stuff to fs/compat.c).

Quoted 10 lines

kernel/compat.c line 423

Quoted 8 lines
kernel/compat.c +6 -0
+{
+	struct file * filp;
 
+	filp = fget(fd);
 
+			/* find the name of the device. */
+			if (path) {
+				struct file *f = fget(fd); 

Is it really necessary to do another fget(fd) ?

This is andi's code, but it seems unneeded, fixed.

Quoted 2 lines

Also, if you are adding this much code, you should add a copyright notice to the top of the file ...

I actually need to copy copyrights from ia32_ioctl, where I took this. Done.
Pavel

Signature · 2 lines
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
#5 lore raw

Hi Pavel,

Sorry I haven't responded earlier. I am willing to help in any way I can, so let me know.

Quoted 6 lines · On Fri, 7 Mar 2003 00:37:21 +0100 Pavel Machek <[email protected]> wrote:

Hi!

Quoted 1 line

For consistancy, this should be called compat_sys_ioctl.

Done. (And moved whole stuff to fs/compat.c).

Great.

Quoted 1 line

This is andi's code, but it seems unneeded, fixed.

I assume Andi will scream if there is something subtle there :-)

Quoted 5 lines
Quoted 2 lines

Also, if you are adding this much code, you should add a copyright notice to the top of the file ...

I actually need to copy copyrights from ia32_ioctl, where I took this. Done.

OK, I will have a deeper look soon.

Signature · 3 lines
Cheers,
Stephen Rothwell                    [email protected]
http://www.canb.auug.org.au/~sfr/
#6 lore raw
In reply to Pavel Machek
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
kernel/compat.c +3 -0
+{
+	struct file * filp;
 
+	filp = fget(fd);
Quoted 7 lines
Quoted 5 lines

kernel/compat.c line 423

Quoted 3 lines
kernel/compat.c +3 -0
+			/* find the name of the device. */
+			if (path) {
+				struct file *f = fget(fd); 

Is it really necessary to do another fget(fd) ?

This is andi's code, but it seems unneeded, fixed.

Yes, it's redundant.

-Andi