,
no additional flushes are required in the various affinity setting
routines.
This patch has been validated with (unreleased) network hardware which
uses MSI-X.
Revised with input from Eric Biederman.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.2
On Fri, 2007-03-30 at 11:56 -0700, Mitch Williams wrote:
> This patch fixes a kernel bug which is triggered when using the
> irqbalance daemon with MSI-X hardware.
>
Grrr. Evolution cut-n-sometimes-paste feature bit me. Will resend with
a proper subject line.
-Mitch
-
To unsubscribe
,
no additional flushes are required in the various affinity setting
routines.
This patch has been validated with (unreleased) network hardware which
uses MSI-X.
Revised with input from Eric Biederman.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.21-
This patch fixes a kernel bug which is triggered when using the
irqbalance daemon with MSI-X hardware.
Because both MSI-X interrupt messages and MSI-X table writes are posted,
it's possible for them to cross while in-flight. This results in
interrupts being received long after the kernel thinks t
with input from Eric Biederman.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.21-rc5-clean/arch/i386/kernel/io_apic.c
linux-2.6.21-rc5/arch/i386/kernel/io_apic.c
--- linux-2.6.21-rc5-clean/arch/i386/kernel/io_apic.c 2007-03-28
10:05:22.0 -0700
2.6.21-rc5.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.21-rc4-clean/arch/i386/kernel/io_apic.c
linux-2.6.21-rc4/arch/i386/kernel/io_apic.c
--- linux-2.6.21-rc4-clean/arch/i386/kernel/io_apic.c 2007-03-19
16:16:30.0 -0700
+++ linux-2.6.21-rc4/
so a read-flush is not necessary for
mask/unmask operations.
This patch has been validated with (unreleased) network hardware which
uses MSI-X.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.20.3-clean/arch/i386/kernel/io_apic.c
linux-2.6.20.3/arch/i386/
so a read-flush is not necessary for
mask/unmask operations.
This patch has been validated with (unreleased) network hardware which
uses MSI-X.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.21-rc4-clean/arch/i386/kernel/io_apic.c
linux-2.6.21-rc4/arch/i386/
This patch generated from 2.6.11-rc1.
Now with tabs.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-2.6.11-clean/fs/sysfs/file.c 2004-12-24 13:33:50.0 -0800
+++ linux-2.6.11/fs/s
On Tue, 1 Feb 2005, Greg KH wrote:
>
> Ick, no. Pulled back out, as it doesn't even compile :(
>
Ick, indeed. Yet another victim of my sneaky attempt to whack one large
patch up into three small patches. Let's just forget this patch ever
existed, since it doesn't add any new functionality an
This patch returns an error code if the caller attempts to open a sysfs
file for appending.
Generated from 2.6.11-rc2.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-2.6.11-clean/fs
key maybe?), I can live without write buffering
right now.
But at the very least, we still need to handle this failure case. I've
tested the following patch and it does resolve the issue. However, it now
limits the size of sysfs writes to the size of the c library's buffer.
----
On Tue, 25 Jan 2005, Josh Boyer wrote:
>
> Could you drop the braces? Coding style thing.
>
Yeah, I'll drop and resubmit. I've got so many different things going on
with these patches, I'm getting confused. I know Greg asked for this same
change a while back but it got lost.
Resubmittal will
This patch causes an error return if the user attempts to seek on a sysfs
file.
The patch was generated from 2.6.11-rc1.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-2.6.11-clean/fs
This patch causes an error to be returned if the caller attempts to open a
sysfs file in append mode.
This patch applies cleanly to 2.6.11-rc1.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -urpN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
---
On Sat, 22 Jan 2005, Greg KH wrote:
>
> On Fri, Jan 21, 2005 at 02:52:29PM -0800, Mitch Williams wrote:
> > This patch buffers writes to sysfs files and flushes them to the
> kobject
> > owner when the file is closed.
>
> Why? This breaks the way things work tod
On Sat, 22 Jan 2005, Greg KH wrote:
>
> On Fri, Jan 21, 2005 at 02:49:39PM -0800, Mitch Williams wrote:
> > This patch causes sysfs to return errors if the caller attempts to
> append
> > to or seek on a sysfs file.
>
> And what happens to it today if you do eithe
My apologies -- I appear to have sent the patches out in reverse order.
Please apply patch 3 before the other two.
This is the first time I've used our automated tools to make small patches
out of big ones, but I think I have it figured out now.
Thanks for your patience.
-Mitch Williams
This patch reverses the semantics of the read fill flag, getting rid of an
extra assignment at allocation time.
Generated from 2.6.11-rc1.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-
This patch buffers writes to sysfs files and flushes them to the kobject
owner when the file is closed.
Generated from 2.6.11-rc1.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-2.6.11-cl
This patch causes sysfs to return errors if the caller attempts to append
to or seek on a sysfs file.
Generated from 2.6.11-rc1.
Signed-off-by: Mitch Williams <[EMAIL PROTECTED]>
diff -uprN -X dontdiff linux-2.6.11-clean/fs/sysfs/file.c
linux-2.6.11/fs/sysfs/file.c
--- linux-2.6.11-cl
21 matches
Mail list logo
