length of the last page as
PAGE_SIZE-iobuf->offset or you'll run over in your S/G.
unmap_kiobuf() to unmap the kiobuf and unpin the pages
If you do a lot of DMA I/O with large buffers you'll need to apply
the 22vmfix patch in the same directory or you'll run out memory
and th
; "\n9:"
> :"=m" (__dummy_lock(&kernel_flag)),
> "=m" (current->lock_depth));
> #endif
> }
>
> Can someone explain what is behind lock_kernel() and
> these changes (#if 1) in 2.4 ?
>
> Is t
overhead and it still works correctly as long as the application
and driver are well behaved.
And I don't agree that we should ever try to do kiobuf() things in
the read/write general case. I think that mapping files into
kernel buffers and copying the pages to/from the page cache
would suffi
here's no way to do this without mmap().
With kiobufs() still being somewhat iffy, it's unfortunate there's
still no good interface to map kernel buffers to user space either...
Something to work on for 2.5 I guess.
--
Eric Lowe
Software Engineer, Systran Corporation
[EMAIL PROTECTED]
ocols.. My experience has been that systems that do a lot
of page pinning in UNIX for I/O don't gain much in performance
anyway, but it _really_ puts a lot of stress on the system compared
to doing it the right way.
--
Eric Lowe
Software Engineer, Systran Corporation
[EMAIL PROTECTED]
.2.17 (with bugfixes), anyone who
needs it e-mail me and I'll send it along. This provides a *nearly*
compatible solution to 2.4, and it's much cleaner than driver-specific
hacks.
--
Eric Lowe
Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe from this list: se
95 is now at the point where if anything kernel side is
> broken it must be fairly obscure or a little used driver
I've been using 2.95.2 to build PPC kernels and it seems to work
OK, but I _know_ that 2.91.66 works with both 2.2 and 2.4 kernels so it should
probably be the recommended one
to go
(e.g. start = buf+begin ; count = end-start)
-if you have to update begin/end non-atomically and check them
only once, use a spinlock, it's the least expensive lock
--
Eric Lowe
Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe from this list: send the
part of the raw I/O patch for 2.2. Replay if you
need a recent one against 2.2.17, I have it.
--
Eric Lowe
Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
y driver or kernel
module.. The size is small enough it probably wouldn't trip
up and cause a panic blocking waiting for more memory, but that's
because you've been lucky so far..
--
Eric Lowe
FibreChannel Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe fr
d be nice to see it fixed though, I'm trying
to help debug the new VM on that machine that just so happens to use
that chipset, and it's hard to do that when it doesn't boot at all. :)
--
Eric Lowe
FibreChannel Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubsc
future versions of the kernel, the buffer
cache will be used less and less, as things like directory entries are
moved into the page cache.
You probably want to start looking at the various kernel guides
and read up on the difference between the page cache and buffer
cache, that's the best p
r lost dirty bits on x86 SMP -and-
Thread: [PATCH] 2.2.18pre5 version of pte dirty bit psmp race atch
--
Eric Lowe
FibreChannel Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Please read the FAQ at http://www.tux.org/lkml/
u think pageable page tables are
something we want to tackle in 2.5.x? 4MB page mappings on x86
could be cool too, as an option...
--
Eric Lowe
FibreChannel Software Engineer, Systran Corporation
[EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel&qu
14 matches
Mail list logo
