| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The example in audio-graph-card2 binding is incomplete, uses
undocumented compatibles strings, and doesn't follow typical .dts
formatting. Rather than try to fix with what would probably be a lengthy
example, just drop the example.
Signed-off-by: Rob Herring <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Yangtao Li <[email protected]>
Acked-by: Thierry Reding <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Yangtao Li <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Use devm_platform_ioremap_resource() to simplify code.
Signed-off-by: Yangtao Li <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Use devm_platform_ioremap_resource_byname() to simplify code.
Signed-off-by: Yangtao Li <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Use devm_platform_ioremap_resource() to simplify code.
Signed-off-by: Yangtao Li <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
If CPU/Codec driver keeps its DAI node, we can directly identify actual
DAI by using snd_soc_get_dai_via_args().
This means we can use multi Component.
This patch enables multi Component support for Simple Card
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
If CPU/Codec driver keeps its DAI node, we can directly identify actual
DAI by using snd_soc_get_dai_via_args().
This means we can use multi Component.
This patch enables multi Component support on Audio Graph Card/Card2.
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
To use multi Component support, we need to check dai_args whether
Card could get DAI from args (CPU/Codec needs set dai_args on DAI driver).
If it could, we need to allocate dai_args for dlc.
This patch adds snd_soc_copy_dai_args() for it.
This is helper function for multi Component support.
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Current snd_soc_is_matching_component() checks "of_node" or "dai_args".
Thus coping "of_node" only is not enough to use CPU as Platform.
This patch adds snd_soc_dlc_use_cpu_as_platform() and help it.
This is helper function for multi Component support.
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
To enable multi Component, Card driver need to get DAI via dai_args
to identify it. This patch adds snd_soc_get_dai_via_args() for it.
This is helper function for multi Component support.
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Current ASoC Card is using dlc (snd_soc_dai_link_component) to find
target DAI / Component to be used.
Current dlc has below 3 items to identify DAI / Component
(a) name for Component
(b) of_node for Component
(c) dai_name for DAI
(a) or (b) is used to identify target Component, and (c) is used
to identify DAI.
One of the biggest issue on it today is dlc needs "name matching"
for "dai_name" (c).
It was not a big deal when we were using platform_device, because we
could specify nessesary "dai_name" via its platform_data.
But we need to find DAI name pointer from whole registered datas and/or
each related driver somehow in case of DT, because we can't specify it.
Therefore, Card driver parses DT and assumes the DAI, and find its name
pointer. How to assume is based on each Component and/or Card.
Next biggest issue is Component node (a)/(b).
Basically, Component is registered when CPU/Codec driver was
probed() (X). Here, 1 Component is possible to have some DAIs.
int xxx_probe(struct platform_device *pdev)
{
...
(X) ret = devm_snd_soc_register_component(pdev->dev,
&component_driver,
&dai_driver, dai_driver_num);
...
}
The image of each data will be like below.
One note here is "driver" is included for later explanation.
+-driver------+
|+-component-+|
|| dai0||
|| dai1||
|| ...||
|+-----------+|
+-------------+
The point here is 1 driver has 1 Component, because basically driver
calles snd_soc_register_component() (= X) once.
Here is the very basic CPU/Codec connection image.
HW image SW image
+-- Board ------------+ +-card--------------------------+
|+-----+ +------+| |+-driver------+ +-driver------+|
|| CPU | <--> |CodecA|| ||+-component-+| |+-component-+||
|+-----+ +------+| ||| dai|<=>|dai |||
+---------------------+ ||+-----------+| |+-----------+||
|+-------------+ +-------------+|
+-------------------------------+
It will be very complex if it has multi DAIs.
Here is intuitive easy to understandable HW / SW example.
HW image SW image
+-- Board ---------------+ +-card--------------------------+
|+--------+ +------+| |+-driver------+ +-driver------+|
|| CPU ch0| <--> |CodecA|| ||+-component-+| |+-component-+||
|| | +------+| ||| ch0 dai|<=>|dai |||
|| | +------+| ||| || |+-----------+||
|| ch1| <--> |CodecB|| ||| || +-------------+|
|+--------+ +------+| ||| || +-driver------+|
+------------------------+ ||| || |+-component-+||
||| ch1 dai|<=>|dai |||
||+-----------+| |+-----------+||
|+-------------+ +-------------+|
+-------------------------------+
It will be handled as multi interface as "one Card".
card0,0: CPU-ch0 - CodecA
card0,1: CPU-ch1 - CodecB
^
But, here is the HW image example which will be more complex
+-- Basic Board ---------+
|+--------+ +------+|
|| CPU ch0| <--> |CodecA||
|| ch1| <-+ +------+|
|+--------+ | |
+-------------|----------+
+-- expansion board -----+
| | +------+|
| +->|CodecB||
| +------+|
+------------------------+
We intuitively think we want to handle these as "2 Sound Cards".
card0,0: CPU-ch0 - CodecA
card1,0: CPU-ch1 - CodecB
^
But below image which we can register today doesn't allow it,
because the same Component will be connected to both Card0/1,
but it will be rejected by (Z).
