| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The only remaining nouveau_drm.master struct member that's being used is
the mutex that protects its object tree. Move that into nouveau_drm and
remove nouveau_drm.master entirely.
A pending series to remove the "ioctl" layer between DRM and NVKM also
removes the need for object handle lookups, and hence this mutex, but
it's still required for the moment.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The nouveau_cli that owns the channel is now stored in nouveau_chan, and
it has a pointer to the drm device already.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
nouveau_chan.device is always the same as nouveau_chan.cli.device,
so there's no need to store it separately.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Both of these are stored in nouveau_cli already, and also allows the
removal of some void casts.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Both of these are stored in nouveau_drm already.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The nouveau_cli pointer is only ever used to eventually access
nouveau_drm, so just store it directly.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This allocates a new nvif_mmu in nouveau_drm, and uses it for TTM
backend memory allocations instead of nouveau_drm.master.mmu,
which is removed by a later commit that removes nouveau_drm.master
entirely.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
These are some dodgy "convenience" macros for the DRM driver to peek
into NVKM state. They're still used in a few places, but don't belong
in nvif/device.h in any case.
Move them to nouveau_drv.h, and modify callers to pass a nouveau_drm
instead of an nvif_device.
v2:
- use drm->nvkm pointer for nvxx_*() macros, removing some void*
v3:
- add some explanation of the nvxx_*() macros
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
There's no good reason the ioremap() that results from nvif_object_map()
should fail, so add a check that the map succeeded, and remove the rd/wr
methods from display channel objects.
As this was the last user of rd/wr methods, the nvif plumbing is removed
at the same time.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The previous commit ensures the device is always mapped, so these
are unneeded.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The next commit removes the nvif rd/wr methods from nvif_device, which
were probably a bad idea, and mostly intended as a fallback if ioremap()
failed (or wasn't available, as was the case in some tools I once used).
The nv04 KMS driver already mapped the device, because it's mostly been
kept alive on life-support for many years and still directly bashes PRI
a lot for modesetting.
Post-nv50, I tried pretty hard to keep PRI accesses out of the DRM code,
but there's still a few random places where we do, and those were using
the rd/wr paths prior to this commit.
This allocates and maps a new nvif_device (which will replace the usage
of nouveau_drm.master.device later on), and replicates this pointer to
all other possible users.
This will be cleaned up by the end of another patch series, after it's
been made safe to do so.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
These were once used by used by userspace tools (with nvkm built as a
library), to access multiple GPUs from a single nvif_client.
The DRM code just uses the driver's default device, so remove the
arguments.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Does nothing. Remove it.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This was once used by userspace tools (with nvkm built as a library),
but is now unused.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This is not, and has never, been used for anything. Remove it.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This was once used by userspace tools (with nvkm built as a library), as
a way to select a "default device".
The DRM code doesn't need this at all as clients only have access to a
single device already, so inherit the value from its parent.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
These are remnants of code long gone. Remove them.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Make use of nouveau_cli.name instead of nvkm_client.name.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This hasn't been used in a while.
Moves io accessors from nvkm/core/os.h to nvif/os.h at the same time to
fix a compile issue that results from <nvkm/core/object.h> no longer
being included.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
These were a cludge used to prevent userspace's nvif ioctl from
accessing objects created by the kernel for the same client.
That interface was removed in a previous patch, so these are no
longer useful for anything.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The tools that used libnvkm no longer exist.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Has been unused for a while now.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
This has never really been used for anything, in part due to never
having reclocking stable enough in general to attempt to implement
dynamic clock changes based on load, etc.
To avoid having to rework its interfaces, remove it entirely.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
All callers now pass "detect=true, mmio=true, subdev_mask=~0ULL",
so remove the function arguments, and associated code.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Hasn't been needed since 2015...
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The previous commit removes the last remnants of userspace's own nvif
instance, so this isn't needed anymore to hide the abi16 objects from
userspace and we can use nouveau_cli.device instead.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
nouveau_usif.c was already stripped right back a couple of years ago,
limiting what userspace could do with it.
