There are two issues with KVM_EXIT_DEBUG on AMD, whose root cause is the
different handling of DR6 on intercepted #DB exceptions on Intel and AMD.
On Intel, #DB exceptions transmit the DR6 value via the exit qualification
field of the VMCS, and the exit qualification only contains the description
of the precise event that caused a vmexit.
On AMD, instead the DR6 field of the VMCB is filled in as if the #DB exception
was to be injected into the guest. This has two effects when guest debugging
is in use:
* the guest DR6 is clobbered
* the kvm_run->debug.arch.dr6 field can accumulate more debug events, rather
than just the last one that happened (the testcase in the next patch covers
this issue).
This patch fixes both issues by emulating, so to speak, the Intel behavior
on AMD processors. The important observation is that (after the previous
patches) the VMCB value of DR6 is only ever observable from the guest is
KVM_DEBUGREG_WONT_EXIT is set. Therefore we can actually set vmcb->save.dr6
to any value we want as long as KVM_DEBUGREG_WONT_EXIT is clear, which it
will be if guest debugging is enabled.
Therefore it is possible to enter the guest with an all-zero DR6,
reconstruct the #DB payload from the DR6 we get at exit time, and let
kvm_deliver_exception_payload move the newly set bits into vcpu->arch.dr6.
Some extra bits may be included in the payload if KVM_DEBUGREG_WONT_EXIT
is set, but this is harmless.
This may not be the most optimized way to deal with this, but it is
simple and, being confined within SVM code, it gets rid of the set_dr6
callback and kvm_update_dr6.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_x86_ops.set_dr6 is only ever called with vcpu->arch.dr6 as the
second argument. Ensure that the VMCB value is synchronized to
vcpu->arch.dr6 on #DB (both "normal" and nested) and nested vmentry, so
that the current value of DR6 is always available in vcpu->arch.dr6.
The get_dr6 callback can just access vcpu->arch.dr6 and becomes redundant.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
As a means for a small code consolidation, but primarily to start
thinking more carefully about internal-vs-external linkage, pull the
pair of i915_sw_fence_await_dma_fence() calls into a common routine.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200508092933.738-2-chris@chris-wilson.co.uk
While we ordinarily do not skip submit-fences due to the accompanying
hook that we want to callback on execution, a submit-fence on the same
timeline is meaningless.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200508092933.738-1-chris@chris-wilson.co.uk
DMA transfer could be completed, but CPU (which handles DMA interrupt)
may get too busy and can't handle the interrupt in a timely manner,
despite of DMA IRQ being raised. In this case the DMA state needs to
synchronized before terminating DMA transfer in order not to miss the
DMA transfer completion.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Boot CPU0 always handle I2C interrupt and under some rare circumstances
(like running KASAN + NFS root) it may stuck in uninterruptible state for
a significant time. In this case we will get timeout if I2C transfer is
running on a sibling CPU, despite of IRQ being raised. In order to handle
this rare condition, the IRQ status needs to be checked after completion
timeout.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
One of the I2C controllers on Tegra SoCs is typically connected to a
system PMIC, which provides controls for critical power supplies for
most platforms.
Some drivers, such as PCI, need to disable these regulators during a
very late stage during suspend and resume them at a very early stage
during resume.
To support these use-cases, keep interrupts disabled during suspend/
resume.
Suggested-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Depending on the board design, the I2C controllers found on Tegra SoCs
may require pinmuxing in order to function. This is done as part of the
driver's runtime suspend/resume operations. However, the PM core does
not allow devices to go into runtime suspend during system sleep to
avoid potential races with the suspend/resume of their parents.
As a result of this, when Tegra SoCs resume from system suspend, their
I2C controllers may have lost the pinmux state in hardware, whereas the
pinctrl subsystem is not aware of this. To fix this, make sure that if
the I2C controller is not runtime suspended, the runtime suspend code is
still executed in order to disable the module clock (which we don't need
to be enabled during sleep) and set the pinmux to the idle state.
Conversely, make sure that the I2C controller is properly resumed when
waking up from sleep so that pinmux settings are properly restored.
This fixes a bug seen with DDC transactions to an HDMI monitor timing
out when resuming from system suspend.
Signed-off-by: Thierry Reding <treding@nvidia.com>
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20200507192228.GA16355@embeddedor
Signed-off-by: Mark Brown <broonie@kernel.org>
When CONFIG_SND_SOC_SOF_BAYTRAIL is not set, gcc warns:
sound/soc/sof/intel/byt.c:85:41: warning: ‘cht_debugfs’ defined but not used [-Wunused-const-variable=]
static const struct snd_sof_debugfs_map cht_debugfs[] = {
^~~~~~~~~~~
Move the variable inside #ifdef
Reported-by: Hulk Robot <hulkci@huawei.com>
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20200507072735.16588-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Make an additional note on DRM format modifiers for x and y tiling. These
format modifiers are defined for BDW+ platforms and therefore definition
is not valid for older gens. This is due to address swizzling for tiled
surfaces is no longer used. For newer platforms main memory controller has
a more effective address swizzling algorithm.
v2: Rephrase comment (Daniel)
Signed-off-by: Mika Kahola <mika.kahola@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20200506120827.12250-1-mika.kahola@intel.com
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
In the current market, the most used bridge chip on the Loongson platform
are RS780E and LS7A, the RS780E bridge chip is already supported by the
mainline kernel.
