| Age | Commit message (Collapse) | Author |
|---|
|
Let's convert cdclk_state to be a proper global state. That allows
us to use the regular atomic old vs. new state accessor, hopefully
making the code less confusing.
We do have to deal with a few more error cases in case the cdclk
state duplication fails. But so be it.
v2: Fix new plane min_cdclk vs. old crtc min_cdclk check
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200121140353.25997-1-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Extract a small helper to compute the active pipes bitmask
based on the old bitmask + the crtcs in the atomic state.
I want to decouple the cdclk state entirely from the current
global state so I want to track the active pipes also inside
the (to be introduced) full cdclk state.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-17-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Now that we have the more formal global state thing let's
use if for memory bandwidth tracking. No real difference
to the current private object usage since we already
tried to avoid taking the single serializing lock needlessly.
But since we're going to roll the global state out to more
things probably a good idea to unify the approaches a bit.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-16-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Our current global state handling is pretty ad-hoc. Let's try to
make it better by imitating the standard drm core private object
approach.
The reason why we don't want to directly use the private objects
is locking; Each private object has its own lock so if we
introduce any global private objects we get serialized by that
single lock across all pipes. The global state apporoach instead
uses a read/write lock type of approach where each individual
crtc lock counts as a read lock, and grabbing all the crtc locks
allows one write access.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-15-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Move intel_atomic_state_free() next to its counterpart.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-13-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Give the cdclk init/uninit functions a _hw suffix to make
it clear they are about initializing the actual hardware.
I'll be wanting to to add a intel_cdclk_init() which is
purely initializing software structures.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-12-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
To make life less confusing let's swap() the entire cdclk state
rather than swapping some parts, copying other parts, and leaving
the rest just as is.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-11-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Use the same structure to store the cdclk state in both
intel_atomic_state and dev_priv. First step towards proper
old vs. new cdclk states.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-10-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
Move all the old vs. new state shenanigans
into intel_set_cdclk_{pre,post}_plane_update() so that the caller
doesn't need to know any of it.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-9-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
I want to have a higher level cdclk state object so let's rename
the current lower level thing to cdclk_config (because I lack
imagination).
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-8-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
intel_cdclk_needs_cd2x_update() is named rather confusingly.
We don't have to do a cd2x update, rather we are allowed to
do one (as opposed to a full PLL reprogramming with its heavy
handed modeset). So let's rename the function to
intel_cdclk_can_cd2x_update().
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-7-ville.syrjala@linux.intel.com
Reviewed-by: Imre Deak <imre.deak@intel.com>
|
|
Move the min_cdclk[] and min_voltage_level[] arrays under the
rest of the cdclk state. And while at it provide a simple
helper (intel_cdclk_clear_state()) to clear the state during
the ww_mutex backoff dance.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-6-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
Move the initial setup of state->{cdclk,min_cdclk[],min_voltage_level[]}
into intel_modeset_calc_cdclk(), and we'll move the counterparts into
intel_cdclk_swap_state(). This encapsulates the cdclk state much better.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-5-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
|
|
The dirty_pipes bitmask is now unused. Get rid of it.
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-4-ville.syrjala@linux.intel.com
|
|
The linetime watermarks really have very little in common with the
plane watermarks. It looks to be cleaner to simply track them in
the crtc_state and program them from the normal modeset/fastset
paths.
The only dark cloud comes from the fact that the register is
still supposedly single buffered. So in theory it might still
need some form of two stage programming. Note that even though
HSW/BDWhave two stage programming we never computed any special
intermediate values for the linetime watermarks, and on SKL+
we don't even have the two stage stuff plugged in since everything
else is double buffered. So let's assume it's all fine and
continue doing what we've been doing.
Actually on HSW/BDW the value should not even change without
a full modeset since it doesn't account for pfit downscaling.
Thus only fastboot might be affected. But on SKL+ the pfit
scaling factor is take into consideration so the value may
change during any fastset.
As a bonus we'll plug this thing into the state
checker/dump now.
v2: Rebase due to bigjoiner prep
v2: Only compute ips linetime for IPS capable pipes.
Bspec says the register values is ignored for other
pipes, but in fact it can't even be written so the
state checker becomes unhappy if we don't compute
it as zero.
Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-3-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
|
|
So that bridge drivers have a way to check/reject an atomic operation.
