I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
Is it any way to load that module? No rt2500 support means no internet for me :-/
Thanks. Cheers!
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
look in dmesg output for a more descriptive reason for the failure
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no longer using the card (though when master mode works I will switch back due to my preference for completely open source drivers instead of partially one way or the other.
I suggest you ask them to fix it up for newer kernels. The problem is in their MODULE_PARAM calls.
Trever
On Tue, 2006-02-21 at 23:09 +0100, Arjan van de Ven wrote:
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
look in dmesg output for a more descriptive reason for the failure
-- "He who chops his own wood, is warm twice." -- Abraham Lincoln
Trever L. Adams wrote:
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no longer using the card (though when master mode works I will switch back due to my preference for completely open source drivers instead of partially one way or the other. I suggest you ask them to fix it up for newer kernels. The problem is in their MODULE_PARAM calls.
http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?p=6906
I got some patch (php script) which I don't know how to apply :-/ Or it does not work for me, I don't know. I hope they will fix that very soon because FC5 is scheduled for 15th of March, which is very soon. I don't want to get stuck on FC4 because of this.
Igor Jagec wrote:
Trever L. Adams wrote:
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no longer using the card (though when master mode works I will switch back due to my preference for completely open source drivers instead of partially one way or the other. I suggest you ask them to fix it up for newer kernels. The problem is in their MODULE_PARAM calls.
http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?p=6906
I got some patch (php script) which I don't know how to apply :-/ Or it does not work for me, I don't know. I hope they will fix that very soon because FC5 is scheduled for 15th of March, which is very soon. I don't want to get stuck on FC4 because of this.
I saw your patch and tried to apply it. It was able to build but failed to insert the module. I updated to fresh CVS this afternoon and I had '/var/log/message' complaints on the ifname parameter sections (when modprobe'd) so i just commented out lines 61 and 62 also removed the ifname option line from /etc/modules.conf. Obviously not the preferred solution but i am at least on the internet on this testing machine. -mf
Michael Favia wrote:
Igor Jagec wrote:
I saw your patch and tried to apply it. It was able to build but failed to insert the module. I updated to fresh CVS this afternoon and I had '/var/log/message' complaints on the ifname parameter sections (when modprobe'd) so i just commented out lines 61 and 62 also removed the
In static_debug_parm.php file?
ifname option line from /etc/modules.conf.
In /etc/modules.conf? Or /etc/modprobe.conf?
Obviously not the preferred solution but i am at least on the internet on this testing machine. -mf
The same problem here after I succesfully applied the patch. Could you copy/paste the lines 61 and 62 here? Is that it:
// -extern int debug;
?
Thanks.
Igor Jagec wrote:
Michael Favia wrote:
Igor Jagec wrote:
I saw your patch and tried to apply it. It was able to build but failed to insert the module. I updated to fresh CVS this afternoon and I had '/var/log/message' complaints on the ifname parameter sections (when modprobe'd) so i just commented out lines 61 and 62 also removed the
In static_debug_parm.php file?
ifname option line from /etc/modules.conf.
In /etc/modules.conf? Or /etc/modprobe.conf?
Yes. Sorry bad with names and rely on TAB completion ALOT apparently.
Obviously not the preferred solution but i am at least on the internet on this testing machine. -mf
The same problem here after I succesfully applied the patch. Could you copy/paste the lines 61 and 62 here? Is that it:
// -extern int debug;
And no in rtmp_main.c Just look for the PARAM lines. See very small attached patch. Please just msg me with any further questions no need to trouble list.
The below patch is NOT the correct solution. The solutions is to fix up the calls.
The new format for module_param is module_param(NAME, TYPE, PERMISSION);
So, the NAME is obvious, it stays the same. The TYPE is one of (byte, short, ushort, int, uint, long, ulong, charp, bool or invbool).
I just now found this http://lwn.net/Articles/22197/ . It answers all the questions of how to do the fix. I wish I would have found this a month ago when I was making the driver work. It would have answered my question in minutes instead of the two hours it took me to track down.