If use the default implementation of __phys_to_dma() and __dma_to_phys()
in dma-direct.h when CONFIG_ARCH_HAS_PHYS_TO_DMA is not set, it works
well used with LS7A on the Loongson single-way and multi-way platform,
and also works well used with RS780E on the Loongson single-way platform,
but the DMA address will be wrong on the non-node0 used with RS780E on
the Loongson multi-way platform.
Just as the description in the code comment, the devices get node id from
40 bit of HyperTransport bus, so we extract 2 bit node id (bit 44~45) from
48 bit address space of Loongson CPU and embed it into HyperTransport bus
(bit 37-38), this operation can be done only at the software level used
with RS780E on the Loongson multi-way platform, because it has no hardware
function to translate address of node id, this is a hardware compatibility
problem.
Device
|
| DMA address
|
Host Bridge
|
| HT bus address (40 bit)
|
CPU
|
| physical address (48 bit)
|
RAM
The LS7A has dma_node_id_offset field in the DMA route config register,
the hardware can use the dma_node_id_offset to translate address of
node id automatically, so we can get correct address when just use the
dma_pfn_offset field in struct device.
For the above reasons, in order to maintain downward compatibility
to support the RS780E bridge chip, it is better to use the platform
dependent implementation of __phys_to_dma() and __dma_to_phys().
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Commit d339cd02b8 ("MIPS: Move unaligned load/store helpers to
inst.h") causes a lot of build failures because macros in asm.h conflict
with various subsystems. Some of these conflictions has been fixed (such
as LONG, PANIC and PRINT) by adjusting asm.h, but some of them is nearly
impossible to fix (such as PTR and END). The only reason of including
asm.h in inst.h is that we need the PTR macro which is used by unaligned
load/store helpers. So in this patch we define a new PTR_STR macro and
use it to replace STR(PTR), then we can stop including asm.h to avoid
various build failures.
Fixes: d339cd02b8 ("MIPS: Move unaligned load/store helpers to inst.h")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Link: https://lore.kernel.org/r/20200507185230.GA14229@embeddedor
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Jan Kara <jack@suse.cz>
First, it should be noted that the CQE timeout (60 seconds) is substantial
so a CQE request that times out is really stuck, and the race between
timeout and completion is extremely unlikely. Nevertheless this patch
fixes an issue with it.
Commit ad73d6fead ("mmc: complete requests from ->timeout")
preserved the existing functionality, to complete the request.
However that had only been necessary because the block layer
timeout handler had been marking the request to prevent it from being
completed normally. That restriction was removed at the same time, the
result being that a request that has gone will have been completed anyway.
That is, the completion was unnecessary.
At the time, the unnecessary completion was harmless because the block
layer would ignore it, although that changed in kernel v5.0.
Note for stable, this patch will not apply cleanly without patch "mmc:
core: Fix recursive locking issue in CQE recovery path"
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Fixes: ad73d6fead ("mmc: complete requests from ->timeout")
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200508062227.23144-1-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In previous commit, the sequence of syt offset and the number of data
blocks per packet is calculated for pool in AMDTP domain structure in
advance of processing outgoing packets.
This commit uses the sequence for outgoing packet processing to obsolete
per-stream processing of the sequence.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-11-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In current implementation, sequence of syt offset and the number of data
blocks is generated when packets for outgoing stream are going to be
queued.
This commit generates and pools the sequence independently of the
processing of outgoing packets for future extension.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-10-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
For future extension, storage is required to store packet sequence in
incoming AMDTP stream to recover media clock for outgoing AMDTP stream.
This commit adds the storage to AMDTP domain for this purpose. The
packet sequence is represented by 'struct seq_desc' which has two
members; syt_offset and the number of data blocks. The size of storage
is decided according to the size of packet queue.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-9-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When calculating the number of data blocks per packet, some states are
stored in AMDTP stream structure. This is inconvenient when reuse the
calculation from non-stream structure.
This commit applies refactoring to helper function for the calculation
so that the function doesn't touch AMDTP stream structure.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-8-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When calculating syt offset, some states are stored in AMDTP stream
structure. This is inconvenient when reuse the calculation from
non-stream structure.
This commit applies refactoring to helper function for the calculation
so that the function doesn't touch AMDTP stream structure.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-7-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In current implementation for outgoing AMDTP packet, the value of syt
field in CIP header is computed when calculating syt offset. For
future extension, it's convenient to split the computation and
calculation.
This commit splits them.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-6-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Although the parameter for packet queue and IRQ timing is calculated when
AMDTP stream starts, the calculated parameters are the same between
streams in AMDTP domain.