The drm_atomic_bridge_chain_check() (which is just a wrapper around
the ->atomic_check() hook) is called in place of
drm_bridge_chain_mode_fixup() (when ->atomic_check() is not implemented,
the core falls back on ->mode_fixup(), so the behavior should stay
the same for existing bridge drivers).
v10:
* Add changelog to the commit message
v8 -> v9:
* No changes
v7:
* Fix a NULL pointer dereference
v5 -> v6:
* No changes
v4:
* Add R-bs
v3:
* No changes
v2:
* Clarify the fact that ->atomic_check() is replacing ->mode_fixup()
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed by: Jernej Skrabec <jernej.skrabec@siol.net>
Tested-by: Jonas Karlman <jonas@kwiboo.se>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-6-boris.brezillon@collabora.com
|
|
This way the drm_bridge_funcs interface is consistent with the rest of
the subsystem.
The drivers implementing those hooks are patched too.
v10:
* Add changelog to the commit message
v8 -> v9:
* No changes
v7:
* Adjust things to the bridge_state changes
v6:
* Also fixed rcar-du/rcar_lvds.c same as analogix/analogix_dp_core.c
v5:
* No changes
v4:
* Rename func params into old_bridge_state
* Add Laurent's Rb
v3:
* Old state clarification moved to a separate patch
v2:
* Pass the old bridge state
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
[narmstrong: renamed state as old_bridge_state in rcar_lvds_atomic_disable]
Reviewed by: Jernej Skrabec <jernej.skrabec@siol.net>
Tested-by: Jonas Karlman <jonas@kwiboo.se>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-5-boris.brezillon@collabora.com
|
|
This is needed to pass a bridge state to all atomic hooks, if we don't
do that, the core can't duplicate/create bridge states.
v10:
* Add changelog to the commit message
v9:
* Add Neil's R-b
* Move earlier in the series
v8:
* No changes
v7:
* New patch
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-4-boris.brezillon@collabora.com
|
|
This is needed to pass a bridge state to all atomic hooks, if we don't
do that, the core can't duplicate/create bridge states.
v10:
* Add changelog to the commit message
v9:
* Add Neil's R-b
* Move earlier in the series
v8:
* No changes
v7:
* New patch
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-3-boris.brezillon@collabora.com
|
|
One of the last remaining objects to not have its atomic state.
This is being motivated by our attempt to support runtime bus-format
negotiation between elements of the bridge chain.
This patch just paves the road for such a feature by adding a new
drm_bridge_state object inheriting from drm_private_obj so we can
re-use some of the existing state initialization/tracking logic.
v10:
* Add changelog to the commit message
v9:
* Clarify the fact that the bridge->atomic_reset() and
{connector,plane,crtc,...}->reset() semantics are different
* Move the drm_atomic_private_obj_init() call back to
drm_bridge_attach()
* Check the presence of ->atomic_duplicate_state instead of
->atomic_reset in drm_atomic_add_encoder_bridges()
* Fix copy&paste errors in the atomic bridge state helpers doc
* Add A-b/R-b tags
v8:
* Move bridge state helpers out of the CONFIG_DEBUGFS section
v7:
* Move helpers, struct-defs, ... to atomic helper files to avoid the
drm -> drm_kms_helper -> drm circular dep
* Stop providing default implementation for atomic state reset,
duplicate and destroy hooks (has to do with the helper/core split)
* Drop all R-b/T-b as helpers have now be moved to other places
v6:
* Made helpers private, removed doc and moved them to satisfy dependencies
* Renamed helpers to _default_
v5:
* Re-introduced the helpers from v4
v4:
* Fix the doc
* Kill default helpers (inlined)
* Fix drm_atomic_get_bridge_state() to check for an ERR_PTR()
* Add Neil's R-b
v3:
* No changes
v2:
* Use drm_for_each_bridge_in_chain()
* Rename helpers to be more consistent with the rest of the DRM API
* Improve/fix the doc
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200128135514.108171-2-boris.brezillon@collabora.com
|
|
Let's store the normal and IPS linetime watermarks individually,
and while at it we'll pimp the register definitions as well.
v2: Deal with gvt
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200120174728.21095-2-ville.syrjala@linux.intel.com
|
|
When doing an incremental send and a file has extents shared with itself
at different file offsets, it's possible for send to emit clone operations
that will fail at the destination because the source range goes beyond the
file's current size. This happens when the file size has increased in the
send snapshot, there is a hole between the shared extents and both shared
extents are at file offsets which are greater the file's size in the
parent snapshot.