+-driver------+
|+-component-+|
+-card0-------------------------+
||| || +-driver------+|
||| || |+-component-+||
||| ch0 dai|<=>|dai |||
||| || |+-----------+||
||| || +-------------+|
+-------------------------------+
|| ||
+-card1-------------------------+
||| || +-driver------+|
||| || |+-component-+||
||| ch1 dai|<=>|dai |||
||| || |+-----------+||
||| || +-------------+|
+-------------------------------+
|+-----------+|
+-------------+
static int soc_probe_component()
{
...
if (component->card) {
(Z) if (component->card != card) {
dev_err(component->dev, ...);
return -ENODEV;
}
return 0;
}
...
}
So, how about to call snd_soc_register_component() (= X) multiple times
on probe() to avoid buplicated component->card limitation, to be like
below ?
+-driver------+
+-card0-------------------------+
|| | +-driver------+|
||+-component-+| |+-component-+||
||| ch0 dai|<=>|dai |||
||+-----------+| |+-----------+||
|| | +-------------+|
+-------------------------------+
| |
+-card1-------------------------+
|| | +-driver------+|
||+-component-+| |+-component-+||
||| ch1 dai|<=>|dai |||
||+-----------+| |+-----------+||
|| | +-------------+|
+-------------------------------+
+-------------+
Yes, looks good. But unfortunately it doesn't help us for now.
Let's see soc_component_to_node() and snd_soc_is_matching_component()
static struct device_node
*soc_component_to_node(struct snd_soc_component *component)
{
...
(A) of_node = component->dev->of_node;
...
}
static int snd_soc_is_matching_component(...)
{
...
(B) if (dlc->of_node && component_of_node != dlc->of_node)
...
}
dlc checkes "of_node" to identify target component (B),
but this "of_node" came from component->dev (A) which is added
by snd_soc_register_component() (X) on probe().
This means we can have different "component->card", but have same
"component->dev" in this case.
Even though we calls snd_soc_register_component() (= X) multiple times,
all Components have same driver's dev, thus it is impossible to
identified the Component.
And if it was impossible to identify Component, it is impossible to
identify DAI on current implementation.
So, how to handle above complex HW image today is 2 patterns.
One is handles it as "1 big sound card".
The SW image is like below.
SW image
+-card--------------------------+
|+-driver------+ +-driver------+|
||+-component-+| |+-component-+||
||| ch0 dai|<=>|dai |||
||| || |+-----------+||
||| || +-------------+|
||| || +-driver------+|
||| || |+-component-+||
||| ch1 dai|<->|dai |||
||+-----------+| |+-----------+||
|+-------------+ +-------------+|
+-------------------------------+
But the problem is not intuitive.
We want to handle it as "2 Cards".
2nd pattern is like below.
SW image
+-card0-------------------------+
|+-driver------+ +-driver------+|
||+-component-+| |+-component-+||
||| ch0 dai|<=>|dai |||
||+-----------+| |+-----------+||
|+-------------+ +-------------+|
+-------------------------------+
+-card1-------------------------+
|+-driver------+ +-driver------+|
||+-component-+| |+-component-+||
||| ch1 dai|<=>|dai |||
||+-----------+| |+-----------+||
|+-------------+ +-------------+|
+-------------------------------+
It handles as "2 Cards", but CPU part needs to be probed as 2 drivers.
It is also not intuitive.
To solve this issue, we need to have multi Component support.
In current implementation, we need to identify Component first
to identify DAI, and it is using name matching to identify DAI.
But how about to be enable to directly identify DAI by unique way
instead of name matching ? In such case, we can directly identify DAI,
then it can identify Component from DAI.
For example Simple-Card / Audio-Graph-Card case, it is specifying DAI
via its node.
Simple-Card
sound-dai = <&cpu-sound>;
Audio-Graph-Card
dais = <&cpu-sound>;
If each CPU/Codec driver keeps this property when probing,
we can identify DAI directly from Card.
Being able to identify DAI directly means being able to identify its
Component as well even though Component has same dev (= B).
This patch adds new "dai_node" for it.
To keeping compatibility, it checks "dai_node" first if it has,
otherwise, use existing method (name matching).
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Current ASoC is specifying and checking DAI name.
But where it came from and how to check was ambiguous.
This patch adds snd_soc_dai_name_get() / snd_soc_dlc_dai_is_match()
and makes it clear.
Signed-off-by: Kuninori Morimoto <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Add Probe register offset for renoir and rembrandt platform to get
position update.
Signed-off-by: V sujith kumar Reddy <[email protected]>
Link: https://lore.kernel.org/r/20230713125709.418851-4-vsujithkumar.reddy@amd.corp-partner.google.com
Signed-off-by: Mark Brown <[email protected]>
|
|
This patch adds acp-probe id as a match id to support probe functionality
for amd platforms.
Signed-off-by: V sujith kumar Reddy <[email protected]>
Link: https://lore.kernel.org/r/20230713125709.418851-3-vsujithkumar.reddy@amd.corp-partner.google.com
Signed-off-by: Mark Brown <[email protected]>
|
|
This patch consist of probe client device registration,stream tag
and dma channel configuration for SOF firmware.
Signed-off-by: V sujith kumar Reddy <[email protected]>
Link: https://lore.kernel.org/r/20230713125709.418851-2-vsujithkumar.reddy@amd.corp-partner.google.com
Signed-off-by: Mark Brown <[email protected]>
|
|
Add compatible string and specific soc data to support rpmsg sound card
on i.MX93 platform.
Signed-off-by: Chancel Liu <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
Add compatible string for i.MX93 platform which supports audio
function through rpmsg channel between Cortex-A and Cortex-M core.
Signed-off-by: Chancel Liu <[email protected]>
Acked-by: Rob Herring <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The maple tree register cache is based on a much more modern data structure
than the rbtree cache and makes optimisation choices which are probably
more appropriate for modern systems than those made by the rbtree cache. In
v6.5 it has also acquired the ability to generate multi-register writes in
sync operations, bringing performance up to parity with the rbtree cache
there.
Update the wsa883x driver to use the more modern data structure.
Signed-off-by: Mark Brown <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The maple tree register cache is based on a much more modern data structure
than the rbtree cache and makes optimisation choices which are probably
more appropriate for modern systems than those made by the rbtree cache. In
v6.5 it has also acquired the ability to generate multi-register writes in
sync operations, bringing performance up to parity with the rbtree cache
there.
Update the wsa881x driver to use the more modern data structure.
Signed-off-by: Mark Brown <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The maple tree register cache is based on a much more modern data structure
than the rbtree cache and makes optimisation choices which are probably
more appropriate for modern systems than those made by the rbtree cache. In
v6.5 it has also acquired the ability to generate multi-register writes in
sync operations, bringing performance up to parity with the rbtree cache
there.
Update the wcd938x driver to use the more modern data structure.
Signed-off-by: Mark Brown <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The maple tree register cache is based on a much more modern data structure
than the rbtree cache and makes optimisation choices which are probably
more appropriate for modern systems than those made by the rbtree cache. In
v6.5 it has also acquired the ability to generate multi-register writes in
sync operations, bringing performance up to parity with the rbtree cache
there.
Update the wcd9335 driver to use the more modern data structure.
Signed-off-by: Mark Brown <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
When some of the da9063 regulators do not have corresponding DT nodes
a null pointer dereference occurs on boot because such regulators have
no init_data causing the pointers calculated in
da9063_check_xvp_constraints() to be invalid.
Do not dereference them in this case.
Fixes: b8717a80e6ee ("regulator: da9063: implement setter for voltage monitoring")
Signed-off-by: Martin Fuzzey <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The SMBus I2C buses have limits on the size of transfers they can do but
do not factor in the register length meaning we may try to do a transfer
longer than our length limit, the core will not take care of this.
Future changes will factor this out into the core but there are a number
of users that assume current behaviour so let's just do something
conservative here.
This does not take account padding bits but practically speaking these
are very rarely if ever used on I2C buses given that they generally run
slowly enough to mean there's no issue.
Cc: [email protected]
Signed-off-by: Mark Brown <[email protected]>
Reviewed-by: Xu Yilun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
When problems were noticed with the register address not being taken
into account when limiting raw transfers with I2C devices we fixed this
in the core. Unfortunately it has subsequently been realised that a lot
of buses were relying on the prior behaviour, partly due to unclear
documentation not making it obvious what was intended in the core. This
is all more involved to fix than is sensible for a fix commit so let's
just drop the original fixes, a separate commit will fix the originally
observed problem in an I2C specific way
Fixes: 3981514180c9 ("regmap: Account for register length when chunking")
Fixes: c8e796895e23 ("regmap: spi-avmm: Fix regmap_bus max_raw_write")
Signed-off-by: Mark Brown <[email protected]>
Reviewed-by: Xu Yilun <[email protected]>
Cc: [email protected]
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
This doesn't check how many bytes the simple_write_to_buffer() writes to
the buffer. The only thing that we know is that the first byte is
initialized and the last byte of the buffer is set to NUL. However
the middle bytes could be uninitialized.
