Thread: [Ndiswrapper-general] Netgear MA111 + Fedora Core 3 == lockup
Status: Beta
Brought to you by:
pgiri
From: John H. <jc...@th...> - 2004-12-13 09:28:20
|
Hello All, I'm trying to get ndiswrapper with an MA111 working on Fedora Core 3. I've tried three kernels, the standard Fedora Core 3 one (which I didn't expect to work), a stock 2.6.9 with 8k stacks which I did expect to work and latterly with the latest (681) of the linuxant 16k kernels. I'm using the MA111 2.0 drivers which reputedly work. It's an MA111v1 which uses the prism chipset. I've tried ndiswrapper-0.12 and CVS-latest. Building 0.12 CVS latest (as of about a two hours before the log messages) gives the log that you'll find at http://www.thehaxbys.plus.com/log.gz. The machine doesn't quite lock up solid; I was on a normal console (ie not X) and the keyboard, including the caps-lock light is totally unresponsive, but I was able to select the letters to make up "reboot" from what was displayed with the mouse! The log is fairly huge, but the hang seems to have come after this lot: (NdisMSetAttributesEx:713): Enter cb66a280, c4870000 10 000001a0, 0 (NdisMSetAttributesEx:752): Exit (NdisResetEvent:2062): Enter c5eb1260 (KeResetEvent:353): Enter event = c5eb1260 (KeResetEvent:360): Exit (NdisWaitEvent:2043): Enter c5eb1260 5000 (KeWaitForSingleObject:375): Enter event = c5eb1260, reason = 0, waitmode = 0, alertable = 0, timeout = c4e7bc64 (KeWaitForSingleObject:377): object type = 0, size = 0 (KeWaitForSingleObject:383): timeout = 50000000 (KeWaitForSingleObject:393): wait ms = 3340547206 (IofCallDriver:634): Enter dev_obj = cfd192c0, irp = cffdc3c0, major_fn = f, ioctl = 220003 (usb_submit_nt_urb:477): Enter nt_urb = cce8f00c, irp = cffdc3c0, length = 72, function = 9 (usb_bulk_or_intr_trans:225): flags = 0, length = 64, buffer = cce8f06c (usb_bulk_or_intr_trans:298): submitting urb c4e9c1e0 on pipe 00000202 (usb_submit_nt_urb:586): Exit (IofCallDriver:658): Exit The times in the log don't show a delay, but I think that's more down to the delay between something happening and it being logged. What looks suspicious to me is that the keyboard lock up seemed to happen about 5 seconds after loding the module and right in the middle of these messages is "timeout = 50000000". I'm not sure what that "wait ms" is about, it's not milliseconds since the machine came up (it had only been up a minute or so before I did a modprobe ndiswrapper). The machine in question is a fairly old athlon slot CPU and so has a fairly old USB controller and I did woneer if this might be the problem especially since the MA111 works flawlessly in my comparatively new laptop. jch |
From: Christian K. <chr...@gm...> - 2004-12-13 13:33:37
|
Hello, two questions: 1. if i find this in an inf-file: ******************* >> [DeviceList] %PRISM_PCMCIA_STR% =3D PRISM_PCMCIA, = PCMCIA\D-Link-DWL-650_Wireless_PC_Card_RevP-DC64 %PCI_DESC_STR% =3D PRISM_PCI1, = PCI\VEN_1260&DEV_3873&SUBSYS_37001186 %PRISM_USB1_STR% =3D PRISM_USB1, USB\VID_2001&PID_3700 %PRISM_USB2_STR% =3D PRISM_USB2, USB\VID_2001&PID_3702 <<******************* and a little bit later this: ******************* >> [PRISM_PCMCIA] AddReg =3D PRISM_PCMCIA.reg, COMMON_PCMCIA.reg, COMMON_NDIS.reg, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy [PRISM_PCMCIA.NT] AddReg =3D PRISM_PCMCIA.reg, COMMON_PCMCIA.reg, COMMON_NDIS.reg.NT, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy.NT BusType =3D 8 Characteristics=3D 0x84 [PRISM_PCMCIA.NT.Services] AddService=3D