kernel.git/drivers/i3c, branch linux-6.4.y

i3c: master: svc: fix probe failure when no i3c device exist

2023-09-13T07:48:39Z

commit 6e13d6528be2f7e801af63c8153b87293f25d736 upstream. I3C masters are expected to support hot-join. This means at initialization time we might not yet discover any device and this should not be treated as a fatal error. During the DAA procedure which happens at probe time, if no device has joined, all CCC will be NACKed (from a bus perspective). This leads to an early return with an error code which fails the probe of the master. Let's avoid this by just telling the core through an I3C_ERROR_M2 return command code that no device was discovered, which is a valid situation. This way the master will no longer bail out and fail to probe for a wrong reason. Cc: stable@vger.kernel.org Fixes: dd3c52846d59 ("i3c: master: svc: Add Silvaco I3C master driver") Signed-off-by: Frank Li Acked-by: Miquel Raynal Link: https://lore.kernel.org/r/20230831141324.2841525-1-Frank.Li@nxp.com Signed-off-by: Alexandre Belloni Signed-off-by: Greg Kroah-Hartman

i3c: master: svc: fix cpu schedule in spin lock

2023-07-19T14:36:33Z

[ Upstream commit 33beadb3b1ab74e69db2c49d9663f3a93a273943 ] pm_runtime_resume_and_get() may call sleep(). It cannot be used in svc_i3c_master_start_xfer_locked(), because it is in a spin lock. Move the pm runtime operations to svc_i3c_master_enqueue_xfer(). Signed-off-by: Clark Wang Fixes: 05be23ef78f7 ("i3c: master: svc: add runtime pm support") Reviewed-by: Miquel Raynal Link: https://lore.kernel.org/r/20230517033030.3068085-2-xiaoning.wang@nxp.com Signed-off-by: Alexandre Belloni Signed-off-by: Sasha Levin

i3c: ast2600: set variable ast2600_i3c_ops storage-class-specifier to static

2023-04-30T21:50:26Z

smatch reports drivers/i3c/master/ast2600-i3c-master.c:121:34: warning: symbol 'ast2600_i3c_ops' was not declared. Should it be static? This variable is only used in its defining file, so it should be static. Signed-off-by: Tom Rix Reviewed-by: Jeremy Kerr Link: https://lore.kernel.org/r/20230429134601.2688558-1-trix@redhat.com Signed-off-by: Alexandre Belloni

i3c: ast2600: fix register setting for 545 ohm pullups

2023-04-28T06:52:23Z

The 2k register setting is zero, OR-ing it in doesn't parallel the 2k and 750 ohm pullups. We need a separate value for the 545 ohm setting. Reported-by: Lukwinski Zbigniew Signed-off-by: Jeremy Kerr Link: https://lore.kernel.org/r/20230428001849.1775559-1-jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: ast2600: enable IBI support

2023-04-28T06:49:50Z

The ast2600 i3c hardware is capable of IBIs, but we need a workaround for a hardware issue with the I3C state machine handling IBI payloads of specific lengths when PEC is not enabled. To avoid this, we need to unconditionally enable PECs, at the consquence of losing a byte of data when the device does not send a PEC. Enable IBIs on the ast2600 platform, including an implementation of the PEC workaround, which prints a warning when triggered. Signed-off-by: Jeremy Kerr Reviewed-by: Joel Stanley Link: https://lore.kernel.org/r/ba923b96d6d129024c975e8a0472c5b2fcb3af32.1680161823.git.jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: dw: Add a platform facility for IBI PEC workarounds

2023-04-28T06:20:07Z

On the AST2600 i3c controller, we'll need to apply a workaround for a hardware issue with IBI payloads. Introduce a platform hook to allow dw i3c platform implementations to modify the DAT entry in IBI enable/disable to allow this workaround in a future change. Signed-off-by: Jeremy Kerr Reviewed-by: Joel Stanley Link: https://lore.kernel.org/r/d5d76a8d2336d2a71886537f42e71d51db184df6.1680161823.git.jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: dw: Add support for in-band interrupts

2023-04-28T06:20:07Z

This change adds support for receiving and dequeueing i3c IBIs. By setting struct dw_i3c_master->ibi_capable before probe, a platform implementation can select the IBI-enabled version of the i3c_master_ops, enabling the global IBI infrastrcture for that controller. Signed-off-by: Jeremy Kerr Reviewed-by: Joel Stanley Link: https://lore.kernel.org/r/79daeefd7ccb7c935d0c159149df21a6c9a73ffa.1680161823.git.jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: dw: Turn DAT array entry into a struct

2023-04-28T06:20:07Z

In an upcoming change, we will want to store additional data about the devices we have in the data address table. Change the type of the DAT entries into a struct, which currently just has the address data. Signed-off-by: Jeremy Kerr Reviewed-by: Joel Stanley Link: https://lore.kernel.org/r/9dc0d9e2857e851a0cf04819df48e5d31921f83e.1680161823.git.jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: dw: Create a generic fifo read function

2023-04-28T06:20:07Z

In a future change we'll want to read from the IBI FIFO too, so turn dw_i3c_read_rx_fifo() into a generic read with the FIFO register as a parameter. Signed-off-by: Jeremy Kerr Reviewed-by: Joel Stanley Link: https://lore.kernel.org/r/827204789583dd86addffb47ecaeab9d67cf95d5.1680161823.git.jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni

i3c: Allow OF-alias-based persistent bus numbering

2023-04-28T06:19:01Z

Parse the /aliases node to assign any fixed bus numbers, as is done with the i2c subsystem. Numbering for non-aliased busses will start after the highest fixed bus number. This allows an alias node such as: aliases { i3c0 = &bus_a, i3c4 = &bus_b, }; to set the numbering for a set of i3c controllers: /* fixed-numbered bus, assigned "i3c-0" */ bus_a: i3c-master { }; /* another fixed-numbered bus, assigned "i3c-4" */ bus_b: i3c-master { }; /* dynamic-numbered bus, likely assigned "i3c-5" */ bus_c: i3c-master { }; If no i3c device aliases are present, the numbering will stay as-is, starting from 0. Signed-off-by: Jeremy Kerr Link: https://lore.kernel.org/r/20230405094149.1513209-1-jk@codeconstruct.com.au Signed-off-by: Alexandre Belloni