A follow-on series removes the nvkm side of these interfaces entirely,
in order to make it less of a nightmare to add/change internal APIs in
the future.
Unfortunately. Userspace uses some of this.
Fortunately, userspace only ever ended up using a fraction of the APIs,
so those are reimplemened here in a more direct manner, and return
-EINVAL to userspace for everything else.
v2:
- simplified struct nouveau_abi16_obj
- added a couple of comments
v3:
- comment harder
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Store a pointer to struct nouveau_cli in struct nouveau_abi16 to
avoid some dubious void casts.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
drm->master isn't really a nouveau_cli, and is mostly just used to get
at an nvif_mmu object to implement a hack around issues with the ioctl
interface to nvkm.
Later patches in this series allocate nvif_device/mmu objects in
nouveau_drm directly, removing the need for master.
A pending series to remove the "ioctl" layer between DRM and NVKM
removes the need for the above-mentioned hack entirely.
The only other member of drm->master that's needed is the nvif_client,
and is a dependency of device/mmu. So the first step is to move its
allocation out of code handling nouveau_cli init.
v2:
- modified slightly due to the addition of tegra/pci cleanup patches
v3:
- move nvif init below drm_dev_alloc() to avoid changing nouveau_name()
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
There's various different places in the drm code that get at the
nvkm_device via various creative (and not very type-safe) means.
One of those being via nvif_device.object.priv.
Another patch series is going to entirely remove the ioctl-like
interfaces beween the drm code and nvkm, and that field will no
longer exist.
This provides a safer replacement for accessing the nvkm_device,
and will used more in upcoming patches to cleanup other cases.
v2:
- fixup printk macros to not oops if used before client ctor
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
HW isn't touched anymore (aside from detection) until the first
nvif_device has been allocated, so we no longer need a separate
probe-only step before kicking efifb (etc) off the HW.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
We almost always want to cast the pointer from dev_get_drvdata() to
'struct nouveau_drm *', so just directly store that pointer instead,
simplifying callers, and fixing some clumsy naming of dev/drm_dev
variables at the same time.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Unify some more of the PCI/Tegra DRM driver init, both as a general
cleanup, and because a subsequent commit changes the pointer stored
via dev_set_drvdata(), and this allows the change to be made in one
place.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The next commit wants to be able to call fini() from an init() failure
path to remove the need to duplicate a bunch of cleanup.
Moving fini() above init() avoids the need for a forward-declaration.
Signed-off-by: Ben Skeggs <[email protected]>
Signed-off-by: Danilo Krummrich <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The minmax infrastructure is overkill for simple constants, and can
cause huge expansions because those simple constants are then used by
other things.
For example, 'pageblock_order' is a core VM constant, but because it was
implemented using 'min_t()' and all the type-checking that involves, it
actually expanded to something like 2.5kB of preprocessor noise.
And when that simple constant was then used inside other expansions:
#define pageblock_nr_pages (1UL << pageblock_order)
#define pageblock_start_pfn(pfn) ALIGN_DOWN((pfn), pageblock_nr_pages)
and we then use that inside a 'max()' macro:
case ISOLATE_SUCCESS:
update_cached = false;
last_migrated_pfn = max(cc->zone->zone_start_pfn,
pageblock_start_pfn(cc->migrate_pfn - 1));
the end result was that one statement expanding to 253kB in size.
There are probably other cases of this, but this one case certainly
stood out.
I've added 'MIN_T()' and 'MAX_T()' macros for this kind of "core simple
constant with specific type" use. These macros skip the type checking,
and as such need to be very sparingly used only for obvious cases that
have active issues like this.
Reported-by: Lorenzo Stoakes <[email protected]>
Link: https://lore.kernel.org/all/[email protected]/
Cc: David Laight <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Some newer ADSP firmware versions on X1E80100 report an extra __le32 at the
end of the battery information request payload, causing qcom_battmgr to
fail to initialize. Adjust the check to ignore the extra field in the info
payload so we can support both old and newer firmware versions.