PRISM, 2, PRISM_DRIVER.Service, PRISM_DRIVER.EventLog [PRISM_PCMCIA.reg] HKR,Ndi = ,DeviceID,0,"PCMCIA\D-Link-DWL-650_Wireless_PC_Card_RevP-DC64" HKR, ,PlatformId,0,"32797" HKR, ,VendorDesc,0,%PRISM_PCMCIA_STR% HKR, ,LoadFirmware,0,"1" ;HKR,NDI\params\LoadFirmware,,0,1 ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D <<******************* What does this mean to me? I*m quite sure that the ndiswrapper-perl-Script won't recognize the correct bustype, because ParseID only looks for USB and PCI? And this explicit 'BusType =3D 8' is ignored? Because there is something else a little bit strange in the next = section: *******************>> ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D [PRISM_PCI1] AddReg =3D PRISM_PCI1.reg, COMMON_PCI.reg, COMMON_NDIS.reg, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy [PRISM_PCI1.NT] AddReg =3D PRISM_PCI1.reg, COMMON_PCI.reg, COMMON_NDIS.reg.NT, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy.NT BusType =3D 5 Characteristics=3D 0x84 [PRISM_PCI1.NT.Services] AddService=3D PRISM, 2, PRISM_DRIVER.Service, PRISM_DRIVER.EventLog [PRISM_PCI1.reg] HKR,Ndi ,DeviceID,0,"PCI\VEN_1260&DEV_3873&SUBSYS_37001186" HKR, ,PlatformId,0,"32804" HKR, ,VendorDesc,0,%PCI_DESC_STR% HKR, ,LoadFirmware,0,"1" ;HKR,NDI\params\LoadFirmware,,0,1 ;=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D [PRISM_USB1] ; Win9x AddReg =3D PRISM_USB1.reg, COMMON_USB.reg, COMMON_NDIS.reg, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy [PRISM_USB1.NT] ; Win2k AddReg =3D PRISM_USB1.reg, COMMON_USB.reg, COMMON_NDIS.reg.NT, = COMMON.reg CopyFiles =3D PRISM_DRIVER.copy.NT BusType =3D 15 ; USB Characteristics=3D 0x84 ; NCF_PHYSICAL | NCF_HAS_UI [PRISM_USB1.NT.Services] AddService =3D PRISM_USB, 2, PRISM_USB_DRIVER.Service, = PRISM_DRIVER.EventLog [PRISM_USB1.reg] HKR,Ndi ,DeviceID,0,"USB\VID_2001&PID_3700" HKR, ,PlatformId,0,"0" HKR, ,VendorDesc,0,%PRISM_USB1_STR% HKR, ,LoadFirmware,0,"1" ;HKR, NDI\params\LoadFirmware,,0,"1" <<******************* What does this mean, Bustype =3D 15? I think, bustype for USB is 0? 2. The structure of the windows-driver-files is as the following: d 0 2004-07-13 15:08:00 Drivers/ d 0 2004-07-13 16:22:00 Drivers/WinXP/ - 666624 2003-10-02 14:47:00 Drivers/WinXP/PRISMUSB.sys - 676352 2003-10-02 14:21:00 Drivers/WinXP/PRISMNDS.sys - 666710 2003-10-02 14:48:00 Drivers/PRISMUSB.sys - 8420 2003-12-04 11:14:00 Drivers/PRISMNIC.cat - 676438 2003-10-02 14:22:00 Drivers/PRISMNDS.sys - 15557 2004-04-20 14:43:00 Drivers/NETPRISM.inf You see, the WinXP-Files are in the WinXP subdirectory. In the inf-file the copy-File and source-Disk-section looks like this: *******************>> ;CopyFiles Section =3D Destination Directory ID -- see layout.inf ;----------------- ------------------------ DefaultDestDir =3D 11 ; Win9x=3D%windir%\system = Win2k=3D%windir%\system32 PRISM_DRIVER.copy =3D 11 ; Win9x=3D%windir%\system PRISM_DRIVER.copy.NT =3D 12 ; Win2k=3D%windir%\system32\drivers [PRISM_DRIVER.copy] PRISMNDS.sys PRISMUSB.sys [PRISM_DRIVER.copy.NT] PRISMNDS.sys PRISMUSB.sys [SourceDisksNames] ;Source Disk ID =3D Disk Name ;-------------- --------- 1 =3D %INSTALL_DISK_STR%,,, [SourceDisksFiles] ; Win9x ;File Name =3D Source Disk ID ;--------- -------------- PRISMNDS.sys =3D 1 PRISMUSB.sys =3D 1 [SourceDisksFiles.X86] ; Win2k/Xp PRISMNDS.sys =3D 1,WinXP PRISMUSB.sys =3D 1,WinXP <<******************* This strange '1,WinXP' substitution is worring me.=20 What does this mean, a win_install-specific-automagic? Is this something someone have to take care about in the=20 next Version of the ndiswrapper-script? I*ve added the whole inf-file as tgz. Greetings, Christian |
