Re: [Patch]: write_kmem above high_memory

[Tigran Aivazian]

> On Tue, 4 Jan 2000, Manfred Spraul wrote:
> > set_fs(KERNEL_DS) + get_user()/__get_user()?
> > 
> > BUT: that's dangerous, because we might read from the mmap'ed io space
> > of a hardware device.

btw, as a side note - this can be worked-around. I implemented VM_IOREMAP
flag in vm_struct *exactly* for that purpose, i.e. to have /proc/kcore
skip io-remapped areas when creating an ELF segment for each element of
vmlist. This flag can be used by /dev/kmem as well. In fact, I wanted
/proc/kcore to use vread() interface - then we don't have to worry about
exporting vmlist_lock outside mm/vmalloc.c (i.e. if Linus accepts the
vmlist_lock idea at some point in the future at all?)

Regards,
------
Tigran A. Aivazian           | http://www.sco.com
Escalations Research Group   | tel: [phone]
Santa Cruz Operation Ltd     | http://www.ocston.org/~tigran