Greetings,
On 11-Sep-01 Alan Cox wrote:
>> This is the second message of this kind that I received. Seems like
>> vmalloc() does weird things on machines with 1GB or more of physical RAM.
>
> It shouldnt do so. The amount of vmalloc memory is partly dependant on the
> amount of RAM because you need address space for the mappings, but that
> is in the other of many megabytes of space.
I need less than 2 MB in total, in 3 chunks; see below.
>> 128KB; plus, they may need to get mmap()ed. So I (abused) the rvmalloc()
>> code from bttv and other drivers, which uses vmalloc(). Vmalloc() itself
>
> rvmalloc uses vmalloc_32 which keeps the memory below 4Gb - that is needed
> on the Bttv at least because >4Gb isnt a valid DMA target for the USB
> controller
A small correction is in order here: the failed function is vmalloc(), not
rvmalloc(). Nevertheless, something weird happens. I�m allocation 3 frame
buffers, each 460812 bytes, in separate calls. I found the previous bug
report, which states:
<<QUOTE>>
derek@monkey:/opt/xawtv-3.58/bin
> ./xawtv
This is xawtv-3.58, running on Linux/i686 (2.4.8)
can't open /dev/video0: Cannot allocate memory
v4l-conf had some trouble, trying to continue anyway
v4l2: open /dev/video0: Cannot allocate memory
v4l: open /dev/video0: Cannot allocate memory
no video grabber device available
----
and in the message log I get:
----
Aug 14 21:12:03 monkey kernel: pwc Failed to allocate frame buffer 1.
----
The info you need is:
Kernel 2.4.8, no patches. I assume that I don't need anything from your site
other than the pwcx module when using this kernel. (The website doesn't make
this clear.)
USB is a "USB Controller: Intel Corporation 82371AB PIIX4 USB (rev 01)"
The machine has dual pentiums and 1GB of RAM - plenty free
<</QUOTE>>
The odd part is that it manages to allocate frame buffer 0, but not buffer 1!
The code itself is in pwc-if.c, around line 312 in pwc_allocate_buffers().
FRAME_SIZE is 460812.
The second bug report said the module works fine when the person limits his
memory to 960 MB (how, I don�t know). With full memory, he gets artefacts
which to me suggests the second frame buffer is allocated, but somehow in
memory which isn�t accessible (every other frame is bright pink --> filled
with 0xFF before YUV->RGB translation?)
Unfortunately I�m not rich enough to put 1GB of RAM in my machine to test
this... ;-(
- Nemosoft
-----------------------------------------------------------------------------
Try SorceryNet! One of the best IRC-networks around! irc.sorcery.net:9000
URL: never IRC: nemosoft IscaBBS (bbs.isca.uiowa.edu): Nemosoft
>> Never mind the daylight <<
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel