| Age | Commit message (Collapse) | Author |
|---|
|
We capture error state not only when the GPU hangs but also on
other situations as in interrupt errors and in situations where
we can kick things forward without GPU reset. There will be log
entry on most of these cases. But as error state capture might be
only thing we have, if dmesg was not captured. Or as in GEN4 case,
interrupt error can trigger error state capture without log entry,
the exact reason why capture was made is hard to decipher.
v2: Split out the the error code stuff to separate patch (Ben)
References: https://bugs.freedesktop.org/show_bug.cgi?id=74193
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
commit 011cf577b2531dfbd2254bd9ec147ad71471abaf
Author: Ben Widawsky <benjamin.widawsky@intel.com>
Date: Tue Feb 4 12:18:55 2014 +0000
drm/i915: Generate a hang error code
added error code debug into dmesg. Store this also
with error state to make matching dmesg logs and error
states easier.
As we need to have full ring state for error code generation,
do full capture always, print hang message into log and then
decide if we need to keep the error state.
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
After finding the guilty batch and request, we can use it to find the
process that submitted the batch and then add the culprit into the error
state.
This is a slightly different approach from Ben's in that instead of
adding the extra information into the struct i915_hw_context, we use the
information already captured in struct drm_file which is then referenced
from the request.
v2: Also capture the workaround buffer for gen2, so that we can compare
its contents against the intended batch for the active request.
v3: Rebase (Mika)
v4: Check for null context (Chris)
checkpatch warnings fixed
Link: http://lists.freedesktop.org/archives/intel-gfx/2013-August/032280.html
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> (v2)
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> (v4)
Acked-by: Ben Widawsky <ben@bwidawsk.net>
Cc: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
In the past, it was possible to have multiple batches per request due to
a stray signal or ENOMEM. As a result we had to scan each active object
(filtered by those having the COMMAND domain) for the one that contained
the ACTHD pointer. This was then made more complicated by the
introduction of ppgtt, whereby ACTHD then pointed into the address space
of the context and so also needed to be taken into account.
This is a fairly robust approach (though the implementation is a little
fragile and depends upon the per-generation setup, registers and
parameters). However, due to the requirements for hangstats, we needed a
robust method for associating batches with a particular request and
having that we can rely upon it for finding the associated batch object
for error capture.
If the batch buffer tracking is not robust enough, that should become
apparent quite quickly through an erroneous error capture. That should
also help to make sure that the runtime reporting to userspace is
robust. It also means that we then report the oldest incomplete batch on
each ring, which can be useful for determining the state of userspace at
the time of a hang.
v2: Use i915_gem_find_active_request (Mika)
v3: remove check for ring->get_seqno, split long lines (Ben)
v4: check that context is available (Chris)
checkpatch warnings fixed
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> (v1)
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> (v3)
Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net> (v3)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
In place of true activity counting, we walk the list of vma associated
with an object managing each on the vm's active/inactive list everytime
we call move-to-inactive. This depends upon the vma->mm_list being
cleared after unbinding, or else we run into difficulty when tracking
the object in multiple vm's - we see a use-after free and corruption of
the mm_list.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
It occured to me that when we're trying to wake up both render
and media wells on VLV, we might end up calling the low level
force_wake_get/put two times even though one call would be
enough. Make that happen by figuring out which wells really
need to be woken up based on the forcewake counts.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by:Deepak S <deepak.s@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
VLV is the only platform where we increment/decrement the forcewake
count around register access. Drop the inc/dec on VLV to make the
forcewake code a bit more unified.
The inc/dec are not necessary since we hold the uncore lock around
the whole operation.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Use the render/media specific forcewake counts to properly restore the
forcewake status after a GPU reset on VLV.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Deepak S <deepak.s@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
After a hang and failed reset, we cannot use the GPU to execute the page
flip instructions. Instead we can force a synchronous mmio flip. (Later,
we can reduce the synchronicity of the mmio flip by moving some of the
delays off to a worker, like the current page flip code; see vblank
tasks.)
References: https://bugs.freedesktop.org/show_bug.cgi?id=72631
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
I could swear this was already happening in the current code...
Also, put the reads and writes in a generic place, so we don't forget
it again when we add runtime PM support to new platforms.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Just to be sure...
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Because we shouldn't be runtime suspended when forcewake is supposed
to be enabled.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
[danvet: Update commit message - no WARN expected since the bugfix for
issues hit with this assert is already in. And resolve conflicts with
the change from worker to timer for the delayed fw release.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Since the addition of dev_priv->mm.busy, there's no more need for
dev_priv->pc8.gpu_idle, so kill it.
Notice that when you remove gpu_idle, hsw_package_c8_gpu_idle and
hsw_package_c8_gpu_busy become identical to hsw_enable_package_c8 and
hsw_disable_package_c8, so just use them.
Also, when we boot the machine, dev_priv->mm.busy initially considers
the machine as idle. This is opposed to dev_priv->pc8.gpu_idle, which
considered it busy. So dev_priv->pc8.disable_count has to be
initalized to 1 now.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
These are places where we read (not write) registers while we're
runtime suspended.
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Otherwise we'll read registers that return 0xffffffff, trigger some
WARNs, think CRT is actually connected (because certain bits are 1),
and fail the drm-resources-equal testcase!
Tested on a SNB machine with runtime PM support (which is not upstream
yet, but is already on my public tree at freedesktop.org, and will
hopefully eventually become upstream).
Testcase: igt/pm_pc8/drm-resources-equal
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
When we call gen6_gt_force_wake_put we don't actually put force_wake,
we just schedule gen6_force_wake_work through mod_delayed_work, and
that will eventually release force_wake.
The problem is that we call intel_runtime_pm_put directly at
gen6_gt_force_wake_put, so most of the times we put our runtime PM
reference before the delayed work happens, so we may runtime suspend
while force_wake is still supposed to be enabled if the graphics
autosuspend_delay_ms is too small.
Now the nice thing about the current code is that after it triggers
the delayed work function it gets a refcount, and it only triggers the
delayed work function if refcount is zero. This guarantees that when
we schedule the funciton, it will run before we try to schedule it
again, which simplifies the problem and allows for the current
solution to work properly (hopefully!).
v2: - Keep the VLV refcounts balanced (Jesse)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Because intel_mark_idle still touches some registers: it needs the
machine to be awake. If you set both the autosuspend and PC8 delays to
zero, you can get a "Device suspended" WARN when gen6_rps_idle touches
registers.
This is not easy to reproduce, but happens once in a while when
running pm_pc8.
Testcase: igt/pm_pc8
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
If we've explicitly stopped the rings for testing purposes, don't ban
the default context. Fixes kms_flip hang tests.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Useful for bug reports.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
MIPI Block #52 which provides configuration details for the MIPI panel
including dphy settings as per panel and tcon specs
Block #53 gives information on panel enable sequences
v2: Address review comemnts from Jani
- Move panel ids from intel_dsi.h to intel_bios.h
- bdb_mipi_config structure improvements for cleaner code
- Adding units for the pps delays, all in ms
- change data structure to be more cleaner and simple
v3: Corrected the unit for pps delays as 100us
Signed-off-by: Shobhit Kumar <shobhit.kumar@intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
We don't want to suffer scheduling delay when turning off the GPU after
waking it up to touch registers. Ideally, we only want to keep the GPU
awake for the register access sequence, with a single forcewake dance on
the first access and release immediately after the last. We set a timer
on the first access so that we only dance once and on the next scheduler
tick, we drop the forcewake again.
This moves the cleanup routine from the common i915 workqueue to a timer
func so that we don't anger powertop, and drop the forcewake again
quicker.
v2: Enable the deferred force_wake_put for regular register reads as
well.
v3: Beautification and make sure we disable forcewake when shutting
down.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This got lost when we shuffled around our internal branch and
GEN7_FEATURES macro. There were no HW changes to support FBC, so we just
need to set the flag.
v2: Don't allow FBC for any pipe but A on platforms with DDI. (Paulo)
Cc: Daisy Sun <daisy.sun@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
We currently call intel_mark_idle() too often, as we do so as a
side-effect of processing the request queue. However, we the calls to
intel_mark_idle() are expected to be paired with a call to
intel_mark_busy() (or else we try to idle the hardware by accessing
registers that are already disabled). Make the idle/busy tracking
explicit to prevent the multiple calls.
v2: We can drop some of the complexity in __i915_add_request() as
queue_delayed_work() already behaves as we want (not requeuing the item
if it is already in the queue) and mark_busy/mark_idle imply that the
idle task is inactive.
v3: We do still need to cancel the pending idle task so that it is sent
again after the current busy load completes (not in the middle of it).
Reported-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Sometimes generic driver code gets forcewake explicitly by
gen6_gt_force_wake_get(), which check forcewake_count before accessing
hardware. However the register access with gen8_write function access
low level hw accessors directly, ignoring the forcewake_count. This
leads to nested forcewake get from hardware, in ring init and possibly
elsewhere, causing forcewake ack clear errors and/or hangs.
Fix this by checking the forcewake count also in gen8_write
v2: Read side doesn't care about shadowed registers,
Remove __needs_put funkiness from gen8_write. (Ville)
Improved commit message.
References: https://bugs.freedesktop.org/show_bug.cgi?id=74007
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Ben Widawsky <benjamin.widawsky@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This way we can reuse the check on other platforms too. Also factor out
a version of the function that doesn't check if the power is on, we'll
need to call this from within the power domain framework.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
On VLV at least the display IRQ register access and functionality
depends on its power well to be on, so move the power domain HW init
before we install the IRQs.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
The power domains framework is internal to the i915 driver, so pass
drm_i915_private instead of drm_device to its functions.
Also remove a dangling intel_set_power_well() declaration.
No functional change.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Both Ville and QA rather immediately complained that with the new
initial_config logic from Jesse not all outputs get enabled. Since the
fbdev emulation pretty much tries to always enable as many outputs as
possible (it even has hotplug handling and all that) fall back if more
outputs could have been enabled.
v2: Fix up my confusion about what enabled means - it's passed from
the fbdev helper, we need to check for a non-zero connector->encoder
link. Spotted by Ville.
v3: Add some debug output as requested by Jesse for debugging fallback
issues.
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75552
Tested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
It started as a simple check whether anything is lit up, but now is't
used to driver the general fallback logic to the default output
configuration selector in the helper library. So rename it for more
clarity.
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This should be impossible due to the wait for outstanding flips that the
caller is meant to perform prior to updating the scanout base. Paranoia
tells me to check anyway.
References: https://bugs.freedesktop.org/show_bug.cgi?id=75502
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This reverts commit 116f2b6da868dec7539103574d0421cd6221e931.
This optimization causes widespread corruption in games, and even in
glxgears, on my ivb:gt1. The corruption appears like z-fighting of
overlapping polygons in the HiZ buffer.
The observation ties in very closely with the description of the
optimization disabled by default on IVB:
"The Hierarchical Z RAW Stall Optimization allows non-overlapping
polygons in the same 8x4 pixel/sample area to be processed without
stalling waiting for the earlier ones to write to Hierarchical Z
buffer."
No reason is given for why it is disabled by default, usually for such
optimizations it is that it is incomplete. However, there is no
indication whether this a gt1 only issue either. Before considering
reenabling this optimization, I would first suggest reproducing the
corruption in piglit.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75623
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chia-I Wu <olv@lunarg.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
To silence locking complaints. This was a rebase failure on my part in
commit fa9fa083d0606cb323f6105c17702460ea0a6780
Author: Jesse Barnes <jbarnes@virtuousgeek.org>
Date: Tue Feb 11 15:28:56 2014 -0800
drm/i915: read out hw state earlier v2
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
In the move over to use BIOS connector configs, we lost the ability to
force a specific set of connectors on or off. Try to remedy that by
dropping back to the old behavior if we detect a hard coded connector
config.
v2: don't deref connector state for disabled connectors (Jesse)
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
In commit e4e0c058a19c41150d12ad2d3023b3cf09c5de67
Author: Eugeni Dodonov <eugeni.dodonov@intel.com>
Date: Wed Feb 8 12:53:50 2012 -0800
drm/i915: gen7: Implement an L3 caching workaround.
the L3 cache aging was disabled. This was part of a shotgun response
to a number of GPU hang bugs, but there appears to be no documentation
to suggest that disabling the L3 cache age was ever required (to prevent
the GPU hangs).
Restoring the L3 cache age is a minor performance win of around 2%
on IVB:GT2. (Note that this value seems to be consistent across a number
of tests and so appears to be above the usual noise.)
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
V2: edit the commit message to contain more info
The W/A spreadsheet says this is still required, but the b-spec says
it's not for BYT-T. So the documentation is not clear. However,
our experience with the other SKUs of BYT-I/M on Android and Linux
suggests that setting this bit actually causes GPU hang for certain
OGL benchmark applications.
Removing this bit completely resolves the GPU hangs.
Signed-off-by: Sinclair Yeh <sinclair.yeh@intel.com>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
With the original PPGTT implementation if the number of PDPs was not a
power of two, the number of pages for the page tables would end up being
rounded up. The code actually had a bug here afaict, but this is a
theoretical bug as I don't believe this can actually occur with the
current code/HW..
With the rework of the page table allocations, there is no longer a
distinction between number of page table pages, and number of page
directory entries. To avoid confusion, kill the redundant (and newer)
struct member.
Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
Simply to match the GEN8 style of PPGTT initialization, split up the
allocations and mappings. Unlike GEN8, we skip a separate dma_addr_t
allocation function, as it is much simpler pre-gen8.
With this code it would be easy to make a more general PPGTT
initialization function with per GEN alloc/map/etc. or use a common
helper, similar to the ringbuffer code. I don't see a benefit to doing
this just yet, but who knows...
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This cleanup is similar to the GEN8 cleanup (though less necessary).
Having everything split will make cleaning the initialization path error
paths easier to understand.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
I keep meaning to do this... by now almost the entire file has been
written by an Intel employee (including Daniel post-2010).
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
This reverts commit 3a2ffb65eec6dbda2fd8151894f51c18b42c8d41.
Now that the code is fixed to use smaller allocations, it should be safe
to let the full GGTT be used on BDW.
The testcase for this is anything which uses more than half of the GTT,
thus eclipsing the old limit.
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
The previous allocation mechanism would get 2 contiguous allocations,
one for the page directories, and one for the page tables. As each page
table is 1 page, and there are 512 of these per page directory, this
goes to 2MB. An unfriendly request at best. Worse still, our HW now
supports 4 page directories, and a 2MB allocation is not allowed.
In order to fix this, this patch attempts to split up each page table
allocation into a single, discrete allocation. There is nothing really
fancy about the patch itself, it just has to manage an extra pointer
indirection, and have a fancier bit of logic to free up the pages.
To accommodate some of the added complexity, two new helpers are
introduced to allocate, and free the page table pages.
NOTE: I really wanted to split the way we do allocations, and the way in
which we identify the page table/page directory being used. I found
splitting this functionality up to be too unwieldy. I apologize in
advance to the reviewer. I'd recommend looking at the result, rather
than the diff.
v2/NOTE2: This patch predated commit:
6f1cc993518462ccf039e195fabd47e7aa5bfd13
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue Dec 31 15:50:31 2013 +0000
drm/i915: Avoid dereference past end of page arr
It fixed the same issue as that patch, but because of the limbo state of
PPGTT, Chris patch was merged instead. The excess churn is a result of
my using my original patch, which has my preferred naming. Primarily
act_* is changed to which_*, but it's mostly the same otherwise. I've
kept the convention Chris used for the pte wrap (I had something
slightly different, and broken - but fixable)
v3: Rename which_p[..]e to drop which_ (Chris)
Remove BUG_ON in inner loop (Chris)
Redo the pde/pdpe wrap logic (Chris)
v4: s/1MB/2MB in commit message (Imre)
Plug leaking gen8_pt_pages in both the error path, as well as general
free case (Imre)
v5: Rename leftover "which_" variables (Imre)
Add the pde = 0 wrap that was missed from v3 (Imre)
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
[danvet: Squash in fixup from Ben.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
i) by the time DM core calls the postsuspend hook the dm_noflush flag
has been cleared. So the old thin_postsuspend did nothing. We need to
use the presuspend hook instead.
ii) There was a race between bios leaving DM core and arriving in the
deferred queue.
thin_presuspend now sets a 'requeue' flag causing all bios destined for
that thin to be requeued back to DM core. Then it requeues all held IO,
and all IO on the deferred queue (destined for that thin). Finally
postsuspend clears the 'requeue' flag.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
The spin lock in requeue_io() was held for too long, allowing deadlock.
