NdisMQueryAdapterResources() : resource allocation problem at WinMe

From: Ashirwad (ash_at_athenasemi.com)
Date: 07/19/04


Date: 19 Jul 2004 09:26:04 -0700

Hi,

I am stuck with a similar problem (as mentioned below in one of the
older postings). Driver works with XP and 2K. Infact it works with
WinME installed on PC where I use PCI-Cardbus(Ricoh) controller to
insert my cardbus wireless lan controller. It doesn't work with WinME
(on IBM Thinkpad A21m laptop).

My NdisMQueryAdapterResources() returns only one resource i.e.
interrupt. Using debugger command 'pci' I see that BAR seems to be
mapped. No difference when I use NdisMPciAssignResources() wrt number
of resources but here I get IO enabled in addition to BusMaster and
Memory (i.e. control word 0x7 in NdisMPCIAssignResources() against 0x6
in NdisMQueryAdapterResources() .

There is one imp print in the debugger in the reboot that follows
crash because of pci-bar not getting mapped. It says"VPICD Severe
Perf Warning: interrupt #9 is supriousing. Get the owners of the
devices on that interrupt to look at their code." This problem exists
till I don't shutdown the m/c and start. Device manager shows that
'SCI IRQ used by ACPI bus' is the interrupt
that prompts that warning. My device is on interrupt 0xB. Don't know
whether this problem is linked to above resource allocation problem.

Hi Daum,

Found you query in this ng. Did you get the solution for this ?

I couldn't find post-followup-message in your posting hence creating a
new thread.

Went through "Another NDIS, PnP, PCI question" thread in
"comp.os.ms-windows.programmer.nt.kernel-mode" which is related
problem but changing eeprom contents from 'other controller' to
'ethernet controller' did not help.

Thanks,
Ashirwad.

>From: daum (daum999@yahoo.com)
>Subject: [Q] Resource allocation problem at WinMe : physical memory
>start .address always 0.
 
  
>View this article only
>Newsgroups: microsoft.public.development.device.drivers
>Date: 2003-05-30 10:57:44 PST
 

>Subject : [Q] Resource allocation problem at WinMe : physical memory
start
>address always 0.

>Hello, everyone.

>I'm developing CardBus NDIS D/D at WinMe, and i have one problem
during
>MiniportInitialize.

>The problem never occurred at Win2K or WinXp.
>It is a problem concerned about resource allocation.
>My Card has 2 memory windows and their sizes are 0x40000 and 0x4000.
>(I think the memory size was reasonable because it is below 1MB)

>After NdisMQueryAdapterResources was called, it returns the status
with
>NDIS_STATUS_SUCCESS.

>But, the start address of the physical memory-CmResourceTypeMemory-
was
>always 0.

>Of course, the memory Length was all OK : 0x40000 and 0x4000.
>And, the returned Flags value was 0x0; so, it was non-prefetchable.
>Please remind that the problem occurred only at WinMe not Win2K or
WinXp.
>I'm not sure it will occur at Win98Se.

>Thank you for your help in advance.

>//DAUM
>This is the only article in this thread



Relevant Pages

  • Re: Interrupte erweitern
    ... Der will dem Controller die Einsprungadresse des Interrupts ... liefert den Interrupt am INT Pin des Prozessors. ... diesem Modus direkt mit dem INTA Eingang des Controllers verbunden ist. ... High- und Low-Byte auf den Datenbus. ...
    (de.sci.electronics)
  • Regression: libata: implement ata_wait_after_reset()
    ... The controller card is this one: ... ALi Corporation ALi M5281 Serial ATA / RAID Host Controller ... Interrupt: pin A routed to IRQ 19 ... spending 2s on 0xff port would incur too ...
    (Linux-Kernel)
  • Re: Question regarding ERR in /proc/interrupts.
    ... not a particular interrupt ... is supposed to put an address on the bus so the CPU can ... by a controller, ... FYI 20 spurious interrupts out of the bazzillion shown isn't ...
    (Linux-Kernel)
  • Re: [RFC][PATCH 4/4] RTC: SMBus support for the M41T80,
    ... The controller implements a command which issues a write ... That would be internal to the m41t80 driver, ... the SOC has enough internal interrupt sources you can almost assume ...
    (Linux-Kernel)
  • Re: Real-time pre-emption kswapd/kscand questions
    ... The FDDI controller may well be the key to the problem here. ... that interrupt might be sufficient to cause a small gap in the analog data ... >acquisition and sends occasional status messages back to a remote host ...
    (comp.os.linux.development.apps)