Tested-by: Srinivas Kandagatla <[email protected]>
Signed-off-by: Stephan Gerhold <[email protected]>
Tested-by: Johan Hovold <[email protected]>
Reviewed-by: Dmitry Baryshkov <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Sebastian Reichel <[email protected]>
|
|
The driver returns -ENODEV when the firmware battmrg service hasn't
started yet, while per-se -ENODEV is fine, we usually use -EAGAIN to
tell the user to retry again later. And the power supply core uses
-EGAIN when the device isn't initialized, let's use the same return.
This notably causes an infinite spam of:
thermal thermal_zoneXX: failed to read out thermal zone (-19)
because the thermal core doesn't understand -ENODEV, but only
considers -EAGAIN as a non-fatal error.
While it didn't appear until now, commit [1] fixes thermal core
and no more ignores thermal zones returning an error at first
temperature update.
[1] 5725f40698b9 ("thermal: core: Call monitor_thermal_zone() if zone temperature is invalid")
Link: https://lore.kernel.org/all/[email protected]/
Cc: [email protected]
Fixes: 29e8142b5623 ("power: supply: Introduce Qualcomm PMIC GLINK power supply")
Signed-off-by: Neil Armstrong <[email protected]>
Tested-by: Stephan Gerhold <[email protected]>
Reviewed-by: Stephan Gerhold <[email protected]>
Link: https://lore.kernel.org/r/20240715-topic-sm8x50-upstream-fix-battmgr-temp-tz-warn-v1-1-16e842ccead7@linaro.org
Signed-off-by: Sebastian Reichel <[email protected]>
|
|
Round constant_charge_voltage writes down to the first supported lower
value, rather then rounding them up to the first supported higher value.
This fixes e.g. writing 4250000 resulting in a value of 4350000 which
might be dangerous, instead writing 4250000 will now result in a safe
4200000 value.
Fixes: 843735b788a4 ("power: axp288_charger: axp288 charger driver")
Cc: [email protected]
Signed-off-by: Hans de Goede <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Sebastian Reichel <[email protected]>
|
|
info->max_cv is in millivolts, divide the microvolt value being written
to constant_charge_voltage by 1000 *before* clamping it to info->max_cv.
Before this fix the code always tried to set constant_charge_voltage
to max_cv / 1000 = 4 millivolt, which ends up in setting it to 4.1V
which is the lowest supported value.
Fixes: 843735b788a4 ("power: axp288_charger: axp288 charger driver")
Cc: [email protected]
Signed-off-by: Hans de Goede <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Sebastian Reichel <[email protected]>
|
|
We have some very fancy min/max macros that have tons of sanity checking
to warn about mixed signedness etc.
This is all things that a sane compiler should warn about, but there are
no sane compiler interfaces for this, and '-Wsign-compare' is broken [1]
and not useful.
So then we compensate (some would say over-compensate) by doing the
checks manually with some truly horrid macro games.
And no, we can't just use __builtin_types_compatible_p(), because the
whole question of "does it make sense to compare these two values" is a
lot more complicated than that.
For example, it makes a ton of sense to compare unsigned values with
simple constants like "5", even if that is indeed a signed type. So we
have these very strange macros to try to make sensible type checking
decisions on the arguments to 'min()' and 'max()'.
But that can cause enormous code expansion if the min()/max() macros are
used with complicated expressions, and particularly if you nest these
things so that you get the first big expansion then expanded again.
The xen setup.c file ended up ballooning to over 50MB of preprocessed
noise that takes 15s to compile (obviously depending on the build host),
largely due to one single line.
So let's split that one single line to just be simpler. I think it ends
up being more legible to humans too at the same time. Now that single
file compiles in under a second.
Reported-and-reviewed-by: Lorenzo Stoakes <[email protected]>
Link: https://lore.kernel.org/all/[email protected]/
Link: https://staticthinking.wordpress.com/2023/07/25/wsign-compare-is-garbage/ [1]
Cc: David Laight <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
We would like to be able to enable vblank_disable_immediate
unconditionally, however there are a handful of cases where a small off
delay is necessary (e.g. with PSR enabled). So, we would like to be able
to adjust the vblank off delay and disable imminent values dynamically
for a given CRTC. Since, it will allow drivers to apply static screen
optimizations more quickly and consequently allow users to benefit more
so from the power savings afforded by the aforementioned optimizations,
while avoiding issues in cases where an off delay is still warranted.
In particular, the PSR case requires a small off delay of 2 frames,
otherwise display firmware isn't able to keep up with all of the
requests made to amdgpu. So, introduce drm_crtc_vblank_on_config() which
is like drm_crtc_vblank_on(), but it allows drivers to specify the
vblank CRTC configuration before enabling vblanking support for a given
CRTC.
Signed-off-by: Hamza Mahfooz <[email protected]>
Acked-by: Daniel Vetter <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Syzbot reported that a buffer state inconsistency was detected in
nilfs_btnode_create_block(), triggering a kernel bug.
It is not appropriate to treat this inconsistency as a bug; it can occur
if the argument block address (the buffer index of the newly created
block) is a virtual block number and has been reallocated due to
corruption of the bitmap used to manage its allocation state.
So, modify nilfs_btnode_create_block() and its callers to treat it as a
possible filesystem error, rather than triggering a kernel bug.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: a60be987d45d ("nilfs2: B-tree node cache")
Signed-off-by: Ryusuke Konishi <[email protected]>
Reported-by: [email protected]
Closes: https://syzkaller.appspot.com/bug?extid=89cc4f2324ed37988b60
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|
|
Post my improvement of the test in e4a4ba415419 ("selftests/mm:
va_high_addr_switch: dynamically initialize testcases to enable LPA2
testing"):
The test begins to fail on 4k and 16k pages, on non-LPA2 systems. To
reduce noise in the CI systems, let us skip the test when higher address
space is not implemented.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: e4a4ba415419 ("selftests/mm: va_high_addr_switch: dynamically initialize testcases to enable LPA2 testing")
Signed-off-by: Dev Jain <[email protected]>
Reviewed-by: Ryan Roberts <[email protected]>
Cc: Anshuman Khandual <[email protected]>
Cc: Mark Brown <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|
|
__rmqueue_pcplist()
It's expected that no page should be left in pcp_list after calling
zone_pcp_disable() in offline_pages(). Previously, it's observed that
offline_pages() gets stuck [1] due to some pages remaining in pcp_list.
Cause:
There is a race condition between drain_pages_zone() and __rmqueue_pcplist()
involving the pcp->count variable. See below scenario:
CPU0 CPU1
---------------- ---------------
spin_lock(&pcp->lock);
__rmqueue_pcplist() {
zone_pcp_disable() {
/* list is empty */
if (list_empty(list)) {
/* add pages to pcp_list */
alloced = rmqueue_bulk()
mutex_lock(&pcp_batch_high_lock)
...
__drain_all_pages() {
drain_pages_zone() {
/* read pcp->count, it's 0 here */
count = READ_ONCE(pcp->count)
/* 0 means nothing to drain */
/* update pcp->count */
pcp->count += alloced << order;
...
...
spin_unlock(&pcp->lock);
In this case, after calling zone_pcp_disable() though, there are still some
pages in pcp_list. And these pages in pcp_list are neither movable nor
isolated, offline_pages() gets stuck as a result.
