linux-stable.git/drivers/bus, branch v6.1.78

bus: mhi: ep: Do not allocate event ring element on stack

2024-02-01T00:17:09+00:00

[ Upstream commit 987fdb5a43a66764808371b54e6047834170d565 ]

It is possible that the host controller driver would use DMA framework to
write the event ring element. So avoid allocating event ring element on the
stack as DMA cannot work on vmalloc memory.

Cc: stable@vger.kernel.org
Fixes: 961aeb689224 ("bus: mhi: ep: Add support for sending events to the host")
Link: https://lore.kernel.org/r/20230901073502.69385-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Sasha Levin

bus: mhi: host: Add spinlock to protect WP access when queueing TREs

2024-02-01T00:16:59+00:00

commit b89b6a863dd53bc70d8e52d50f9cfaef8ef5e9c9 upstream.

Protect WP accesses such that multiple threads queueing buffers for
incoming data do not race.

Meanwhile, if CONFIG_TRACE_IRQFLAGS is enabled, irq will be enabled once
__local_bh_enable_ip is called as part of write_unlock_bh. Hence, let's
take irqsave lock after TRE is generated to avoid running write_unlock_bh
when irqsave lock is held.

Cc: stable@vger.kernel.org
Fixes: 189ff97cca53 ("bus: mhi: core: Add support for data transfer")
Signed-off-by: Bhaumik Bhatt 
Signed-off-by: Qiang Yu 
Reviewed-by: Jeffrey Hugo 
Tested-by: Jeffrey Hugo 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/1702276972-41296-2-git-send-email-quic_qianyu@quicinc.com
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Greg Kroah-Hartman

bus: mhi: host: Drop chan lock before queuing buffers

2024-02-01T00:16:59+00:00

commit 01bd694ac2f682fb8017e16148b928482bc8fa4b upstream.

Ensure read and write locks for the channel are not taken in succession by
dropping the read lock from parse_xfer_event() such that a callback given
to client can potentially queue buffers and acquire the write lock in that
process. Any queueing of buffers should be done without channel read lock
acquired as it can result in multiple locks and a soft lockup.

Cc:  # 5.7
Fixes: 1d3173a3bae7 ("bus: mhi: core: Add support for processing events from client device")
Signed-off-by: Qiang Yu 
Reviewed-by: Jeffrey Hugo 
Tested-by: Jeffrey Hugo 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/1702276972-41296-3-git-send-email-quic_qianyu@quicinc.com
[mani: added fixes tag and cc'ed stable]
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Greg Kroah-Hartman

bus: mhi: host: Add alignment check for event ring read pointer

2024-02-01T00:16:59+00:00

commit eff9704f5332a13b08fbdbe0f84059c9e7051d5f upstream.

Though we do check the event ring read pointer by "is_valid_ring_ptr"
to make sure it is in the buffer range, but there is another risk the
pointer may be not aligned.  Since we are expecting event ring elements
are 128 bits(struct mhi_ring_element) aligned, an unaligned read pointer
could lead to multiple issues like DoS or ring buffer memory corruption.

So add a alignment check for event ring read pointer.

Fixes: ec32332df764 ("bus: mhi: core: Sanity check values from remote device before use")
cc: stable@vger.kernel.org
Signed-off-by: Krishna chaitanya chundru 
Reviewed-by: Jeffrey Hugo 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/20231031-alignment_check-v2-1-1441db7c5efd@quicinc.com
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Greg Kroah-Hartman

bus: ti-sysc: Flush posted write only after srst_udelay

2024-01-01T12:39:07+00:00

commit f71f6ff8c1f682a1cae4e8d7bdeed9d7f76b8f75 upstream.

Commit 34539b442b3b ("bus: ti-sysc: Flush posted write on enable before
reset") caused a regression reproducable on omap4 duovero where the ISS
target module can produce interconnect errors on boot. Turns out the
registers are not accessible until after a delay for devices needing
a ti,sysc-delay-us value.

Let's fix this by flushing the posted write only after the reset delay.
We do flushing also for ti,sysc-delay-us using devices as that should
trigger an interconnect error if the delay is not properly configured.

Let's also add some comments while at it.

Fixes: 34539b442b3b ("bus: ti-sysc: Flush posted write on enable before reset")
Cc: stable@vger.kernel.org
Signed-off-by: Tony Lindgren 
Signed-off-by: Greg Kroah-Hartman

bus: ti-sysc: Fix SYSC_QUIRK_SWSUP_SIDLE_ACT handling for uart wake-up

2023-10-06T12:56:48+00:00

[ Upstream commit e5deb8f76e64d94ccef715e75ebafffd0c312d80 ]

The uarts should be tagged with SYSC_QUIRK_SWSUP_SIDLE instead of
SYSC_QUIRK_SWSUP_SIDLE_ACT. The difference is that SYSC_QUIRK_SWSUP_SIDLE
is used to force idle target modules rather than block idle during usage.

