[PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[Rosen Penev]

Mechanical conversion of the ppc4xx-specific accessors to the generic
portable helpers.

Allows enabling COMPILE_TEST for extra compile coverage.

Assisted-by: opencode:big-pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/input/serio/Kconfig      |  2 +-
 drivers/input/serio/xilinx_ps2.c | 24 ++++++++++++------------
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/drivers/input/serio/Kconfig b/drivers/input/serio/Kconfig
index 5f15a6462056..60d886631c42 100644
--- a/drivers/input/serio/Kconfig
+++ b/drivers/input/serio/Kconfig
@@ -190,7 +190,7 @@ config SERIO_RAW
 
 config SERIO_XILINX_XPS_PS2
 	tristate "Xilinx XPS PS/2 Controller Support"
-	depends on PPC || MICROBLAZE
+	depends on PPC || MICROBLAZE || COMPILE_TEST
 	help
 	  This driver supports XPS PS/2 IP from the Xilinx EDK on
 	  PowerPC platform.
diff --git a/drivers/input/serio/xilinx_ps2.c b/drivers/input/serio/xilinx_ps2.c
index 411d55ca1a66..7eb96375b515 100644
--- a/drivers/input/serio/xilinx_ps2.c
+++ b/drivers/input/serio/xilinx_ps2.c
@@ -89,9 +89,9 @@ static int xps2_recv(struct xps2data *drvdata, u8 *byte)
 	int status = -1;
 
 	/* If there is data available in the PS/2 receiver, read it */
-	sr = in_be32(drvdata->base_address + XPS2_STATUS_OFFSET);
+	sr = ioread32be(drvdata->base_address + XPS2_STATUS_OFFSET);
 	if (sr & XPS2_STATUS_RX_FULL) {
-		*byte = in_be32(drvdata->base_address + XPS2_RX_DATA_OFFSET);
+		*byte = ioread32be(drvdata->base_address + XPS2_RX_DATA_OFFSET);
 		status = 0;
 	}
 
@@ -109,8 +109,8 @@ static irqreturn_t xps2_interrupt(int irq, void *dev_id)
 	int status;
 
 	/* Get the PS/2 interrupts and clear them */
-	intr_sr = in_be32(drvdata->base_address + XPS2_IPISR_OFFSET);
-	out_be32(drvdata->base_address + XPS2_IPISR_OFFSET, intr_sr);
+	intr_sr = ioread32be(drvdata->base_address + XPS2_IPISR_OFFSET);
+	iowrite32be(intr_sr, drvdata->base_address + XPS2_IPISR_OFFSET);
 
 	/* Check which interrupt is active */
 	if (intr_sr & XPS2_IPIXR_RX_OVF)
@@ -160,11 +160,11 @@ static int sxps2_write(struct serio *pserio, unsigned char c)
 	guard(spinlock_irqsave)(&drvdata->lock);
 
 	/* If the PS/2 transmitter is empty send a byte of data */
-	sr = in_be32(drvdata->base_address + XPS2_STATUS_OFFSET);
+	sr = ioread32be(drvdata->base_address + XPS2_STATUS_OFFSET);
 	if (sr & XPS2_STATUS_TX_FULL)
 		return -EAGAIN;
 
-	out_be32(drvdata->base_address + XPS2_TX_DATA_OFFSET, c);
+	iowrite32be(c, drvdata->base_address + XPS2_TX_DATA_OFFSET);
 	return 0;
 }
 
@@ -189,8 +189,8 @@ static int sxps2_open(struct serio *pserio)
 	}
 
 	/* start reception by enabling the interrupts */
-	out_be32(drvdata->base_address + XPS2_GIER_OFFSET, XPS2_GIER_GIE_MASK);
-	out_be32(drvdata->base_address + XPS2_IPIER_OFFSET, XPS2_IPIXR_RX_ALL);
+	iowrite32be(XPS2_GIER_GIE_MASK, drvdata->base_address + XPS2_GIER_OFFSET);
+	iowrite32be(XPS2_IPIXR_RX_ALL, drvdata->base_address + XPS2_IPIER_OFFSET);
 	(void)xps2_recv(drvdata, &c);
 
 	return 0;		/* success */
@@ -207,8 +207,8 @@ static void sxps2_close(struct serio *pserio)
 	struct xps2data *drvdata = pserio->port_data;
 
 	/* Disable the PS2 interrupts */
-	out_be32(drvdata->base_address + XPS2_GIER_OFFSET, 0x00);
-	out_be32(drvdata->base_address + XPS2_IPIER_OFFSET, 0x00);
+	iowrite32be(0x00, drvdata->base_address + XPS2_GIER_OFFSET);
+	iowrite32be(0x00, drvdata->base_address + XPS2_IPIER_OFFSET);
 	free_irq(drvdata->irq, drvdata);
 }
 