Trever
On Thu, 2006-02-23 at 09:16 -0600, Michael Favia wrote:
Index: rtmp_main.c
RCS file: /cvsroot/rt2400/source/rt2500/Module/rtmp_main.c,v retrieving revision 1.64 diff -u -r1.64 rtmp_main.c --- rtmp_main.c 9 Feb 2006 23:42:10 -0000 1.64 +++ rtmp_main.c 22 Feb 2006 23:33:41 -0000 @@ -58,8 +58,8 @@ MODULE_PARM_DESC(debug, "Enable level: accepted values: 1 to switch debug on, 0 to switch debug off.");
static char *ifname = NULL ; -MODULE_PARM(ifname, "s"); -MODULE_PARM_DESC(ifname, "Network device name (default ra%d)"); +//MODULE_PARM(ifname, "s"); +//MODULE_PARM_DESC(ifname, "Network device name (default ra%d)");
// Following information will be show when you run 'modinfo' MODULE_AUTHOR("http://rt2x00.serialmonkey.com");
-- "Stop searching forever. Happiness is just next to you." -- Unknown
It did not help :-/ I gues I'll have to wait for rt2x00 guys, or Ralink guys, to solve that problem. I'm just warried if they don't fix that untill FC5 is released. Thanks anyway.
Trever L. Adams kaže:
The below patch is NOT the correct solution. The solutions is to fix up the calls.
The new format for module_param is module_param(NAME, TYPE, PERMISSION);
So, the NAME is obvious, it stays the same. The TYPE is one of (byte, short, ushort, int, uint, long, ulong, charp, bool or invbool).
I just now found this http://lwn.net/Articles/22197/ . It answers all the questions of how to do the fix. I wish I would have found this a month ago when I was making the driver work. It would have answered my question in minutes instead of the two hours it took me to track down.
Trever
On Thu, 2006-02-23 at 09:16 -0600, Michael Favia wrote:
Index: rtmp_main.c
RCS file: /cvsroot/rt2400/source/rt2500/Module/rtmp_main.c,v retrieving revision 1.64 diff -u -r1.64 rtmp_main.c --- rtmp_main.c 9 Feb 2006 23:42:10 -0000 1.64 +++ rtmp_main.c 22 Feb 2006 23:33:41 -0000 @@ -58,8 +58,8 @@ MODULE_PARM_DESC(debug, "Enable level: accepted values: 1 to switch debug on, 0 to switch debug off.");
static char *ifname = NULL ; -MODULE_PARM(ifname, "s"); -MODULE_PARM_DESC(ifname, "Network device name (default ra%d)"); +//MODULE_PARM(ifname, "s"); +//MODULE_PARM_DESC(ifname, "Network device name (default ra%d)");
// Following information will be show when you run 'modinfo' MODULE_AUTHOR("http://rt2x00.serialmonkey.com");
-- "Stop searching forever. Happiness is just next to you." -- Unknown
Le mercredi 22 février 2006 à 00:00 -0700, Trever L. Adams a écrit :
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no longer using the card (though when master mode works I will switch back due to my preference for completely open source drivers instead of partially one way or the other.
I suggest you ask them to fix it up for newer kernels. The problem is in their MODULE_PARAM calls.
Trever
On Tue, 2006-02-21 at 23:09 +0100, Arjan van de Ven wrote:
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
look in dmesg output for a more descriptive reason for the failure
-- "He who chops his own wood, is warm twice." -- Abraham Lincoln
This is the report you speak about ? http://sourceforge.net/tracker/index.php?func=detail&aid=1427910&gro... If yes could you answer questions asked for the final submission ? If not could you give us the link to the report you did, please ? Thanks Eric
No, that is not it. I no longer have the book mark, but it was in the support forums (vichu is the user name). I cannot remember which team member was reading my comments. They fixed up the USB part but didn't fix up the module_param stuff. Basically, in newer kernels, this has changed and the latest one now has depricated the older method. instead of "i" or what not for type it is int. Beyond that, I cannot remember what the changes were.
Trever
On Wed, 2006-02-22 at 22:44 +0100, Eric Tanguy wrote:
Le mercredi 22 février 2006 à 00:00 -0700, Trever L. Adams a écrit :
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no longer using the card (though when master mode works I will switch back due to my preference for completely open source drivers instead of partially one way or the other.
I suggest you ask them to fix it up for newer kernels. The problem is in their MODULE_PARAM calls.