Example:
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt/sdb
$ xfs_io -f -c "pwrite -S 0xf1 0 64K" /mnt/sdb/foobar
$ btrfs subvolume snapshot -r /mnt/sdb /mnt/sdb/base
$ btrfs send -f /tmp/1.snap /mnt/sdb/base
# Create a 320K extent at file offset 512K.
$ xfs_io -c "pwrite -S 0xab 512K 64K" /mnt/sdb/foobar
$ xfs_io -c "pwrite -S 0xcd 576K 64K" /mnt/sdb/foobar
$ xfs_io -c "pwrite -S 0xef 640K 64K" /mnt/sdb/foobar
$ xfs_io -c "pwrite -S 0x64 704K 64K" /mnt/sdb/foobar
$ xfs_io -c "pwrite -S 0x73 768K 64K" /mnt/sdb/foobar
# Clone part of that 320K extent into a lower file offset (192K).
# This file offset is greater than the file's size in the parent
# snapshot (64K). Also the clone range is a bit behind the offset of
# the 320K extent so that we leave a hole between the shared extents.
$ xfs_io -c "reflink /mnt/sdb/foobar 448K 192K 192K" /mnt/sdb/foobar
$ btrfs subvolume snapshot -r /mnt/sdb /mnt/sdb/incr
$ btrfs send -p /mnt/sdb/base -f /tmp/2.snap /mnt/sdb/incr
$ mkfs.btrfs -f /dev/sdc
$ mount /dev/sdc /mnt/sdc
$ btrfs receive -f /tmp/1.snap /mnt/sdc
$ btrfs receive -f /tmp/2.snap /mnt/sdc
ERROR: failed to clone extents to foobar: Invalid argument
The problem is that after processing the extent at file offset 256K, which
refers to the first 128K of the 320K extent created by the buffered write
operations, we have 'cur_inode_next_write_offset' set to 384K, which
corresponds to the end offset of the partially shared extent (256K + 128K)
and to the current file size in the receiver. Then when we process the
extent at offset 512K, we do extent backreference iteration to figure out
if we can clone the extent from some other inode or from the same inode,
and we consider the extent at offset 256K of the same inode as a valid
source for a clone operation, which is not correct because at that point
the current file size in the receiver is 384K, which corresponds to the
end of last processed extent (at file offset 256K), so using a clone
source range from 256K to 256K + 320K is invalid because that goes past
the current size of the file (384K) - this makes the receiver get an
-EINVAL error when attempting the clone operation.
So fix this by excluding clone sources that have a range that goes beyond
the current file size in the receiver when iterating extent backreferences.
A test case for fstests follows soon.
Fixes: 11f2069c113e02 ("Btrfs: send, allow clone operations within the same file")
CC: stable@vger.kernel.org # 5.5+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
We ran into a deadlock in production with the fixup worker. The stack
traces were as follows:
Thread responsible for the writeout, waiting on the page lock
[<0>] io_schedule+0x12/0x40
[<0>] __lock_page+0x109/0x1e0
[<0>] extent_write_cache_pages+0x206/0x360
[<0>] extent_writepages+0x40/0x60
[<0>] do_writepages+0x31/0xb0
[<0>] __writeback_single_inode+0x3d/0x350
[<0>] writeback_sb_inodes+0x19d/0x3c0
[<0>] __writeback_inodes_wb+0x5d/0xb0
[<0>] wb_writeback+0x231/0x2c0
[<0>] wb_workfn+0x308/0x3c0
[<0>] process_one_work+0x1e0/0x390
[<0>] worker_thread+0x2b/0x3c0
[<0>] kthread+0x113/0x130
[<0>] ret_from_fork+0x35/0x40
[<0>] 0xffffffffffffffff
Thread of the fixup worker who is holding the page lock
[<0>] start_delalloc_inodes+0x241/0x2d0
[<0>] btrfs_start_delalloc_roots+0x179/0x230
[<0>] btrfs_alloc_data_chunk_ondemand+0x11b/0x2e0
[<0>] btrfs_check_data_free_space+0x53/0xa0
[<0>] btrfs_delalloc_reserve_space+0x20/0x70
[<0>] btrfs_writepage_fixup_worker+0x1fc/0x2a0
[<0>] normal_work_helper+0x11c/0x360
[<0>] process_one_work+0x1e0/0x390
[<0>] worker_thread+0x2b/0x3c0
[<0>] kthread+0x113/0x130
[<0>] ret_from_fork+0x35/0x40
[<0>] 0xffffffffffffffff
Thankfully the stars have to align just right to hit this. First you
have to end up in the fixup worker, which is tricky by itself (my
reproducer does DIO reads into a MMAP'ed region, so not a common
operation). Then you have to have less than a page size of free data
space and 0 unallocated space so you go down the "commit the transaction
to free up pinned space" path. This was accomplished by a random
balance that was running on the host. Then you get this deadlock.
I'm still in the process of trying to force the deadlock to happen on
demand, but I've hit other issues. I can still trigger the fixup worker
path itself so this patch has been tested in that regard, so the normal
case is fine.