The SYSC_QUIRK_SWSUP_SIDLE_ACT should disable autoidle and wake-up when
a target module is active, and configure autoidle and wake-up when a
target module is inactive. We are missing configuring the target module
on sysc_disable_module(), and missing toggling of the wake-up bit.

Let's fix the issue to allow uart wake-up to work.

Fixes: fb685f1c190e ("bus: ti-sysc: Handle swsup idle mode quirks")
Tested-by: Dhruva Gole 
Tested-by: Kevin Hilman 
Signed-off-by: Tony Lindgren 
Signed-off-by: Sasha Levin

bus: ti-sysc: Fix missing AM35xx SoC matching

2023-10-06T12:56:46+00:00

[ Upstream commit 11729caa520950e17cd81bc43ffc477c46cf791e ]

Commit feaa8baee82a ("bus: ti-sysc: Implement SoC revision handling")
created a list of SoC types searching for strings based on names
and wildcards which associates the SoC to different families.

The OMAP34xx and OMAP35xx are treated as SOC_3430 while
OMAP36xx and OMAP37xx are treated as SOC_3630, but the AM35xx
isn't listed.

The AM35xx is mostly an OMAP3430, and a later commit a12315d6d270
("bus: ti-sysc: Make omap3 gpt12 quirk handling SoC specific") looks
for the SOC type and behaves in a certain way if it's SOC_3430.

This caused a regression on the AM3517 causing it to return two
errors:

 ti-sysc: probe of 48318000.target-module failed with error -16
 ti-sysc: probe of 49032000.target-module failed with error -16

Fix this by treating the creating SOC_AM35 and inserting it between
the SOC_3430 and SOC_3630.  If it is treaed the same way as the
SOC_3430 when checking the status of sysc_check_active_timer,
the error conditions will disappear.

Fixes: a12315d6d270 ("bus: ti-sysc: Make omap3 gpt12 quirk handling SoC specific")
Fixes: feaa8baee82a ("bus: ti-sysc: Implement SoC revision handling")
Signed-off-by: Adam Ford 
Message-ID: <20230906233442.270835-1-aford173@gmail.com>
Signed-off-by: Tony Lindgren 
Signed-off-by: Sasha Levin

bus: ti-sysc: Use fsleep() instead of usleep_range() in sysc_reset()

2023-10-06T12:56:46+00:00

[ Upstream commit d929b2b7464f95ec01e47f560b1e687482ba8929 ]

The am335x-evm started producing boot errors because of subtle timing
changes:

Unhandled fault: external abort on non-linefetch (0x1008) at 0xf03c1010
...
sysc_reset from sysc_probe+0xf60/0x1514
sysc_probe from platform_probe+0x5c/0xbc
...

The fix consists in using the appropriate sleep function in sysc reset.
For flexible sleeping, fsleep is recommended. Here, sysc delay parameter
can take any value in [0 - 255] us range. As a result, fsleep() should
be used, calling udelay() for a sysc delay lower than 10 us.

Signed-off-by: Julien Panis 
Fixes: e709ed70d122 ("bus: ti-sysc: Fix missing reset delay handling")
Message-ID: <20230821-fix-ti-sysc-reset-v1-1-5a0a5d8fae55@baylibre.com>
Signed-off-by: Tony Lindgren 
Signed-off-by: Sasha Levin

bus: ti-sysc: Configure uart quirks for k3 SoC

2023-09-23T09:11:05+00:00

[ Upstream commit 03a711d3cb83692733f865312f49e665c49de6de ]

Enable the uart quirks similar to the earlier SoCs. Let's assume we are
likely going to need a k3 specific quirk mask separate from the earlier
SoCs, so let's not start changing the revision register mask at this point.

Note that SYSC_QUIRK_LEGACY_IDLE will be needed until we can remove the
need for pm_runtime_irq_safe() from 8250_omap driver.

Reviewed-by: Nishanth Menon 
Signed-off-by: Tony Lindgren 
Signed-off-by: Sasha Levin

bus: mhi: host: Skip MHI reset if device is in RDDM

2023-09-19T10:27:58+00:00

commit cabce92dd805945a090dc6fc73b001bb35ed083a upstream.

In RDDM EE, device can not process MHI reset issued by host. In case of MHI
power off, host is issuing MHI reset and polls for it to get cleared until
it times out. Since this timeout can not be avoided in case of RDDM, skip
the MHI reset in this scenarios.

Cc: 
Fixes: a6e2e3522f29 ("bus: mhi: core: Add support for PM state transitions")
Signed-off-by: Qiang Yu 
Reviewed-by: Jeffrey Hugo 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/1684390959-17836-1-git-send-email-quic_qianyu@quicinc.com
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Greg Kroah-Hartman