... yes ... not currently, and that is the problem ... depends on what the new api's look like.. Sam...
2238
Daniela Engert
engertd
Jan 23, 2003 6:45 pm
On Thu, 23 Jan 2003 16:38:46 -0000, "barf1234de" ... They *are* ! ... Read the interrupt line register. If the value happens to be 0 (or may be FFh) assume the...
2239
barf1234de <lars.e...
barf1234de
Jan 25, 2003 11:22 am
Hi, ok, managed to read the base addresses of I/O port and memory ranges and detected their lengths from PCI config space. Now, I want to use VMAlloc to map...
2240
Sam Detweiler
sam@...
Jan 25, 2003 1:46 pm
You need to VMLock the memory when you need to use a consistant physical to logical (Virtual 16:16 or Linear 0:32) addressing map. Like you start an I/O,...
2241
Daniela Engert
engertd
Jan 25, 2003 5:35 pm
... Well, you map the *physical* address range to a *linear* address in the system arena which the CPU accesses at a *virtual* address by constructing it from...
2242
barf1234de <lars.e...
barf1234de
Jan 26, 2003 1:45 pm
Hi, ... the ... Ok, I knew but I used the wrong words. By the way, since base addr of DosFlatDS is 0 (isn't it ?), if I have a physical addr, could I construct...
2243
Sam Detweiler
sam@...
Jan 26, 2003 2:46 pm
... In this case the VMAlloc does the VMLock for you Sam...
2244
Ruediger Ihle
ruediger_ihle
Jan 27, 2003 11:28 am
... Not that I know of. ... Unfortunately, yes. WATCOM should be able to inline this, others com- pilers most likely not. That's the reason why I use a dual...
2245
Daniela Engert
engertd
Jan 27, 2003 4:48 pm
... No, you can't, because of the way how x86 CPUs work: virtual address -> segmentation unit -> linear address -> paging unit -> physical address The mapping...
2246
Sam Detweiler
sam@...
Jan 27, 2003 4:52 pm
... Well, If the VMAlloc allocates the linear address in the process address space, and NOT the global address space, then the VMGlobalToProcess is NOT...
2247
Daniela Engert
engertd
Jan 27, 2003 5:16 pm
... True. In this case the driver writer needs to take care of the fact that the linear address is valid in this particular process context only; i.e. [at task...
2248
Sam Detweiler
sam@...
Jan 27, 2003 5:20 pm
... unless they feel 'lucky'.. remember, IF the active process at the tim eof the interrupt IS the one with the proper mapping it will work JUST fine.. its not...
2249
barf1234de <lars.e...
barf1234de
Jan 27, 2003 6:36 pm
Hi, ok, understood. The scenario was to let an application call the driver through and IOCTL, the driver allocates memory via VMALLOC (MAPPHYS|PROCESS) and...
2250
barf1234de <lars.e...
barf1234de
Jan 27, 2003 7:11 pm
Hi, (if I find the time) I want to write a MAC driver. I got the OS/2 MAC documentation and I got the documentation for the device. Now, there is this thing...
2251
Holger Veit
holger.veit@...
Jan 27, 2003 7:37 pm
... So, this is exactly what the MAP$ subdriver of XF86SUP.SYS does; with the exception that it is passed the physical address through the ioctl param packet,...
2252
Veit Kannegieser
veit.kannegieser@...
Jan 27, 2003 7:44 pm
I am about to releas a package that has - an patch for pcmcia.sys to hande CardBus cards - a "generic" CardBus enabler client written for the Realtek 8139 NIC...
2253
Veit Kannegieser
veit.kannegieser@...
Jan 27, 2003 7:59 pm
... Since it the device does hopefully not uses the CPU to transfer the data to the chip, but perhaps uses Bus master DMA, it will the physical location of the...
2254
Daniela Engert
engertd
Jan 27, 2003 8:00 pm
... Thank you very much! May be you are aware of a new cardbus driver stack due to be released rsn? This one is supposed to support card services level 8 (i.e....
2255
Sam Detweiler
sam@...
Jan 27, 2003 8:03 pm
... the ID's aren't in the spec.. you have to request them separately.. Sam...
2256
Daniela Engert
engertd
Jan 27, 2003 8:05 pm
On Mon, 27 Jan 2003 19:11:05 -0000, "barf1234de" ... Well, in this case you should coordinate your efforts with Mr.NIC to avoid duplicated work. While he is...
2257
Adrian Gschwend
netlabsorg
Jan 27, 2003 8:30 pm
... found at least some of them (hope this is the stuff you are looking for): http://www.jedec.org/DOWNLOAD/search/JEP106l.pdf the document is from December...
2258
Markus Montkowski
hugh_os2
Jan 27, 2003 8:40 pm
When's that up? I have cardbus USB 2.0 controler and it doesn't show any tuples with the version of pcmcia.sys I have, also tried your CB bridge enabler and ...
2259
Veit Kannegieser
veit.kannegieser@...
Jan 27, 2003 8:55 pm
... Thanks. I will see if i can verify the guessed numbers by looking for a diagnostics program for any platform that has the table build in... Veit...
2260
Veit Kannegieser
veit.kannegieser@...
Jan 27, 2003 8:55 pm
... Interesting. Currently CardServices/Socket services leak IO address ranges in the Resource manager when a CardBus card is removed before doing a release...
2261
Sam Detweiler
sam@...
Jan 27, 2003 9:10 pm
... I can handle that one Sam...
2262
Chris Myers
myers@...
Jan 27, 2003 9:11 pm
Here's a list of PCMCIA manufacturer IDs. http://www.pcmcia.org/tupleidlist.htm Best Regards, Chris Myers Senior Project Engineer Quatech, Inc. ...
2263
Veit Kannegieser
veit.kannegieser@...
Jan 27, 2003 9:48 pm
... Only the first bank seems to match. ... I have used some .h file from that. Will try again with an newer copy. (num:$0101; str:'3Com'...
2264
barf1234de <lars.e...
barf1234de
Jan 28, 2003 8:37 am
Hi, ... link ... clearly ... Oops, I haven't read the NDIS spec completely. The data structures passed from the OS to the driver are a bit different than the...
2265
serge sterck
serge_sterck
Jan 28, 2003 10:54 am
Can somebody provide me the source of a working sess driver or better has anyone the source of the fwatch.sys to modify him to provide a speed > 9600 on a...
2266
Veit Kannegieser
veit.kannegieser@...
Jan 28, 2003 2:51 pm
... Ok. I will report if i see it with the updated Card services release. Veit Kannegieser...