Fixes: 87826df0ec36 ("btrfs: delalloc for page dirtied out-of-band in fixup worker")
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Now that we only return 0 or -EAGAIN from btrfs_writepage_cow_fixup, we
do not need this -EBUSY case.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
For COW, btrfs expects pages dirty pages to have been through a few setup
steps. This includes reserving space for the new block allocations and marking
the range in the state tree for delayed allocation.
A few places outside btrfs will dirty pages directly, especially when unmapping
mmap'd pages. In order for these to properly go through COW, we run them
through a fixup worker to wait for stable pages, and do the delalloc prep.
87826df0ec36 added a window where the dirty pages were cleaned, but pending
more action from the fixup worker. We clear_page_dirty_for_io() before
we call into writepage, so the page is no longer dirty. The commit
changed it so now we leave the page clean between unlocking it here and
the fixup worker starting at some point in the future.
During this window, page migration can jump in and relocate the page. Once our
fixup work actually starts, it finds page->mapping is NULL and we end up
freeing the page without ever writing it.
This leads to crc errors and other exciting problems, since it screws up the
whole statemachine for waiting for ordered extents. The fix here is to keep
the page dirty while we're waiting for the fixup worker to get to work.
This is accomplished by returning -EAGAIN from btrfs_writepage_cow_fixup
if we queued the page up for fixup, which will cause the writepage
function to redirty the page.
Because we now expect the page to be dirty once it gets to the fixup
worker we must adjust the error cases to call clear_page_dirty_for_io()
on the page. That is the bulk of the patch, but it is not the fix, the
fix is the -EAGAIN from btrfs_writepage_cow_fixup. We cannot separate
these two changes out because the error conditions change with the new
expectations.
Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
inc_block_group_ro does a calculation to see if we have enough room left
over if we mark this block group as read only in order to see if it's ok
to mark the block group as read only.
The problem is this calculation _only_ works for data, where our used is
always less than our total. For metadata we will overcommit, so this
will almost always fail for metadata.
Fix this by exporting btrfs_can_overcommit, and then see if we have
enough space to remove the remaining free space in the block group we
are trying to mark read only. If we do then we can mark this block
group as read only.
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
For some reason we've translated the do_chunk_alloc that goes into
btrfs_inc_block_group_ro to force in inc_block_group_ro, but these are
two different things.
force for inc_block_group_ro is used when we are forcing the block group
read only no matter what, for example when the underlying chunk is
marked read only. We need to not do the space check here as this block
group needs to be read only.
btrfs_inc_block_group_ro() has a do_chunk_alloc flag that indicates that
we need to pre-allocate a chunk before marking the block group read
only. This has nothing to do with forcing, and in fact we _always_ want
to do the space check in this case, so unconditionally pass false for
force in this case.
Then fixup inc_block_group_ro to honor force as it's expected and
documented to do.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Raviu reported that running his regular fs_trim segfaulted with the
following backtrace:
[ 237.525947] assertion failed: prev, in ../fs/btrfs/extent_io.c:1595
[ 237.525984] ------------[ cut here ]------------
[ 237.525985] kernel BUG at ../fs/btrfs/ctree.h:3117!
[ 237.525992] invalid opcode: 0000 [#1] SMP PTI
[ 237.525998] CPU: 4 PID: 4423 Comm: fstrim Tainted: G U OE 5.4.14-8-vanilla #1
[ 237.526001] Hardware name: ASUSTeK COMPUTER INC.
[ 237.526044] RIP: 0010:assfail.constprop.58+0x18/0x1a [btrfs]
[ 237.526079] Call Trace:
[ 237.526120] find_first_clear_extent_bit+0x13d/0x150 [btrfs]
[ 237.526148] btrfs_trim_fs+0x211/0x3f0 [btrfs]
[ 237.526184] btrfs_ioctl_fitrim+0x103/0x170 [btrfs]
[ 237.526219] btrfs_ioctl+0x129a/0x2ed0 [btrfs]
[ 237.526227] ? filemap_map_pages+0x190/0x3d0
[ 237.526232] ? do_filp_open+0xaf/0x110
[ 237.526238] ? _copy_to_user+0x22/0x30
[ 237.526242] ? cp_new_stat+0x150/0x180
[ 237.526247] ? do_vfs_ioctl+0xa4/0x640
[ 237.526278] ? btrfs_ioctl_get_supported_features+0x30/0x30 [btrfs]
[ 237.526283] do_vfs_ioctl+0xa4/0x640
[ 237.526288] ? __do_sys_newfstat+0x3c/0x60
[ 237.526292] ksys_ioctl+0x70/0x80
[ 237.526297] __x64_sys_ioctl+0x16/0x20
[ 237.526303] do_syscall_64+0x5a/0x1c0
[ 237.526310] entry_SYSCALL_64_after_hwframe+0x49/0xbe
That was due to btrfs_fs_device::aloc_tree being empty. Initially I
thought this wasn't possible and as a percaution have put the assert in
find_first_clear_extent_bit. Turns out this is indeed possible and could
happen when a file system with SINGLE data/metadata profile has a 2nd
device added. Until balance is run or a new chunk is allocated on this
device it will be completely empty.