Don't worry, due to other issues addressed in the following "dm thin:
fix noflush suspend IO queueing" commit, this code was never called.
Fix this by taking the spin lock for a much shorter period of time.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
Ideally a thin pool would never run out of data space; the low water
mark would trigger userland to extend the pool before we completely run
out of space. However, many small random IOs to unprovisioned space can
consume data space at an alarming rate. Adjust your low water mark if
you're frequently seeing "out-of-data-space" mode.
Before this fix, if data space ran out the pool would be put in
PM_READ_ONLY mode which also aborted the pool's current metadata
transaction (data loss for any changes in the transaction). This had a
side-effect of needlessly compromising data consistency. And retry of
queued unserviceable bios, once the data pool was resized, could
initiate changes to potentially inconsistent pool metadata.
Now when the pool's data space is exhausted transition to a new pool
mode (PM_OUT_OF_DATA_SPACE) that allows metadata to be changed but data
may not be allocated. This allows users to remove thin volumes or
discard data to recover data space.
The pool is no longer put in PM_READ_ONLY mode in response to the pool
running out of data space. And PM_READ_ONLY mode no longer aborts the
pool's current metadata transaction. Also, set_pool_mode() will now
notify userspace when the pool mode is changed.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
|
|
If a thin metadata operation fails the current transaction will abort,
whereby causing potential for IO layers up the stack (e.g. filesystems)
to have data loss. As such, set THIN_METADATA_NEEDS_CHECK_FLAG in the
thin metadata's superblock which:
1) requires the user verify the thin metadata is consistent (e.g. use
thin_check, etc)
2) suggests the user verify the thin data is consistent (e.g. use fsck)
The only way to clear the superblock's THIN_METADATA_NEEDS_CHECK_FLAG is
to run thin_repair.
On metadata operation failure: abort current metadata transaction, set
pool in read-only mode, and now set the needs_check flag.
As part of this change, constraints are introduced or relaxed:
* don't allow a pool to transition to write mode if needs_check is set
* don't allow data or metadata space to be resized if needs_check is set
* if a thin pool's metadata space is exhausted: the kernel will now
force the user to take the pool offline for repair before the kernel
will allow the metadata space to be extended.
Also, update Documentation to include information about when the thin
provisioning target commits metadata, how it handles metadata failures
and running out of space.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Joe Thornber <ejt@redhat.com>
|
|
If the open stateid could not be recovered, or the file locks were lost,
then we should fail the truncate() operation altogether.
Reported-by: Andy Adamson <andros@netapp.com>
Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
|
Signed-off-by: Andy Adamson <andros@netapp.com>
Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
|
In commit 5521abfdcf4d6 (NFSv4: Resend the READ/WRITE RPC call
if a stateid change causes an error), we overloaded the return value of
nfs4_select_rw_stateid() to cause it to return -EWOULDBLOCK if an RPC
call is outstanding that would cause the NFSv4 lock or open stateid
to change.
That is all redundant when we actually copy the stateid used in the
read/write RPC call that failed, and check that against the current
stateid. It is doubly so, when we consider that in the NFSv4.1 case,
we also set the stateid's seqid to the special value '0', which means
'match the current valid stateid'.
Reported-by: Andy Adamson <andros@netapp.com>
Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
|
When nfs4_set_rw_stateid() can fails by returning EIO to indicate that
the stateid is completely invalid, then it makes no sense to have it
trigger a retry of the READ or WRITE operation. Instead, we should just
have it fall through and attempt a recovery.
This fixes an infinite loop in which the client keeps replaying the same
bad stateid back to the server.
Reported-by: Andy Adamson <andros@netapp.com>
Link: http://lkml.kernel.org/r/1393954269-3974-1-git-send-email-andros@netapp.com
Cc: stable@vger.kernel.org # 3.10+
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
