| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
This now should reflect "all things kernel" and thus could not only
contain patches but other things. If we have so many patches that it
makes sense to namespace them further, we can still use kernel/patches
for that purpose which is way better than "kpatches".
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
I've also simplified the Nix expression a bit using the exact names from
the upstream URL instead of using antiquotations, because it's easier to
copy & paste them.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
| |
We can now drop the kernel 4.4 patch and use the upstream version,
because kernel 4.4 has been released since quite a while.
The names for the patches changed a bit on the BFQ scheduler page but
they represent the same commits nevertheless.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
The __GFP_WAIT macro has been renamed in torvalds/linux@71baba4 and
torvalds/linux@d0164ad introduced the gfpflags_allow_blocking() helper,
which we're going to use to avoid false positives.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
| |
I have no idea why the upstream patch went from v7r9 to v7r8 now,
essentially reverting the changes described in a8b6eeb. But after
checking the patch against kernel 4.4, there is quite a lot more to fix
up in r9 rather than r8, so I'm crossing fingers and go with this
instead.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
| |
We need to add our own patch and don't want to clutter up the kpatches
directory.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 7f9f88e90b8ab41a97a86fa4ff8a501e0e0eea27.
The reason why I'm reverting this is because while it indeed fixes the
volume mapping, it doesn't fix my main problem that it takes a few
seconds to go from 100% volume to the desired volume level that's
currently set.
So instead of applying this patch and maintaining it until it may
eventually hit mainline, it's better to debug the original issue rather
than applying a patch that _may_ fix an unrelated issue.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
| |
It's already in version 4.3-rc5, but the following patch seems to be
more correct:
http://mailman.alsa-project.org/pipermail/alsa-devel/2015-August/096516.html
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From the changelog:
* IMPROVEMENT: Changed BFQ to use the blkio controller instead of its
own controller. BFQ now registers itself as a policy to
the blkio controller and implements its hierarchical
scheduling support using data structures that already
exist in blk-cgroup. The bfqio controller's code is
completely removed.
* CODE IMPROVEMENTS: Applied all suggestions from Tejun Heo,
received on the last submission to LKML:
https://lkml.org/lkml/2014/5/27/314
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
| |
This was in since the initial commit (6c224e4) of the bfqsched Nix
expression and I didn't notice it for all those upgrades.
Basically, this is just a no-op, because nowadays the sed expression
doesn't rewrite anything anymore, but it might be harmful if it should
match something in a future version of the bfqsched patch set.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
| |
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changes for v7r7:
* BUGFIX: Prevent the OOM queue from being involved in the queue
cooperation mechanism. In fact, since the requests temporarily
redirected to the OOM queue could be redirected again to dedicated
queues at any time, the state needed to correctly handle merging with
the OOM queue would be quite complex and expensive to maintain.
Besides, in such a critical condition as an out of memory, the
benefits of queue merging may be little relevant, or even negligible.
* IMPROVEMENT: Let the OOM queue be initialized only once. Previously,
the OOM queue was reinitialized, at each request enqueue, with the
parameters related to the process that issued that request.
Depending on the parameters of the processes doing I/O, this could
easily cause the OOM queue to be moved continuously across service
trees, or even across groups. It also caused the parameters of the
OOM queue to be continuously reset in any case.
* CODE IMPROVEMENT. Performed some minor code cleanups, and added some
BUG_ON()s that, if the weight of an entity becomes inconsistent,
should better help understand why.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
| |
Although we're building for 3.18, the patch should still apply.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
|
|
|
|
|
|
| |
Includes a patch for 3.17-rc1 from:
https://groups.google.com/forum/#!topic/bfq-iosched/piEzoNMGmk8
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|
|
Those packages aren't actually overrides and thus it doesn't make sense
to categorize them into "overrides".
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
|