There is no need to use simple_write_to_buffer(). This code does not
support partial writes but instead passes "pos = 0" as the starting
offset regardless of what the user passed as "*ppos". Just use the
copy_from_user() function and initialize the whole buffer.
Fixes: 671e0b90051e ("ASoC: SOF: Clone the trace code to ipc3-dtrace as fw_tracing implementation")
Signed-off-by: Dan Carpenter <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
In commit 2cb1e0259f50 ("ASoC: cs42l51: re-hook of_match_table
pointer"), 9 years ago, some random guy fixed the cs42l51 after it was
split into a core part and an I2C part to properly match based on a
Device Tree compatible string.
However, the fix in this commit is wrong: the MODULE_DEVICE_TABLE(of,
....) is in the core part of the driver, not the I2C part. Therefore,
automatic module loading based on module.alias, based on matching with
the DT compatible string, loads the core part of the driver, but not
the I2C part. And threfore, the i2c_driver is not registered, and the
codec is not known to the system, nor matched with a DT node with the
corresponding compatible string.
In order to fix that, we move the MODULE_DEVICE_TABLE(of, ...) into
the I2C part of the driver. The cs42l51_of_match[] array is also moved
as well, as it is not possible to have this definition in one file,
and the MODULE_DEVICE_TABLE(of, ...) invocation in another file, due
to how MODULE_DEVICE_TABLE works.
Thanks to this commit, the I2C part of the driver now properly
autoloads, and thanks to its dependency on the core part, the core
part gets autoloaded as well, resulting in a functional sound card
without having to manually load kernel modules.
Fixes: 2cb1e0259f50 ("ASoC: cs42l51: re-hook of_match_table pointer")
Cc: [email protected]
Signed-off-by: Thomas Petazzoni <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new() which already returns void. Eventually after all drivers
are converted, .remove_new() is renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/[email protected]
|
|
Get a similar baseline to my other branches, and fixes for people using
the branch.
|
|
Driver unload hits a hang during stress testing of load/unload.
stack trace snippet -
tasklet_kill at ffffffff9aabb8b2
bnxt_qplib_nq_stop_irq at ffffffffc0a805fb [bnxt_re]
bnxt_qplib_disable_nq at ffffffffc0a80c5b [bnxt_re]
bnxt_re_dev_uninit at ffffffffc0a67d15 [bnxt_re]
bnxt_re_remove_device at ffffffffc0a6af1d [bnxt_re]
tasklet_kill can hang if the tasklet is scheduled after it is disabled.
Modified the sequences to disable the interrupt first and synchronize
irq before disabling the tasklet.
Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <[email protected]>
Signed-off-by: Selvin Xavier <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
HW may generate completions that indicates QP is destroyed.
Driver should not be scheduling any more completion handlers
for this QP, after the QP is destroyed. Since CQs are active
during the QP destroy, driver may still schedule completion
handlers. This can cause a race where the destroy_cq and poll_cq
running simultaneously.
Snippet of kernel panic while doing bnxt_re driver load unload in loop.
This indicates a poll after the CQ is freed.
[77786.481636] Call Trace:
[77786.481640] <TASK>
[77786.481644] bnxt_re_poll_cq+0x14a/0x620 [bnxt_re]
[77786.481658] ? kvm_clock_read+0x14/0x30
[77786.481693] __ib_process_cq+0x57/0x190 [ib_core]
[77786.481728] ib_cq_poll_work+0x26/0x80 [ib_core]
[77786.481761] process_one_work+0x1e5/0x3f0
[77786.481768] worker_thread+0x50/0x3a0
[77786.481785] ? __pfx_worker_thread+0x10/0x10
[77786.481790] kthread+0xe2/0x110
[77786.481794] ? __pfx_kthread+0x10/0x10
[77786.481797] ret_from_fork+0x2c/0x50
To avoid this, complete all completion handlers before returning the
destroy QP. If free_cq is called soon after destroy_qp, IB stack
will cancel the CQ work before invoking the destroy_cq verb and
this will prevent any race mentioned.
Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <[email protected]>
Signed-off-by: Selvin Xavier <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
Commit 21c2fe94abb2 ("RDMA/mthca: Combine special QP struct with mthca QP")
introduced a new struct mthca_sqp which doesn't contain struct mthca_qp
any longer. Placing a pointer of this new struct into qptable leads
to crashes, because mthca_poll_one() expects a qp pointer. Fix this
by putting the correct pointer into qptable.
Fixes: 21c2fe94abb2 ("RDMA/mthca: Combine special QP struct with mthca QP")
Signed-off-by: Thomas Bogendoerfer <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
8d037973d48c ("RDMA/core: Refactor rdma_bind_addr") intoduces as regression
on irdma devices on certain tests which uses rdma CM, such as cmtime.
No connections can be established with the MAD QP experiences a fatal
error on the active side.
The cma destination address is not updated with the dst_addr when ULP
on active side calls rdma_bind_addr followed by rdma_resolve_addr.
The id_priv state is 'bound' in resolve_prepare_src and update is skipped.
This leaves the dgid passed into irdma driver to create an Address Handle
(AH) for the MAD QP at 0. The create AH descriptor as well as the ARP cache
entry is invalid and HW throws an asynchronous events as result.
[ 1207.656888] resolve_prepare_src caller: ucma_resolve_addr+0xff/0x170 [rdma_ucm] daddr=200.0.4.28 id_priv->state=7
[....]
[ 1207.680362] ice 0000:07:00.1 rocep7s0f1: caller: irdma_create_ah+0x3e/0x70 [irdma] ah_id=0 arp_idx=0 dest_ip=0.0.0.0
destMAC=00:00:64:ca:b7:52 ipvalid=1 raw=0000:0000:0000:0000:0000:ffff:0000:0000
[ 1207.682077] ice 0000:07:00.1 rocep7s0f1: abnormal ae_id = 0x401 bool qp=1 qp_id = 1, ae_src=5
[ 1207.691657] infiniband rocep7s0f1: Fatal error (1) on MAD QP (1)
Fix this by updating the CMA destination address when the ULP calls
a resolve address with the CM state already bound.
Fixes: 8d037973d48c ("RDMA/core: Refactor rdma_bind_addr")
Signed-off-by: Shiraz Saleem <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
KCSAN detects a data race on cqp_request->request_done memory location
which is accessed locklessly in irdma_handle_cqp_op while being
updated in irdma_cqp_ce_handler.
Annotate lockless intent with READ_ONCE/WRITE_ONCE to avoid any
compiler optimizations like load fusing and/or KCSAN warning.
[222808.417128] BUG: KCSAN: data-race in irdma_cqp_ce_handler [irdma] / irdma_wait_event [irdma]
[222808.417532] write to 0xffff8e44107019dc of 1 bytes by task 29658 on cpu 5:
[222808.417610] irdma_cqp_ce_handler+0x21e/0x270 [irdma]
[222808.417725] cqp_compl_worker+0x1b/0x20 [irdma]
[222808.417827] process_one_work+0x4d1/0xa40
[222808.417835] worker_thread+0x319/0x700
[222808.417842] kthread+0x180/0x1b0
[222808.417852] ret_from_fork+0x22/0x30
[222808.417918] read to 0xffff8e44107019dc of 1 bytes by task 29688 on cpu 1:
[222808.417995] irdma_wait_event+0x1e2/0x2c0 [irdma]
[222808.418099] irdma_handle_cqp_op+0xae/0x170 [irdma]
[222808.418202] irdma_cqp_cq_destroy_cmd+0x70/0x90 [irdma]
[222808.418308] irdma_puda_dele_rsrc+0x46d/0x4d0 [irdma]
[222808.418411] irdma_rt_deinit_hw+0x179/0x1d0 [irdma]
[222808.418514] irdma_ib_dealloc_device+0x11/0x40 [irdma]
[222808.418618] ib_dealloc_device+0x2a/0x120 [ib_core]
[222808.418823] __ib_unregister_device+0xde/0x100 [ib_core]
[222808.418981] ib_unregister_device+0x22/0x40 [ib_core]
[222808.419142] irdma_ib_unregister_device+0x70/0x90 [irdma]
[222808.419248] i40iw_close+0x6f/0xc0 [irdma]
[222808.419352] i40e_client_device_unregister+0x14a/0x180 [i40e]
[222808.419450] i40iw_remove+0x21/0x30 [irdma]
[222808.419554] auxiliary_bus_remove+0x31/0x50
[222808.419563] device_remove+0x69/0xb0
[222808.419572] device_release_driver_internal+0x293/0x360
[222808.419582] driver_detach+0x7c/0xf0
[222808.419592] bus_remove_driver+0x8c/0x150
[222808.419600] driver_unregister+0x45/0x70
[222808.419610] auxiliary_driver_unregister+0x16/0x30
[222808.419618] irdma_exit_module+0x18/0x1e [irdma]
[222808.419733] __do_sys_delete_module.constprop.0+0x1e2/0x310
[222808.419745] __x64_sys_delete_module+0x1b/0x30
[222808.419755] do_syscall_64+0x39/0x90
[222808.419763] entry_SYSCALL_64_after_hwframe+0x63/0xcd
[222808.419829] value changed: 0x01 -> 0x03
Fixes: 915cc7ac0f8e ("RDMA/irdma: Add miscellaneous utility definitions")
Signed-off-by: Shiraz Saleem <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
CQP completion statistics is read lockesly in irdma_wait_event and
irdma_check_cqp_progress while it can be updated in the completion
thread irdma_sc_ccq_get_cqe_info on another CPU as KCSAN reports.