Solution:
Expand the scope of the pcp->lock to also protect pcp->count in
drain_pages_zone(), to ensure no pages are left in the pcp list after
zone_pcp_disable()
[1] https://lore.kernel.org/linux-mm/[email protected]/
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 4b23a68f9536 ("mm/page_alloc: protect PCP lists with a spinlock")
Signed-off-by: Li Zhijian <[email protected]>
Reported-by: Yao Xingtao <[email protected]>
Reviewed-by: Vlastimil Babka <[email protected]>
Cc: David Hildenbrand <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|
|
Oliver Sand reported a performance regression caused by commit
98c9daf5ae6b ("mm: memcg: guard memcg1-specific members of struct
mem_cgroup_per_node"), which puts some fields of the mem_cgroup_per_node
structure under the CONFIG_MEMCG_V1 config option. Apparently it causes a
false cache sharing between lruvec and lru_zone_size members of the
structure. Fix it by adding an explicit padding after the lruvec member.
Even though the padding is not required with CONFIG_MEMCG_V1 set, it seems
like the introduced memory overhead is not significant enough to warrant
another divergence in the mem_cgroup_per_node layout, so the padding is
added unconditionally.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 98c9daf5ae6b ("mm: memcg: guard memcg1-specific members of struct mem_cgroup_per_node")
Signed-off-by: Roman Gushchin <[email protected]>
Reported-by: kernel test robot <[email protected]>
Closes: https://lore.kernel.org/oe-lkp/[email protected]
Tested-by: Oliver Sang <[email protected]>
Acked-by: Shakeel Butt <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Muchun Song <[email protected]>
Cc: Roman Gushchin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|
|
Outline and export free_reserved_page() because modules use it and it in
turn uses page_ext_{get|put} which should not be exported. The same
result could be obtained by outlining {get|put}_page_tag_ref() but that
would have higher performance impact as these functions are used in more
performance critical paths.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: dcfe378c81f7 ("lib: introduce support for page allocation tagging")
Signed-off-by: Suren Baghdasaryan <[email protected]>
Reported-by: kernel test robot <[email protected]>
Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
Suggested-by: Christoph Hellwig <[email protected]>
Suggested-by: Vlastimil Babka <[email protected]>
Acked-by: Vlastimil Babka <[email protected]>
Cc: Kees Cook <[email protected]>
Cc: Kent Overstreet <[email protected]>
Cc: Pasha Tatashin <[email protected]>
Cc: Sourav Panda <[email protected]>
Cc: <[email protected]> [6.10]
Signed-off-by: Andrew Morton <[email protected]>
|
|
The decompression code parses a huffman tree and counts the number of
symbols for a given bit length. In rare cases, there may be >= 256
symbols with a given bit length, causing the unsigned char to overflow.
This causes a decompression failure later when the code tries and fails to
find the bit length for a given symbol.
Since the maximum number of symbols is 258, use unsigned short instead.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: bc22c17e12c1 ("bzip2/lzma: library support for gzip, bzip2 and lzma decompression")
Signed-off-by: Ross Lagerwall <[email protected]>
Cc: Alain Knaff <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|
|
xarray can't support arbitrary page cache size. the largest and supported
page cache size is defined as MAX_PAGECACHE_ORDER by commit 099d90642a71
("mm/filemap: make MAX_PAGECACHE_ORDER acceptable to xarray"). However,
it's possible to have 512MB page cache in the huge memory's collapsing
path on ARM64 system whose base page size is 64KB. 512MB page cache is
breaking the limitation and a warning is raised when the xarray entry is
split as shown in the following example.