In this case find_first_clear_extent_bit should return the full range
[0, -1ULL] and let the caller handle this i.e for trim the end will be
capped at the size of actual device.
Link: https://lore.kernel.org/linux-btrfs/izW2WNyvy1dEDweBICizKnd2KDwDiDyY2EYQr4YCwk7pkuIpthx-JRn65MPBde00ND6V0_Lh8mW0kZwzDiLDv25pUYWxkskWNJnVP0kgdMA=@protonmail.com/
Fixes: 45bfcfc168f8 ("btrfs: Implement find_first_clear_extent_bit")
CC: stable@vger.kernel.org # 5.2+
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
There exists a deadlock with range_cyclic that has existed forever. If
we loop around with a bio already built we could deadlock with a writer
who has the page locked that we're attempting to write but is waiting on
a page in our bio to be written out. The task traces are as follows
PID: 1329874 TASK: ffff889ebcdf3800 CPU: 33 COMMAND: "kworker/u113:5"
#0 [ffffc900297bb658] __schedule at ffffffff81a4c33f
#1 [ffffc900297bb6e0] schedule at ffffffff81a4c6e3
#2 [ffffc900297bb6f8] io_schedule at ffffffff81a4ca42
#3 [ffffc900297bb708] __lock_page at ffffffff811f145b
#4 [ffffc900297bb798] __process_pages_contig at ffffffff814bc502
#5 [ffffc900297bb8c8] lock_delalloc_pages at ffffffff814bc684
#6 [ffffc900297bb900] find_lock_delalloc_range at ffffffff814be9ff
#7 [ffffc900297bb9a0] writepage_delalloc at ffffffff814bebd0
#8 [ffffc900297bba18] __extent_writepage at ffffffff814bfbf2
#9 [ffffc900297bba98] extent_write_cache_pages at ffffffff814bffbd
PID: 2167901 TASK: ffff889dc6a59c00 CPU: 14 COMMAND:
"aio-dio-invalid"
#0 [ffffc9003b50bb18] __schedule at ffffffff81a4c33f
#1 [ffffc9003b50bba0] schedule at ffffffff81a4c6e3
#2 [ffffc9003b50bbb8] io_schedule at ffffffff81a4ca42
#3 [ffffc9003b50bbc8] wait_on_page_bit at ffffffff811f24d6
#4 [ffffc9003b50bc60] prepare_pages at ffffffff814b05a7
#5 [ffffc9003b50bcd8] btrfs_buffered_write at ffffffff814b1359
#6 [ffffc9003b50bdb0] btrfs_file_write_iter at ffffffff814b5933
#7 [ffffc9003b50be38] new_sync_write at ffffffff8128f6a8
#8 [ffffc9003b50bec8] vfs_write at ffffffff81292b9d
#9 [ffffc9003b50bf00] ksys_pwrite64 at ffffffff81293032
I used drgn to find the respective pages we were stuck on
page_entry.page 0xffffea00fbfc7500 index 8148 bit 15 pid 2167901
page_entry.page 0xffffea00f9bb7400 index 7680 bit 0 pid 1329874
As you can see the kworker is waiting for bit 0 (PG_locked) on index
7680, and aio-dio-invalid is waiting for bit 15 (PG_writeback) on index
8148. aio-dio-invalid has 7680, and the kworker epd looks like the
following
crash> struct extent_page_data ffffc900297bbbb0
struct extent_page_data {
bio = 0xffff889f747ed830,
tree = 0xffff889eed6ba448,
extent_locked = 0,
sync_io = 0
}
Probably worth mentioning as well that it waits for writeback of the
page to complete while holding a lock on it (at prepare_pages()).
Using drgn I walked the bio pages looking for page
0xffffea00fbfc7500 which is the one we're waiting for writeback on
bio = Object(prog, 'struct bio', address=0xffff889f747ed830)
for i in range(0, bio.bi_vcnt.value_()):
bv = bio.bi_io_vec[i]
if bv.bv_page.value_() == 0xffffea00fbfc7500:
print("FOUND IT")
which validated what I suspected.