Make completion statistics an atomic variable to reflect coherent updates
to it. This will also avoid load/store tearing logic bug potentially
possible by compiler optimizations.
[77346.170861] BUG: KCSAN: data-race in irdma_handle_cqp_op [irdma] / irdma_sc_ccq_get_cqe_info [irdma]
[77346.171383] write to 0xffff8a3250b108e0 of 8 bytes by task 9544 on cpu 4:
[77346.171483] irdma_sc_ccq_get_cqe_info+0x27a/0x370 [irdma]
[77346.171658] irdma_cqp_ce_handler+0x164/0x270 [irdma]
[77346.171835] cqp_compl_worker+0x1b/0x20 [irdma]
[77346.172009] process_one_work+0x4d1/0xa40
[77346.172024] worker_thread+0x319/0x700
[77346.172037] kthread+0x180/0x1b0
[77346.172054] ret_from_fork+0x22/0x30
[77346.172136] read to 0xffff8a3250b108e0 of 8 bytes by task 9838 on cpu 2:
[77346.172234] irdma_handle_cqp_op+0xf4/0x4b0 [irdma]
[77346.172413] irdma_cqp_aeq_cmd+0x75/0xa0 [irdma]
[77346.172592] irdma_create_aeq+0x390/0x45a [irdma]
[77346.172769] irdma_rt_init_hw.cold+0x212/0x85d [irdma]
[77346.172944] irdma_probe+0x54f/0x620 [irdma]
[77346.173122] auxiliary_bus_probe+0x66/0xa0
[77346.173137] really_probe+0x140/0x540
[77346.173154] __driver_probe_device+0xc7/0x220
[77346.173173] driver_probe_device+0x5f/0x140
[77346.173190] __driver_attach+0xf0/0x2c0
[77346.173208] bus_for_each_dev+0xa8/0xf0
[77346.173225] driver_attach+0x29/0x30
[77346.173240] bus_add_driver+0x29c/0x2f0
[77346.173255] driver_register+0x10f/0x1a0
[77346.173272] __auxiliary_driver_register+0xbc/0x140
[77346.173287] irdma_init_module+0x55/0x1000 [irdma]
[77346.173460] do_one_initcall+0x7d/0x410
[77346.173475] do_init_module+0x81/0x2c0
[77346.173491] load_module+0x1232/0x12c0
[77346.173506] __do_sys_finit_module+0x101/0x180
[77346.173522] __x64_sys_finit_module+0x3c/0x50
[77346.173538] do_syscall_64+0x39/0x90
[77346.173553] entry_SYSCALL_64_after_hwframe+0x63/0xcd
[77346.173634] value changed: 0x0000000000000094 -> 0x0000000000000095
Fixes: 915cc7ac0f8e ("RDMA/irdma: Add miscellaneous utility definitions")
Signed-off-by: Shiraz Saleem <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
On code inspection, there are many instances in the driver where
CEQE and AEQE fields written to by HW are read without guaranteeing
that the polarity bit has been read and checked first.
Add a read barrier to avoid reordering of loads on the CEQE/AEQE fields
prior to checking the polarity bit.
Fixes: 3f49d6842569 ("RDMA/irdma: Implement HW Admin Queue OPs")
Signed-off-by: Shiraz Saleem <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Leon Romanovsky <[email protected]>
|
|
aesp10-ppc.S and ghashp10-ppc.S are autogenerated and not tracked by
git, so they should be ignored. This is doing the same as the P8 files
in drivers/crypto/vmx/.gitignore but for the P10 files in
arch/powerpc/crypto.