This commit moves the calculation and decide the parameters when AMDTP
domain starts.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-5-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In current implementation, AMDTP domain structure and AMDTP stream
structure has one way of reference from the former to the latter. For
future extension, bidirectional reference is needed.
This commit adds a member into stream structure to refer to domain
structure to which the stream belongs.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-4-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In descriptor of isochronous context in 1394 OHCI, the field of second
has 3 bit, thus the maximum value is 8. The value is used for correct
cycle calculation.
This commit replaces hard-coded value with macro to obsolete magic
number.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-3-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Although the value of FDF is used just for outgoing stream, the assignment
to union member is done for both directions of stream. At present this
causes no issue because the value of same position is reassigned later for
opposite stream. However, it's better to add if statement.
Fixes: d3d10a4a1b ("ALSA: firewire-lib: use union for directional parameters")
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200508043635.349339-2-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Acked-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20200507185245.GA14270@embeddedor
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Suspending failed because there's no mode if the CRTC is being
disabled. Early-out in this case. This fixes runtime PM for ast.
v3:
* fixed commit message
v2:
* added Tested-by/Reported-by tags
* added Fixes tags and CC (Sam)
* improved comment
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reported-by: Cary Garrett <cogarre@gmail.com>
Tested-by: Cary Garrett <cogarre@gmail.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Fixes: b48e1b6ffd ("drm/ast: Add CRTC helpers for atomic modesetting")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: <stable@vger.kernel.org> # v5.6+
Link: https://patchwork.freedesktop.org/patch/msgid/20200507090640.21561-1-tzimmermann@suse.de
There are several "flavors" of HW that have the same HW type, but
can be told apart after reading a certain perph register. This
is easy to do in runtime, but more complicated to do when looking
at the logs offline.
To make it easier to tell apart these "flavors" when looking at
the dumped dbg info, add these bits to the HW type, allowing
simple differentiation.
Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.330ea11d17ae.Ie59b25430a308090b15112ac6deedf4fbf487ff1@changeid
We don't really expect fragmented RBs, and don't seem to be seeing
them in practice since that would've caused a crash. Nevertheless,
we should be expecting the hardware to send them.
Parse the flag indicating a fragmented buffer, but then discard it
and any fragments thereof, at least for now. We need to do more
work in the higher layers to properly deal with this, since we may
not get "normal" firmware notifications that are fragmented, only
RX, and then we need to put it back together and add the necessary
API to report a chain of things to the higher layers, this doesn't
fit into the struct iwl_rx_cmd_buffer today.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.e78a59f70b1d.Ica656a98a4e4220d73edc97600edd680cbc97241@changeid
We can currently end up transmitting on an unallocated queue, if
the allocation fails. Stop doing that, by simply not transmitting.
We don't have any better strategy here, unfortunately, but the
previous commits make that much less likely.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.dcf1801f25ef.I6d71e13ea042765800f2ee41401b8eb282527c34@changeid
When we have 256 block-ack support, we may need to be very fast
to provide a lot of frames to the hardware to transmit, but that
cannot be guaranteed. Use a longer queue size to have more time,
and the next possible queue size is 1024 since it must be a power
of two.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.851866c7e4c4.I13fa678929431f1694fd202c1da40aa476ab70fe@changeid
Since the recent patch in this area, we no longer allocate 64k
for a single queue, but only 1k, which still means a full page.
Use a DMA pool to reduce this further, since we will have a lot
of queues in a typical system that can share pages.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.6e84c79aea30.Ie9a417132812d110ec1cc87852f101477c01cfcb@changeid
We currently attempt to allocate queues that are 512 entries long,
but that requires 32 KiB memory, which may not be available, at
least not contiguously. If we fail to allocate, attempt to use a
smaller queue all the way down to 16 entries (which fit into a
single page).
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.c8548d7cc08a.I5059c410e628726cbce98d6311b690c632d00f97@changeid
The hardware needs a byte-count table with the size of each frame
on the queue to build A-MPDUs, but:
* newer generation no longer have the duplicated space at the end,
they can deal with the wrap properly - and we don't even fill
the dup anyway
* we have a maximum queue size of 512 right now and don't use the
theoretical hardware maximum of 65536.
Together, this reduces the byte count table DMA allocation from
64KiB (65536*2 + 64*2 rounded up) to 1 KiB (though that might be
rounded up to a full 4 KiB page by the allocator, not sure it can
share the allocations.)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200425130140.c263b787b5ab.I059507a9760b1ce1d45d84dcaa91629a5cfb58e0@changeid
Used for debugging what FW API we are using to understand misalignment
with API changes.
The output looks like this as a yaml format
fw_api_ver:
0x0001:
name: MVM_ALIVE
cmd_ver: 99
notif_ver: 4
0x0108:
name: PHY_CONTEXT_CMD
cmd_ver: 2
notif_ver: 0
...
Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200424194456.18bf540ab8e0.I6217488f1740f0e6accd0cecd09dfd46bad88426@changeid
