[lkml]   [2000]   [Aug]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectPCI Memory Mapping

First, please CC your emails to me, since I'm not subscribed.
Secondly, I know C, but I'm not anywhere near what you'd call a kernel
hacker. So dumb things down for me, but not _too_ much :)

Also note that I use Linux 2.2.15, as much as I would want to use 2.4,
I don't feel like it would be keen of me to switch to a development kernel
until this release kernel's driver works.

Anyway, What I'm doing is scanning the PCI bus, and finding the one
or more cards that have 16mb of on-board memory. base_address[0] if you
will. I AND that with PCI_BASE_ADDRESS_IO_MASK, and register a device
major and minor for the device.

A userland program will now attempt to mmap the device. It requests
exactly 16mb. Everything seems ok until I try to read or write to this
memory. It gives me a kernel complaint:

do_wp_page: bogus page at address 4017400 (72000000)

base_address[0] is 0xF2000008, and with the mask, F2000000. According to
the device's documentation, both are acceptable (for what they are).

I've also pasted in my entire mmap handler:
(please excuse me if this email is a bit long)

int device_mmap(struct file * filp, struct vm_area_struct * vma)
int id = MINOR(filp->f_dentry->d_inode->i_rdev)-1;
/* this id is supposed to be -1 */
e4DevT *dev = &devTable[id];

unsigned long off = vma->vm_offset;
unsigned long physical=dev->ioaddr+off+PAGE_OFFSET;
unsigned long vsize=vma->vm_end - vma->vm_start;
unsigned long psize=dev->ioaddr - off;

if (off & (PAGE_SIZE-1))
printk("mmap: page unaligned\n");
return -ENXIO;
if (vsize > psize)
printk("too high\n");
return -EINVAL;
if (remap_page_range(vma->vm_start, physical, vsize, vma->vm_page_prot))
return 0;

printk("mmap (remap_page_range) device #%d successful\n",id);
/* this message is printed */

return 0;

Thanks for your assistance.

P.S. There needs to be reference/documentation for all these functions,
something easier to grep.

-Charles <>

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:58    [W:0.035 / U:2.024 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site