Allow the user to restrict the RC5 address
PATCH 1/1 Allow the user to restrict the RC5 address
Mauro,
This patch allows users with multiple remotes to specify an RC5 address for a remote from which key codes will be accepted. If no address is specified, the default value of 0 accepts key codes from any remote. This replaces the current hard-coded address checks, which are too restrictive.
drivers/media/video/ir-kbd-i2c.new renamed from drivers/media/video/ir-kbd-i2c.c +5 -2
@@ -58,6 +58,9 @@ module_param(hauppauge, int, 0644); /* Choose Hauppauge remote */ MODULE_PARM_DESC(hauppauge, "Specify Hauppauge remote: 0=black, 1=grey (defaults to 0)"); +static unsigned int device; +module_param(device, uint, 0644); /* RC5 device address */ +MODULE_PARM_DESC(device, "Specify device address: 0=any (defaults to 0)"); #define DEVNAME "ir-kbd-i2c" #define dprintk(level, fmt, arg...) if (debug >= level) \
@@ -104,8 +107,8 @@ /* invalid key press */ return 0; - if (dev!=0x1e && dev!=0x1f) - /* not a hauppauge remote */ + if (device && device != dev) + /* not an acceptable remote */ return 0; if (!range)
Quoted 7 lines · On Thursday 26 March 2009 03:34:53 Udo A. Steinberg wrote:
Mauro,
This patch allows users with multiple remotes to specify an RC5 address for a remote from which key codes will be accepted. If no address is specified, the default value of 0 accepts key codes from any remote. This replaces the current hard-coded address checks, which are too restrictive.
I think this should be reviewed by Steve Toth first (CC-ed him).
One thing that this patch breaks is if you have multiple Hauppauge remotes, some sending 0x1e, some 0x1f. With this patch I can't use both, only one.
It might be better to have an option to explicitly allow old Hauppauge remotes that send 0x00.
Or an heuristic: if you see a remote with device 0x1e or 0x1f, then filter remotes with device 0x00 afterwards.
A third option might be to extend the hauppauge option, allowing you to specify '2=Old WinTV remote' or something like that, and if set device 0x00 will be allowed.
Regards,
Hans
Quoted 31 lines
Signed-off-by Udo Steinberg <[email protected]>---
linux-2.6.29/drivers/media/video/ir-kbd-i2c.c2009-03-24
00:12:14.000000000 +0100 +++
linux-2.6.29/drivers/media/video/ir-kbd-i2c.new 2009-03-26
03:12:11.000000000 +0100 @@ -58,6 +58,9 @@
module_param(hauppauge, int, 0644); /* Choose Hauppauge remote */
MODULE_PARM_DESC(hauppauge, "Specify Hauppauge remote: 0=black, 1=grey
(defaults to 0)");drivers/media/video/ir-kbd-i2c.new renamed from drivers/media/video/ir-kbd-i2c.c +3
-0+static unsigned int device; +module_param(device, uint, 0644); /* RC5 device address */ +MODULE_PARM_DESC(device, "Specify device address: 0=any (defaults to0)");
#define DEVNAME "ir-kbd-i2c"
#definedprintk(level, fmt, arg...)if (debug >= level) \drivers/media/video/ir-kbd-i2c.new renamed from drivers/media/video/ir-kbd-i2c.c +2
-2@@ -104,8 +107,8 @@ /* invalid key press */ return 0; - if (dev!=0x1e && dev!=0x1f) - /* not a hauppauge remote */ + if (device && device != dev) + /* not an acceptable remote */ return 0; if (!range)
Signature · 1 line
Hans Verkuil - video4linux developer - sponsored by TANDBERG
Quoted 11 lines · Hans Verkuil
Quoted 8 lines · On Thursday 26 March 2009 03:34:53 Udo A. Steinberg wrote:
Mauro,
This patch allows users with multiple remotes to specify an RC5 address for a remote from which key codes will be accepted. If no address is specified, the default value of 0 accepts key codes from any remote. This replaces the current hard-coded address checks, which are too restrictive.
Quoted 5 lines
I think this should be reviewed by Steve Toth first (CC-ed him).
One thing that this patch breaks is if you have multiple Hauppauge remotes, some sending
0x1e, some0x1f. With this patch I can't use both, only one.
Hans, thanks for bringing this to my attention.
Mauro, This patch is a regression, although a small one. it probably needs a little more work.
I too tend to have multiple remotes, I don't think it's that unusual for long standing Hauppauge customers to have many boards with many types of remotes.
Quoted 3 lines
It might be better to have an option to explicitly allow old Hauppauge remotes that send
0x00.
I could live with this. It relegates older remotes but those remotes are no longer made. This feels like a good compromise.
- Steve
On Thu, 26 Mar 2009 12:45:18 -0400 Steven Toth (ST) wrote:
ST> I too tend to have multiple remotes, I don't think it's that unusual for ST> long standing Hauppauge customers to have many boards with many types of ST> remotes.
ST>
ST> > It might be better to have an option to explicitly allow old Hauppauge ST> > remotes that send 0x00.
ST> >
ST> I could live with this. It relegates older remotes but those remotes ST> are no longer made. This feels like a good compromise.
How about changing the parameter such that it is a filter mask? The default value of 0x0 would accept all remotes. For non-zero values, each bit set in the parameter would filter the device address corresponding to that bit, e.g. 0x1 would filter address 0x0, 0x80000000 would filter address 0x1f, etc.
Cheers,
- Udo
Quoted 12 lines · On Thursday 26 March 2009 19:45:53 Udo A. Steinberg wrote:
On Thu, 26 Mar 2009 12:45:18 -0400 Steven Toth (ST) wrote:
ST> I too tend to have multiple remotes, I don't think it's that unusual for ST> long standing Hauppauge customers to have many boards with many types of ST> remotes.
ST>
ST> > It might be better to have an option to explicitly allow old Hauppauge ST> > remotes that send0x00.
ST> >
ST> I could live with this. It relegates older remotes but those remotes ST> are no longer made. This feels like a good compromise.
Quoted 5 lines
How about changing the parameter such that it is a filter mask? The default value of
0x0would accept all remotes. For non-zero values, each bit set in the parameter would filter the device address corresponding to that bit, e.g.0x1would filter address0x0,0x80000000would filter address0x1f, etc.
Seems way to complicated to me. Remember that end users have no idea about the device addresses, so it should be a simple to understand module option. So I'd just go with Steve's preference.
Regards,
Hans
Quoted 4 lines
Cheers,
- Udo
Signature · 1 line
Hans Verkuil - video4linux developer - sponsored by TANDBERG