[root@dhcp-10-26-1-207 ~]# cat /proc/1/smaps | grep KernelPageSize
KernelPageSize: 64 kB
[root@dhcp-10-26-1-207 ~]# cat /tmp/test.c
:
int main(int argc, char **argv)
{
const char *filename = TEST_XFS_FILENAME;
int fd = 0;
void *buf = (void *)-1, *p;
int pgsize = getpagesize();
int ret = 0;
if (pgsize != 0x10000) {
fprintf(stdout, "System with 64KB base page size is required!\n");
return -EPERM;
}
system("echo 0 > /sys/devices/virtual/bdi/253:0/read_ahead_kb");
system("echo 1 > /proc/sys/vm/drop_caches");
/* Open the xfs file */
fd = open(filename, O_RDONLY);
assert(fd > 0);
/* Create VMA */
buf = mmap(NULL, TEST_MEM_SIZE, PROT_READ, MAP_SHARED, fd, 0);
assert(buf != (void *)-1);
fprintf(stdout, "mapped buffer at 0x%p\n", buf);
/* Populate VMA */
ret = madvise(buf, TEST_MEM_SIZE, MADV_NOHUGEPAGE);
assert(ret == 0);
ret = madvise(buf, TEST_MEM_SIZE, MADV_POPULATE_READ);
assert(ret == 0);
/* Collapse VMA */
ret = madvise(buf, TEST_MEM_SIZE, MADV_HUGEPAGE);
assert(ret == 0);
ret = madvise(buf, TEST_MEM_SIZE, MADV_COLLAPSE);
if (ret) {
fprintf(stdout, "Error %d to madvise(MADV_COLLAPSE)\n", errno);
goto out;
}
/* Split xarray entry. Write permission is needed */
munmap(buf, TEST_MEM_SIZE);
buf = (void *)-1;
close(fd);
fd = open(filename, O_RDWR);
assert(fd > 0);
fallocate(fd, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE,
TEST_MEM_SIZE - pgsize, pgsize);
out:
if (buf != (void *)-1)
munmap(buf, TEST_MEM_SIZE);
if (fd > 0)
close(fd);
return ret;
}
[root@dhcp-10-26-1-207 ~]# gcc /tmp/test.c -o /tmp/test
[root@dhcp-10-26-1-207 ~]# /tmp/test
------------[ cut here ]------------
WARNING: CPU: 25 PID: 7560 at lib/xarray.c:1025 xas_split_alloc+0xf8/0x128
Modules linked in: nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib \
nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct \
nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 \
ip_set rfkill nf_tables nfnetlink vfat fat virtio_balloon drm fuse \
xfs libcrc32c crct10dif_ce ghash_ce sha2_ce sha256_arm64 virtio_net \
sha1_ce net_failover virtio_blk virtio_console failover dimlib virtio_mmio
CPU: 25 PID: 7560 Comm: test Kdump: loaded Not tainted 6.10.0-rc7-gavin+ #9
Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20240524-1.el9 05/24/2024
pstate: 83400005 (Nzcv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
pc : xas_split_alloc+0xf8/0x128
lr : split_huge_page_to_list_to_order+0x1c4/0x780
sp : ffff8000ac32f660
x29: ffff8000ac32f660 x28: ffff0000e0969eb0 x27: ffff8000ac32f6c0
x26: 0000000000000c40 x25: ffff0000e0969eb0 x24: 000000000000000d
x23: ffff8000ac32f6c0 x22: ffffffdfc0700000 x21: 0000000000000000
x20: 0000000000000000 x19: ffffffdfc0700000 x18: 0000000000000000
x17: 0000000000000000 x16: ffffd5f3708ffc70 x15: 0000000000000000
x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
x11: ffffffffffffffc0 x10: 0000000000000040 x9 : ffffd5f3708e692c
x8 : 0000000000000003 x7 : 0000000000000000 x6 : ffff0000e0969eb8
x5 : ffffd5f37289e378 x4 : 0000000000000000 x3 : 0000000000000c40
x2 : 000000000000000d x1 : 000000000000000c x0 : 0000000000000000
Call trace:
xas_split_alloc+0xf8/0x128
split_huge_page_to_list_to_order+0x1c4/0x780
truncate_inode_partial_folio+0xdc/0x160
truncate_inode_pages_range+0x1b4/0x4a8
truncate_pagecache_range+0x84/0xa0
xfs_flush_unmap_range+0x70/0x90 [xfs]
xfs_file_fallocate+0xfc/0x4d8 [xfs]
vfs_fallocate+0x124/0x2f0
ksys_fallocate+0x4c/0xa0
__arm64_sys_fallocate+0x24/0x38
invoke_syscall.constprop.0+0x7c/0xd8
do_el0_svc+0xb4/0xd0
el0_svc+0x44/0x1d8
el0t_64_sync_handler+0x134/0x150
el0t_64_sync+0x17c/0x180