The fix for this is simple, flush the epd before we loop back around to
the beginning of the file during writeout.
Fixes: b293f02e1423 ("Btrfs: Add writepages support")
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
There is a race between adding and removing elements to the tree mod log
list and rbtree that can lead to use-after-free problems.
Consider the following example that explains how/why the problems happens:
1) Task A has mod log element with sequence number 200. It currently is
the only element in the mod log list;
2) Task A calls btrfs_put_tree_mod_seq() because it no longer needs to
access the tree mod log. When it enters the function, it initializes
'min_seq' to (u64)-1. Then it acquires the lock 'tree_mod_seq_lock'
before checking if there are other elements in the mod seq list.
Since the list it empty, 'min_seq' remains set to (u64)-1. Then it
unlocks the lock 'tree_mod_seq_lock';
3) Before task A acquires the lock 'tree_mod_log_lock', task B adds
itself to the mod seq list through btrfs_get_tree_mod_seq() and gets a
sequence number of 201;
4) Some other task, name it task C, modifies a btree and because there
elements in the mod seq list, it adds a tree mod elem to the tree
mod log rbtree. That node added to the mod log rbtree is assigned
a sequence number of 202;
5) Task B, which is doing fiemap and resolving indirect back references,
calls btrfs get_old_root(), with 'time_seq' == 201, which in turn
calls tree_mod_log_search() - the search returns the mod log node
from the rbtree with sequence number 202, created by task C;
6) Task A now acquires the lock 'tree_mod_log_lock', starts iterating
the mod log rbtree and finds the node with sequence number 202. Since
202 is less than the previously computed 'min_seq', (u64)-1, it
removes the node and frees it;
7) Task B still has a pointer to the node with sequence number 202, and
it dereferences the pointer itself and through the call to
__tree_mod_log_rewind(), resulting in a use-after-free problem.
This issue can be triggered sporadically with the test case generic/561
from fstests, and it happens more frequently with a higher number of
duperemove processes. When it happens to me, it either freezes the VM or
it produces a trace like the following before crashing:
[ 1245.321140] general protection fault: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC PTI
[ 1245.321200] CPU: 1 PID: 26997 Comm: pool Not tainted 5.5.0-rc6-btrfs-next-52 #1
[ 1245.321235] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014
[ 1245.321287] RIP: 0010:rb_next+0x16/0x50
[ 1245.321307] Code: ....
[ 1245.321372] RSP: 0018:ffffa151c4d039b0 EFLAGS: 00010202
[ 1245.321388] RAX: 6b6b6b6b6b6b6b6b RBX: ffff8ae221363c80 RCX: 6b6b6b6b6b6b6b6b
[ 1245.321409] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff8ae221363c80
[ 1245.321439] RBP: ffff8ae20fcc4688 R08: 0000000000000002 R09: 0000000000000000
[ 1245.321475] R10: ffff8ae20b120910 R11: 00000000243f8bb1 R12: 0000000000000038
[ 1245.321506] R13: ffff8ae221363c80 R14: 000000000000075f R15: ffff8ae223f762b8
[ 1245.321539] FS: 00007fdee1ec7700(0000) GS:ffff8ae236c80000(0000) knlGS:0000000000000000
[ 1245.321591] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1245.321614] CR2: 00007fded4030c48 CR3: 000000021da16003 CR4: 00000000003606e0
[ 1245.321642] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1245.321668] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 1245.321706] Call Trace:
[ 1245.321798] __tree_mod_log_rewind+0xbf/0x280 [btrfs]
[ 1245.321841] btrfs_search_old_slot+0x105/0xd00 [btrfs]
[ 1245.321877] resolve_indirect_refs+0x1eb/0xc60 [btrfs]
[ 1245.321912] find_parent_nodes+0x3dc/0x11b0 [btrfs]
[ 1245.321947] btrfs_check_shared+0x115/0x1c0 [btrfs]
[ 1245.321980] ? extent_fiemap+0x59d/0x6d0 [btrfs]
[ 1245.322029] extent_fiemap+0x59d/0x6d0 [btrfs]
[ 1245.322066] do_vfs_ioctl+0x45a/0x750
[ 1245.322081] ksys_ioctl+0x70/0x80
[ 1245.322092] ? trace_hardirqs_off_thunk+0x1a/0x1c
[ 1245.322113] __x64_sys_ioctl+0x16/0x20
[ 1245.322126] do_syscall_64+0x5c/0x280
[ 1245.322139] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 1245.322155] RIP: 0033:0x7fdee3942dd7
[ 1245.322177] Code: ....