Trever
On Tue, 2006-02-21 at 23:09 +0100, Arjan van de Ven wrote:
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
look in dmesg output for a more descriptive reason for the failure
-- "He who chops his own wood, is warm twice." -- Abraham Lincoln
This is the report you speak about ? http://sourceforge.net/tracker/index.php?func=detail&aid=1427910&gro... If yes could you answer questions asked for the final submission ? If not could you give us the link to the report you did, please ? Thanks Eric
-- "Magazines all too frequently lead to books and should be regarded by the prudent as the heavy petting of literature." -- Fran Lebowitz
I found it in the forum : http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=675&start=0&po... But it seems to not be relevant to this problem : "That topic is about the rt2x00 driver not the legacy driver. And the patch was not for rt2x00 but for the ieee80211 stack."
Eric
No, that is not it. I no longer have the book mark, but it was in the support forums (vichu is the user name). I cannot remember which team member was reading my comments. They fixed up the USB part but didn't fix up the module_param stuff. Basically, in newer kernels, this has changed and the latest one now has depricated the older method. instead of "i" or what not for type it is int. Beyond that, I cannot remember what the changes were.
Trever
On Wed, 2006-02-22 at 22:44 +0100, Eric Tanguy wrote:
Le mercredi 22 février 2006 à 00:00 -0700, Trever L. Adams a écrit :
This is a bug in rt2500. I reported it. It has to do with module_param call not following the new standard. They have not fixed this problem even though I reported it with a fix. I patched my own but am no
longer
using the card (though when master mode works I will switch back due
to
my preference for completely open source drivers instead of partially one way or the other.
I suggest you ask them to fix it up for newer kernels. The problem is
in
their MODULE_PARAM calls.
Trever
On Tue, 2006-02-21 at 23:09 +0100, Arjan van de Ven wrote:
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok,
but
the rt2500 module can't be loaded. On FC5t2's default kernel
everything
went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
look in dmesg output for a more descriptive reason for the failure
-- "He who chops his own wood, is warm twice." -- Abraham Lincoln
This is the report you speak about ? http://sourceforge.net/tracker/index.php?func=detail&aid=1427910&gro... If yes could you answer questions asked for the final submission ? If not could you give us the link to the report you did, please ? Thanks Eric
-- "Magazines all too frequently lead to books and should be regarded by the prudent as the heavy petting of literature." -- Fran Lebowitz
-- fedora-devel-list mailing list fedora-devel-list@redhat.com https://www.redhat.com/mailman/listinfo/fedora-devel-list
On Thu, 2006-02-23 at 16:43 +0100, Eric TANGUY wrote:
I found it in the forum : http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=675&start=0&po... But it seems to not be relevant to this problem : "That topic is about the rt2x00 driver not the legacy driver. And the patch was not for rt2x00 but for the ieee80211 stack."
Eric
At that point only the ieee80211 stack had the problems. The topic was some one else's. I just followed up and found some of the answers my self. The problem still seems to be the same, just in a different part of the code. See my last message in this thread.
BTW, ANY 2.6 kernel will use this stuff. So, if you don't mind only hitting 2.6 and up, you can just fix it the way I describe, no ifdefs, and all will be good.
Trever -- "Love is friendship set on fire." -- French Proverb
On Tue, 2006-02-21 at 21:53 +0100, Igor Jagec wrote:
I compiled beta3 and cvs drivers on FC5t3 and everything went ok, but the rt2500 module can't be loaded. On FC5t2's default kernel everything went ok (both, compiling and loading the rt2500 module).
[root@localhost Module]# /sbin/modprobe rt2500 FATAL: Error inserting rt2500 (/lib/modules/2.6.15-1.1955_FC5/extra/rt2500.ko): Invalid argument
Is it any way to load that module? No rt2500 support means no internet for me :-/
Thanks. Cheers!
-- Igor Jagec
You can fix it with my patch:
for beta3 http://dclam.com/share/param_patch_b3.txt for cvs http://dclam.com/share/param_patch_cvs.txt
go to Modules dir and: $ patch -p2 < ../../param_patch_b3.txt
that assumes you are using beta 3, and that the patch is two levels down.
Dick Dishkuvek wrote:
You can fix it with my patch: for beta3 http://dclam.com/share/param_patch_b3.txt for cvs http://dclam.com/share/param_patch_cvs.txt go to Modules dir and: $ patch -p2 < ../../param_patch_b3.txt that assumes you are using beta 3, and that the patch is two levels down.
It did help!!! Yesssss!!! Cheers mate, and thanks again!!!