Re: CONFIG_BIGMEM and rawio

Manfred Spraul (manfreds@colorfullife.com)
Fri, 03 Sep 1999 01:25:06 +0200

"Stephen C. Tweedie" wrote:
> Exactly why it needs to be per-device. Once the high memory limit is
> raised above 4G, this will become even more important as not all devices
> will physically be able to support 64-bit PCI addressing.

I doubt that bounce buffers are necessary for 4 GB:
* most fast disk controllers use DMA
* noone would combine a 4 GB memory computer with an old IDE disk.

But perhaps you should add them for 64-bit PCI.

Btw, if you add per-device flags, please add a central "block device
busy"-flag:
* raw-io does not set the blocksize for VFS mounts, but it ignores md
volumes and swap partitions.
* "mount -t ext2 /dev/swappartition" makes swapping unstable.
* an accidential "swapon /dev/root" makes the root volume unusable: ext2
tries to read 1024 bytes, but the sector size was switched to 4096.

--
	Manfred


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/