Signed-off-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/[email protected]
|
|
with asm goto"
This partly reverts commit 1e688dd2a3d6759d416616ff07afc4bb836c4213.
That commit aimed at optimising the code around generation of
WARN_ON/BUG_ON but this leads to a lot of dead code erroneously
generated by GCC.
That dead code becomes a problem when we start using objtool validation
because objtool will abort validation with a warning as soon as it
detects unreachable code. This is because unreachable code might
be the indication that objtool doesn't properly decode object text.
text data bss dec hex filename
9551585 3627834 224376 13403795 cc8693 vmlinux.before
9535281 3628358 224376 13388015 cc48ef vmlinux.after
Once this change is reverted, in a standard configuration (pmac32 +
function tracer) the text is reduced by 16k which is around 1.7%
We already had problem with it when starting to use objtool on powerpc
as a replacement for recordmcount, see commit 93e3f45a2631 ("powerpc:
Fix __WARN_FLAGS() for use with Objtool")
There is also a problem with at least GCC 12, on ppc64_defconfig +
CONFIG_CC_OPTIMIZE_FOR_SIZE=y + CONFIG_DEBUG_SECTION_MISMATCH=y :
LD .tmp_vmlinux.kallsyms1
powerpc64-linux-ld: net/ipv4/tcp_input.o:(__ex_table+0xc4): undefined reference to `.L2136'
make[2]: *** [scripts/Makefile.vmlinux:36: vmlinux] Error 1
make[1]: *** [/home/chleroy/linux-powerpc/Makefile:1238: vmlinux] Error 2
Taking into account that other problems are encountered with that
'asm goto' in WARN_ON(), including build failures, keeping that
change is not worth it allthough it is primarily a compiler bug.
Revert it for now.
mpe: Retain EMIT_WARN_ENTRY as a synonym for EMIT_BUG_ENTRY to reduce
churn, as there are now nearly as many uses of EMIT_WARN_ENTRY as
EMIT_BUG_ENTRY.
Signed-off-by: Christophe Leroy <[email protected]>
Acked-by: Naveen N Rao <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/[email protected]
|
|
Commit 8ef7b9e1765a ("powerpc/pseries/vas: Close windows with DLPAR
core removal") unmaps the window paste address and issues HCALL to
close window in the hypervisor for migration or DLPAR core removal
events. So holds mmap_mutex and then mmap lock before unmap the
paste address. But if the user space issue mmap paste address at
the same time with the migration event, coproc_mmap() is called
after holding the mmap lock which can trigger deadlock when trying
to acquire mmap_mutex in coproc_mmap().
t1: mmap() call to mmap t2: Migration event
window paste address
do_mmap2() migration_store()
ksys_mmap_pgoff() pseries_migrate_partition()
vm_mmap_pgoff() vas_migration_handler()
Acquire mmap lock reconfig_close_windows()
do_mmap() lock mmap_mutex
mmap_region() Acquire mmap lock
call_mmap() //Wait for mmap lock
coproc_mmap() unmap vma
lock mmap_mutex update window status
//wait for mmap_mutex Release mmap lock
mmap vma unlock mmap_mutex
update window status
unlock mmap_mutex
...
Release mmap lock
Fix this deadlock issue by holding mmap lock first before mmap_mutex
in reconfig_close_windows().
Fixes: 8ef7b9e1765a ("powerpc/pseries/vas: Close windows with DLPAR core removal")
Signed-off-by: Haren Myneni <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/[email protected]
|
|
Most of the protocol modules for the pata_parport driver are missing a
module description, causing warnings such as:
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/ata/pata_parport/aten.o
when compiling with W=1. Add the missing MODULE_DESCRIPTION()
definitions to avoid these warnings. While at it, also add the missing
MODULE_AUTHOR() definitions.
Signed-off-by: Damien Le Moal <[email protected]>
|
|
|
|
Pull xtensa fixes from Max Filippov:
- fix interaction between unaligned exception handler and load/store
exception handler
- fix parsing ISS network interface specification string
- add comment about etherdev freeing to ISS network driver
* tag 'xtensa-20230716' of https://github.com/jcmvbkbc/linux-xtensa:
xtensa: fix unaligned and load/store configuration interaction
xtensa: ISS: fix call to split_if_spec
xtensa: ISS: add comment about etherdev freeing
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fix from Borislav Petkov:
- Fix a lockdep warning when the event given is the first one, no event
group exists yet but the code still goes and iterates over event
siblings
* tag 'perf_urgent_for_v6.5_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86: Fix lockdep warning in for_each_sibling_event() on SPR
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull objtool fixes from Borislav Petkov:
- Mark copy_iovec_from_user() __noclone in order to prevent gcc from
doing an inter-procedural optimization and confuse objtool
- Initialize struct elf fully to avoid build failures
* tag 'objtool_urgent_for_v6.5_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
iov_iter: Mark copy_iovec_from_user() noclone
objtool: initialize all of struct elf
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fixes from Borislav Petkov:
- Remove a cgroup from under a polling process properly
- Fix the idle sibling selection
* tag 'sched_urgent_for_v6.5_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/psi: use kernfs polling functions for PSI trigger polling
sched/fair: Use recent_used_cpu to test p->cpus_ptr
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl
Pull pin control fixes from Linus Walleij:
"I'm mostly on vacation but what would vacation be without a few
critical fixes so people can use their gaming laptops when hiding away
from the sun (or rain)?