[ 1245.322258] RSP: 002b:00007fdee1ec6c88 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[ 1245.322294] RAX: ffffffffffffffda RBX: 00007fded40210d8 RCX: 00007fdee3942dd7
[ 1245.322314] RDX: 00007fded40210d8 RSI: 00000000c020660b RDI: 0000000000000004
[ 1245.322337] RBP: 0000562aa89e7510 R08: 0000000000000000 R09: 00007fdee1ec6d44
[ 1245.322369] R10: 0000000000000073 R11: 0000000000000246 R12: 00007fdee1ec6d48
[ 1245.322390] R13: 00007fdee1ec6d40 R14: 00007fded40210d0 R15: 00007fdee1ec6d50
[ 1245.322423] Modules linked in: ....
[ 1245.323443] ---[ end trace 01de1e9ec5dff3cd ]---
Fix this by ensuring that btrfs_put_tree_mod_seq() computes the minimum
sequence number and iterates the rbtree while holding the lock
'tree_mod_log_lock' in write mode. Also get rid of the 'tree_mod_seq_lock'
lock, since it is now redundant.
Fixes: bd989ba359f2ac ("Btrfs: add tree modification log functions")
Fixes: 097b8a7c9e48e2 ("Btrfs: join tree mod log code with the code holding back delayed refs")
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
|
|
Local IRQs are reset by a normal cpu reset. The initial cpu reset and
the clear cpu reset, as superset of the normal reset, both clear the
IRQs too.
Let's inject an interrupt to a vCPU before calling a reset and see if
it is gone after the reset.
We choose to inject only an emergency interrupt at this point and can
extend the test to other types of IRQs later.
Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
[minor fixups]
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Link: https://lore.kernel.org/r/20200131100205.74720-7-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
Test if the registers end up having the correct values after a normal,
initial and clear reset.
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Link: https://lore.kernel.org/r/20200131100205.74720-6-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
Add library access to more registers.
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Link: https://lore.kernel.org/r/20200131100205.74720-5-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
The architecture states that we need to reset local IRQs for all CPU
resets. Because the old reset interface did not support the normal CPU
reset we never did that on a normal reset.
Let's implement an interface for the missing normal and clear resets
and reset all local IRQs, registers and control structures as stated
in the architecture.
Userspace might already reset the registers via the vcpu run struct,
but as we need the interface for the interrupt clearing part anyway,
we implement the resets fully and don't rely on userspace to reset the
rest.
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Link: https://lore.kernel.org/r/20200131100205.74720-4-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
The code seems to be quite old and uses lots of unneeded spaces for
alignment, which doesn't really help with readability.
Let's:
* Get rid of the extra spaces
* Remove the ULs as they are not needed on 0s
* Define constants for the CR 0 and 14 initial values
* Use the sizeof of the gcr array to memset it to 0
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Link: https://lore.kernel.org/r/20200131100205.74720-3-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
The initial CPU reset clobbers the userspace fpc and the store status
ioctl clobbers the guest acrs + fpr. As these calls are only done via
ioctl (and not via vcpu_run), no CPU context is loaded, so we can (and
must) act directly on the sync regs, not on the thread context.
Cc: stable@kernel.org
Fixes: e1788bb995be ("KVM: s390: handle floating point registers in the run ioctl not in vcpu_put/load")
Fixes: 31d8b8d41a7e ("KVM: s390: handle access registers in the run ioctl not in vcpu_put/load")
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
Link: https://lore.kernel.org/r/20200131100205.74720-2-frankja@linux.ibm.com
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
There is no ENOTSUPP for userspace.
Reported-by: Julian Wiedmann <jwi@linux.ibm.com>
Fixes: 519783935451 ("KVM: s390: introduce ais mode modify function")
Fixes: 2c1a48f2e5ed ("KVM: S390: add new group for flic")
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
|
|
CONFIG_ENABLE_WARN_DEPRECATED is gone since
commit 771c035372a0 ("deprecate the '__deprecated' attribute warnings
entirely and for good").
CONFIG_IOSCHED_DEADLINE and CONFIG_IOSCHED_CFQ are gone since
commit f382fb0bcef4 ("block: remove legacy IO schedulers").
The IOSCHED_DEADLINE was replaced by MQ_IOSCHED_DEADLINE and it will be
now enabled by default (along with MQ_IOSCHED_KYBER).
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200130195223.3843-1-krzk@kernel.org
|
|
Enable more hardening options.
Note BUG_ON_DATA_CORRUPTION selects DEBUG_LIST and is essentially just
a synonym for it.
DEBUG_SG, DEBUG_NOTIFIERS, DEBUG_LIST, DEBUG_CREDENTIALS and
SCHED_STACK_END_CHECK should all be low overhead and just add a few
extra checks.
