[ECOS] cf_get_CIS () not right?

[Trenton D. Adams]

I was looking at cf_get_CIS () and noticed that when creating a buffer
from a section in the CIS such as CF_CISTPL_CONFIG, it doesn't seem to
do it quite right from what I understand of the CIS.

For the CF_CISTPL_CONFIG section, the first LINK field that comes right
after the CF_CISTPL_CONFIG(0x1a) id will only indicate the size of the
first part of the CONFIG section.  The next part starts with a
CF_CISTPL_CFTABLE_ENTRY (0x1b).  In the case of my card, I only have one
CF_CISTPL_CFTABLE_ENTRY and the CONFIG section ends at the end of that.

Below is the CF_CISTPL_CONFIG section for my WaveLAN card.  The 0x07 for
the second byte does not indicate the full size of the CF_CISTPL_CONFIG
section.  It only indicates the size of the section before the first
CF_CISTPL_CFTABLE_ENTRY.  

0xbf3d8 <stack+30648>:  0x1a    0x07    0x03    0x01    0xe0    0x03
0x00    0x00
0xbf3e0 <stack+30656>:  0x01    0x1b    0x0f    0xc1    0x01    0x19
0x76    0xc5
0xbf3e8 <stack+30664>:  0x4b    0xd5    0x19    0x36    0x36    0x05
0x46    0x7f
0xbf3f0 <stack+30672>:  0xff    0xff

The buffer above was obtained with the new cf_get_CIS () below.

bool
cf_get_CIS(struct cf_slot *slot, unsigned char id, 
           unsigned char *buf, int *len, int *ptr)
{
    int i, size;
    unsigned char *cis = slot->attr;
    unsigned char *cis_end = cis + slot->attr_length;
    bool bFound = false;
    *len = 0;

    cis += *ptr;
    while (cis < cis_end) 
    {
        if (*cis == 0xFF) 
        {
            break;
        }
        if (*cis == id || bFound) 
        {
            size = *(cis+2) + 2;
            for (i = 0;  i < size;  i++) 
            {
                *buf++ = *cis;
                cis += 2;
            }
            *len += size;
            *ptr = (unsigned long)(cis - slot->attr);
            bFound = true;
        } 
        else 
        {
            // Skip to next entry
            cis += (*(cis+2) * 2) + 4;
        }
    }
    return bFound;
}


Is this correct!?!?

p.s.
I think there's something wrong with the cf_parse_XXX functions too, but
I haven't fiddled with them much yet.

Trenton D. Adams
Extreme Engineering
#17, 6025 - 12 St. SE
Calgary, Alberta, Canada
T2H 2K1

Phone: 403 640 9494 ext-208
Fax: 403 640 9599

http://www.extremeeng.com