Is there any reason why the debug parameter in the iwlwifi driver
should be read-only?
(Patch inlined and attached since I'm sure it will get whitespace
broken).
The debug parameter of the Intel 3945ABG driver should be writable
via sysfs.
Signed-off-by: SL Baur <[email protected]>
---
This patch is against Linus' 2.6.25 release.
--- linux-2.6/drivers/net/wireless/iwlwifi/iwl3945-base.c~ 2008-04-17
10:15:33.000000000 -0700
+++ linux-2.6/drivers/net/wireless/iwlwifi/iwl3945-base.c 2008-04-20
23:49:35.000000000 -0700
@@ -8855,7 +8855,7 @@ MODULE_PARM_DESC(disable, "manually disa
module_param_named(hwcrypto, iwl3945_param_hwcrypto, int, 0444);
MODULE_PARM_DESC(hwcrypto,
"using hardware crypto engine (default 0 [software])\n");
-module_param_named(debug, iwl3945_param_debug, int, 0444);
+module_param_named(debug, iwl3945_param_debug, int, 0644);
MODULE_PARM_DESC(debug, "debug output mask");
module_param_named(disable_hw_scan, iwl3945_param_disable_hw_scan, int, 0444);
MODULE_PARM_DESC(disable_hw_scan, "disable hardware scanning (default 0)");
-sb
On 4/21/08, Tomas Winkler <[email protected]> wrote:
> On Mon, Apr 21, 2008 at 10:10 AM, SL Baur <[email protected]> wrote:
> > Is there any reason why the debug parameter in the iwlwifi driver
> > should be read-only?
> it doesn't have any affect on debug output though. This variable is
> copied to to another one in pci proble
> so unless you have 2 cards there is no use for it.
> Run time debug output can be handled through debug_level sysfs entry
> /sys/bus/pci/drivers/iwl4965/debug_level
> Yet this will be changed as well to be per device instead of per driver
O.K. The driver help recommended turning on debug for first time usage
and I was surprised when I didn't get any debug output. And I see now
that /proc/bus/pci/drivers/iwl3945/debug_level *is* writable. I'm lost in a
twisty maze of /sys directories, all subtlely different.
The driver itself works fine, but I'm having trouble getting decent speed out
of it, similar to the horrible performance I had under another operating
system, not quite as bad but still worse than a MacBook Pro. Any clues on what
I can look for to see why it so relatively slow? I'm going to put it to the WoW
test as soon as I can get a modern userland & Wine on the machine, it's running
RHEL 4 now. The machine is a Lenovo T60.
-sb
On Mon, Apr 21, 2008 at 10:10 AM, SL Baur <[email protected]> wrote:
> Is there any reason why the debug parameter in the iwlwifi driver
> should be read-only?
>
> (Patch inlined and attached since I'm sure it will get whitespace
> broken).
>
>
> The debug parameter of the Intel 3945ABG driver should be writable
> via sysfs.
>
> Signed-off-by: SL Baur <[email protected]>
> ---
> This patch is against Linus' 2.6.25 release.
>
> --- linux-2.6/drivers/net/wireless/iwlwifi/iwl3945-base.c~ 2008-04-17
> 10:15:33.000000000 -0700
> +++ linux-2.6/drivers/net/wireless/iwlwifi/iwl3945-base.c 2008-04-20
> 23:49:35.000000000 -0700
> @@ -8855,7 +8855,7 @@ MODULE_PARM_DESC(disable, "manually disa
> module_param_named(hwcrypto, iwl3945_param_hwcrypto, int, 0444);
> MODULE_PARM_DESC(hwcrypto,
> "using hardware crypto engine (default 0 [software])\n");
> -module_param_named(debug, iwl3945_param_debug, int, 0444);
> +module_param_named(debug, iwl3945_param_debug, int, 0644);
> MODULE_PARM_DESC(debug, "debug output mask");
> module_param_named(disable_hw_scan, iwl3945_param_disable_hw_scan, int, 0444);
> MODULE_PARM_DESC(disable_hw_scan, "disable hardware scanning (default 0)");
>
> -sb
>
it doesn't have any affect on debug output though. This variable is
copied to to another one in pci proble
so unless you have 2 cards there is no use for it.
Run time debug output can be handled through debug_level sysfs entry
/sys/bus/pci/drivers/iwl4965/debug_level
Yet this will be changed as well to be per device instead of per driver
Tomas