Fix it by correcting the supported page cache orders, different sets for
DAX and other files. With it corrected, 512MB page cache becomes
disallowed on all non-DAX files on ARM64 system where the base page size
is 64KB. After this patch is applied, the test program fails with error
-EINVAL returned from __thp_vma_allowable_orders() and the madvise()
system call to collapse the page caches.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 6b24ca4a1a8d ("mm: Use multi-index entries in the page cache")
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: David Hildenbrand <[email protected]>
Reviewed-by: Ryan Roberts <[email protected]>
Acked-by: Zi Yan <[email protected]>
Cc: Baolin Wang <[email protected]>
Cc: Barry Song <[email protected]>
Cc: Don Dutile <[email protected]>
Cc: Matthew Wilcox (Oracle) <[email protected]>
Cc: Peter Xu <[email protected]>
Cc: Ryan Roberts <[email protected]>
Cc: William Kucharski <[email protected]>
Cc: <[email protected]> [5.17+]
Signed-off-by: Andrew Morton <[email protected]>
|
|
machines
Yves-Alexis Perez reported commit 4ef9ad19e176 ("mm: huge_memory: don't
force huge page alignment on 32 bit") didn't work for x86_32 [1]. It is
because x86_32 uses CONFIG_X86_32 instead of CONFIG_32BIT.
!CONFIG_64BIT should cover all 32 bit machines.
[1] https://lore.kernel.org/linux-mm/CAHbLzkr1LwH3pcTgM+aGQ31ip2bKqiqEQ8=FQB+t2c3dhNKNHA@mail.gmail.com/
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 4ef9ad19e176 ("mm: huge_memory: don't force huge page alignment on 32 bit")
Signed-off-by: Yang Shi <[email protected]>
Reported-by: Yves-Alexis Perez <[email protected]>
Tested-by: Yves-Alexis Perez <[email protected]>
Acked-by: David Hildenbrand <[email protected]>
Cc: Ben Hutchings <[email protected]>
Cc: Christoph Lameter <[email protected]>
Cc: Jiri Slaby <[email protected]>
Cc: Matthew Wilcox (Oracle) <[email protected]>
Cc: Rik van Riel <[email protected]>
Cc: Salvatore Bonaccorso <[email protected]>
Cc: Suren Baghdasaryan <[email protected]>
Cc: <[email protected]> [6.8+]
Signed-off-by: Andrew Morton <[email protected]>
|
|
Commit 3bd786f76de2 ("mm: convert do_set_pte() to set_pte_range()")
replaced do_set_pte() with set_pte_range() and that introduced a
regression in the following faulting path of non-anonymous vmas which
caused the PTE for the faulting address to be marked as old instead of
young.
handle_pte_fault()
do_pte_missing()
do_fault()
do_read_fault() || do_cow_fault() || do_shared_fault()
finish_fault()
set_pte_range()
The polarity of prefault calculation is incorrect. This leads to prefault
being incorrectly set for the faulting address. The following check will
incorrectly mark the PTE old rather than young. On some architectures
this will cause a double fault to mark it young when the access is
retried.
if (prefault && arch_wants_old_prefaulted_pte())
entry = pte_mkold(entry);
On a subsequent fault on the same address, the faulting path will see a
non NULL vmf->pte and instead of reaching the do_pte_missing() path, PTE
will then be correctly marked young in handle_pte_fault() itself.
Due to this bug, performance degradation in the fault handling path will
be observed due to unnecessary double faulting.
Link: https://lkml.kernel.org/r/[email protected]
Fixes: 3bd786f76de2 ("mm: convert do_set_pte() to set_pte_range()")
Signed-off-by: Ram Tummala <[email protected]>
Reviewed-by: Yin Fengwei <[email protected]>
Cc: Alistair Popple <[email protected]>
Cc: Matthew Wilcox (Oracle) <[email protected]>
Cc: Yin Fengwei <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
|