From: Christian K. <chr...@gm...> - 2004-12-13 13:45:14
Attachments:
NETPRISM.tgz
|
> I*ve added the whole inf-file as tgz. :-( |
From: Ralf H. <Ral...@ch...> - 2004-12-14 14:16:23
|
* Christian Kneisel <chr...@gm...>: >=20 > > I*ve added the whole inf-file as tgz. >=20 > :-( Say, how often is this mail going to go around? --=20 Ralf Hildebrandt (i.A. des IT-Zentrum) Ralf.Hildebrandt@charite.= de Charite - Universit=E4tsmedizin Berlin Tel. +49 (0)30-450 570= -155 Gemeinsame Einrichtung von FU- und HU-Berlin Fax. +49 (0)30-450 570-9= 62 IT-Zentrum Standort CBF send no mail to spamtrap@charite.= de |
From: Christian K. <chr...@gm...> - 2004-12-14 16:06:50
|
>>> I*ve added the whole inf-file as tgz. >>=20 >> :-( >=20 > Say, how often is this mail going to go around? I only forget in the previous mail to add it and I don*t know why you asked it?=20 You*ve found something in this inf-file and why=20 it isn*t installed correctly? |
From: Dominique R. <do...@le...> - 2004-12-13 15:24:09
|
Le Mon, Dec 13, 2004 at 09:28:06AM +0000, John Haxby [jc...@th...= ] a =E9crit: > kernels. I'm using the MA111 2.0 drivers which reputedly work. It's=20 > an MA111v1 which uses the prism chipset. This should work with prism2_usb, I think. |
From: John H. <jc...@th...> - 2004-12-14 15:21:10
|
Dominique Rousseau wrote: >Le Mon, Dec 13, 2004 at 09:28:06AM +0000, John Haxby [jc...@th....u= k] a =E9crit: > =20 > >>kernels. I'm using the MA111 2.0 drivers which reputedly work. It's=20 >>an MA111v1 which uses the prism chipset. >> =20 >> > >This should work with prism2_usb, I think. > =20 > I was rather hoping to get away from wlan-ng and its painful=20 installation and configuration process. HostAP, the other likely=20 source of something called prism_usb doesn't claim support for the MA111=20 although I guess it's worth a try. jch |
From: Adrian Irving-B. <wis...@wi...> - 2004-12-14 16:17:53
|
On Tue, Dec 14, 2004 at 03:20:58PM +0000, John Haxby wrote: > I was rather hoping to get away from wlan-ng and its painful > installation and configuration process. Yes. Not adhering to the Linux Wireless Extensions was a rather bad idea on their part. For me, at least, it makes their driverset fairly useless. > HostAP, the other likely source of something called prism_usb > doesn't claim support for the MA111 although I guess it's > worth a try. As I recall, HostAP was not only what the name implies, but also something you could add on top of the Prism drivers to give them LWE support (see above). But last time I checked, HostAP supports nothing USB. |
From: John H. <jc...@th...> - 2004-12-19 19:51:10
|
John Haxby wrote: > Hello All, > > I'm trying to get ndiswrapper with an MA111 working on Fedora Core > 3. I've tried three kernels, the standard Fedora Core 3 one (which I > didn't expect to work), a stock 2.6.9 with 8k stacks which I did > expect to work and latterly with the latest (681) of the linuxant 16k > kernels. I'm using the MA111 2.0 drivers which reputedly work. It's > an MA111v1 which uses the prism chipset. In case anyone was interested, the problem was (and indeed, still is) a bad interaction between the ohci_hcd driver and the antique AMD-756 USB device (1022:740c, rev 06). I bought a cheap USB 2.0 PCI card and everything is fine. Indeed, it's happy running with the Fedora Core kernel, including its 4k stacks. Which is nice. jch |