- Fix a really annoying interrupt storm in the AMD driver affecting
Asus TUF gaming notebooks
- Fix device tree parsing in the Renesas driver"
* tag 'pinctrl-v6.5-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
pinctrl: amd: Unify debounce handling into amd_pinconf_set()
pinctrl: amd: Drop pull up select configuration
pinctrl: amd: Use amd_pinconf_set() for all config options
pinctrl: amd: Only use special debounce behavior for GPIO 0
pinctrl: renesas: rzg2l: Handle non-unique subnode names
pinctrl: renesas: rzv2m: Handle non-unique subnode names
|
|
Pull smb client fixes from Steve French:
- Two reconnect fixes: important fix to address inFlight count to leak
(which can leak credits), and fix for better handling a deleted share
- DFS fix
- SMB1 cleanup fix
- deferred close fix
* tag '6.5-rc1-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6:
cifs: fix mid leak during reconnection after timeout threshold
cifs: is_network_name_deleted should return a bool
smb: client: fix missed ses refcounting
smb: client: Fix -Wstringop-overflow issues
cifs: if deferred close is disabled then close files immediately
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
- Fix Speculation_Store_Bypass reporting in /proc/self/status on
Power10
- Fix HPT with 4K pages since recent changes by implementing pmd_same()
- Fix 64-bit native_hpte_remove() to be irq-safe
Thanks to Aneesh Kumar K.V, Nageswara R Sastry, and Russell Currey.
* tag 'powerpc-6.5-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/mm/book3s64/hash/4k: Add pmd_same callback for 4K page size
powerpc/64e: Fix obtool warnings in exceptions-64e.S
powerpc/security: Fix Speculation_Store_Bypass reporting on Power10
powerpc/64s: Fix native_hpte_remove() to be irq-safe
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
Pull hardening fixes from Kees Cook:
- Remove LTO-only suffixes from promoted global function symbols
(Yonghong Song)
- Remove unused .text..refcount section from vmlinux.lds.h (Petr Pavlu)
- Add missing __always_inline to sparc __arch_xchg() (Arnd Bergmann)
- Claim maintainership of string routines
* tag 'hardening-v6.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
sparc: mark __arch_xchg() as __always_inline
MAINTAINERS: Foolishly claim maintainership of string routines
kallsyms: strip LTO-only suffixes from promoted global functions
vmlinux.lds.h: Remove a reference to no longer used sections .text..refcount
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull probe fixes from Masami Hiramatsu:
- fprobe: Add a comment why fprobe will be skipped if another kprobe is
running in fprobe_kprobe_handler().
- probe-events: Fix some issues related to fetch-arguments:
- Fix double counting of the string length for user-string and
symstr. This will require longer buffer in the array case.
- Fix not to count error code (minus value) for the total used
length in array argument. This makes the total used length
shorter.
- Fix to update dynamic used data size counter only if fetcharg uses
the dynamic size data. This may mis-count the used dynamic data
size and corrupt data.
- Revert "tracing: Add "(fault)" name injection to kernel probes"
because that did not work correctly with a bug, and we agreed the
current '(fault)' output (instead of '"(fault)"' like a string)
explains what happened more clearly.
- Fix to record 0-length (means fault access) data_loc data in fetch
function itself, instead of store_trace_args(). If we record an
array of string, this will fix to save fault access data on each
entry of the array correctly.
* tag 'probes-fixes-v6.5-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing/probes: Fix to record 0-length data_loc in fetch_store_string*() if fails
Revert "tracing: Add "(fault)" name injection to kernel probes"
tracing/probes: Fix to update dynamic data counter if fetcharg uses it
tracing/probes: Fix not to count error code to total length
tracing/probes: Fix to avoid double count of the string length on the array
fprobes: Add a comment why fprobe_kprobe_handler exits if kprobe is running
|