SLAB_FREELIST_RANDOM, and SLUB_DEBUG_ON will add some overhead to the
SLAB allocator, but nothing that should be meaningful for skiroot.
Unselecting SLAB_MERGE_DEFAULT causes the SLAB to use more memory, but
the skiroot kernel shouldn't be memory constrained on any of our
systems, all it does is run a small bootloader.
Disabling merging has some security/robustness benefit as it means a
user-after-free or overflow will be limited to the objects in that
slab, rather than potentially affecting objects from unrelated slabs
that have been merged.
Note also that slab merging is disabled anyway by enabling
SLUB_DEBUG_ON, because of the SLAB_NEVER_MERGE mask.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-9-mpe@ellerman.id.au
|
|
If the skiroot kernel crashes we don't want it sitting at an xmon
prompt forever. Instead it's more helpful to reboot and bring the
boot loader back up, and if the crash was transient we can then boot
successfully.
Similarly if we panic we should reboot, with a short timeout in case
someone is watching the console.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-8-mpe@ellerman.id.au
|
|
This turns on HARDENED_USERCOPY with HARDENED_USERCOPY_PAGESPAN, and
FORTIFY_SOURCE.
It also enables SECURITY_LOCKDOWN_LSM with _EARLY and
LOCK_DOWN_KERNEL_FORCE_INTEGRITY options enabled. This still allows
xmon to be used in read-only mode.
MODULE_SIG is selected by lockdown, so it is still enabled.
Because we're setting LOCK_DOWN_KERNELFORCE_INTEGRITY=y we also need
to enable KEXEC_FILE=y so that kexec continues to work.
Signed-off-by: Joel Stanley <joel@jms.id.au>
[mpe: Switch to lockdown integrity mode per oohal, enable KEXEC_FILE
as reported by jms]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-7-mpe@ellerman.id.au
|
|
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-6-mpe@ellerman.id.au
|
|
It's default n so we don't need to disable it.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-5-mpe@ellerman.id.au
|
|
Commit bdd08fff4915 ("HID: logitech: Add depends on LEDS_CLASS to
Logitech Kconfig entry") made HID_LOGITECH depend on LEDS_CLASS which
we do not enable, meaning we are not actually enabling those drivers
any more.
The Kconfig help text suggests USB HID compliant Logictech devices
will continue to work without HID_LOGITECH, so just drop it.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-4-mpe@ellerman.id.au
|
|
The HP network driver moved to staging in commit 52340b82cf1a ("hp100:
Move 100BaseVG AnyLAN driver to staging") meaning we don't need to
disable it any more in our defconfigs.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-3-mpe@ellerman.id.au
|
|
The NET_CADENCE symbol was renamed to NET_VENDOR_CADENCE, so we don't
need to disable the former, see commit 0df5f81c481e ("net: ethernet:
Add missing VENDOR to Cadence and Packet Engines symbols").
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-2-mpe@ellerman.id.au
|
|
The QLGE driver moved to staging in commit 955315b0dc8c ("qlge: Move
drivers/net/ethernet/qlogic/qlge/ to drivers/staging/qlge/"), meaning
our defconfigs that enable it have no effect as we don't enable
CONFIG_STAGING.
It sounds like the device is obsolete, so drop the driver.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Acked-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20200121043000.16212-1-mpe@ellerman.id.au
|
|
Clang warns (trimmed for brevity):
../drivers/power/avs/qcom-cpr.c:570:13: warning: variable 'reg_mask' is
used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
../drivers/power/avs/qcom-cpr.c:520:13: warning: variable 'new_uV' is
used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
Due to the fact that Clang's static analysis happens before any
optimization passes are taken into account, it cannot see that both
branches in the if statement must be taken because dir cannot be
something other than UP or DOWN due to the check at the top of this
function. Change the else if condition to else to fix this false
positive.
Fixes: bf6910abf548 ("power: avs: Add support for CPR (Core Power Reduction)")
Link: https://github.com/ClangBuiltLinux/linux/issues/840
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Currently CONFIG_QCOM_CPR=y implicitly depends on CONFIG_HAS_IOMEM=y;
consequently, on architectures without IOMEM we get the following build
error:
/usr/bin/ld: drivers/power/avs/qcom-cpr.o: in function `cpr_probe':
drivers/power/avs/qcom-cpr.c:1690: undefined reference to `devm_ioremap_resource'
Fix the build error by adding the unspecified dependency.
Signed-off-by: Brendan Higgins <brendanhiggins@google.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The supply going to the emmc/flash is named vccio6, not vccio0 and while
the code does this correctly already, the comments and error output do not.
So just change these values to the correct ones.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