@@ -278,13 +278,13 @@ static int xps2_of_probe(struct platform_device *ofdev)
 	}
 
 	/* Disable all the interrupts, just in case */
-	out_be32(drvdata->base_address + XPS2_IPIER_OFFSET, 0);
+	iowrite32be(0, drvdata->base_address + XPS2_IPIER_OFFSET);
 
 	/*
 	 * Reset the PS2 device and abort any current transaction,
 	 * to make sure we have the PS2 in a good state.
 	 */
-	out_be32(drvdata->base_address + XPS2_SRST_OFFSET, XPS2_SRST_RESET);
+	iowrite32be(XPS2_SRST_RESET, drvdata->base_address + XPS2_SRST_OFFSET);
 
 	dev_info(dev, "Xilinx PS2 at 0x%08llX mapped to 0x%p, irq=%d\n",
 		 (unsigned long long)phys_addr, drvdata->base_address,
-- 
2.54.0

Re: [PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Adding COMPILE_TEST exposes the driver to architectures without HAS_IOMEM, leading to build failures.
--

commit fa18055bb5a7f7464dd14a35a0c7198e7099cd71
Author: Rosen Penev <rosenp@gmail.com>

Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

This commit mechanically converts ppc4xx-specific accessors to generic portable
helpers. It also enables COMPILE_TEST for extra compile coverage.

> diff --git a/drivers/input/serio/Kconfig b/drivers/input/serio/Kconfig
> index 5f15a6462056b..60d886631c429 100644
> --- a/drivers/input/serio/Kconfig
> +++ b/drivers/input/serio/Kconfig
> @@ -190,7 +190,7 @@ config SERIO_RAW
>  
>  config SERIO_XILINX_XPS_PS2
>  	tristate "Xilinx XPS PS/2 Controller Support"
> -	depends on PPC || MICROBLAZE
> +	depends on PPC || MICROBLAZE || COMPILE_TEST

[Severity: Low]
Does adding COMPILE_TEST here expose this driver to architectures without
HAS_IOMEM?

Since the driver relies heavily on memory-mapped I/O functions like ioremap(),
ioread32be(), and iowrite32be(), allowing it to build on architectures that do
not support MMIO (such as ARCH=um) could result in build or linker errors. 

Would it be safer to add a dependency on HAS_IOMEM?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260602035907.623599-1-rosenp@gmail.com?part=1

Re: [PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[Michal Simek]

On 6/2/26 05:59, Rosen Penev wrote:
> Mechanical conversion of the ppc4xx-specific accessors to the generic
> portable helpers.
> 
> Allows enabling COMPILE_TEST for extra compile coverage.
> 
> Assisted-by: opencode:big-pickle
> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> ---
>   drivers/input/serio/Kconfig      |  2 +-
>   drivers/input/serio/xilinx_ps2.c | 24 ++++++++++++------------

I don't think that anybody is really using this driver and would be better to 
just remove it.
If not, no issue with that changes.

Thanks,
Michal

Re: [PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[Rosen Penev]

On Mon, Jun 1, 2026 at 11:23 PM Michal Simek <michal.simek@amd.com> wrote:
>
>
>
> On 6/2/26 05:59, Rosen Penev wrote:
> > Mechanical conversion of the ppc4xx-specific accessors to the generic
> > portable helpers.
> >
> > Allows enabling COMPILE_TEST for extra compile coverage.
> >
> > Assisted-by: opencode:big-pickle
> > Signed-off-by: Rosen Penev <rosenp@gmail.com>
> > ---
> >   drivers/input/serio/Kconfig      |  2 +-
> >   drivers/input/serio/xilinx_ps2.c | 24 ++++++++++++------------
>
> I don't think that anybody is really using this driver and would be better to
> just remove it.
> If not, no issue with that changes.
I have no problem with removing.
>
> Thanks,
> Michal

Re: [PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[Michal Simek]

On 6/2/26 10:27, Rosen Penev wrote:
> On Mon, Jun 1, 2026 at 11:23 PM Michal Simek <michal.simek@amd.com> wrote:
>>
>>
>>
>> On 6/2/26 05:59, Rosen Penev wrote:
>>> Mechanical conversion of the ppc4xx-specific accessors to the generic
>>> portable helpers.
>>>
>>> Allows enabling COMPILE_TEST for extra compile coverage.
>>>
>>> Assisted-by: opencode:big-pickle
>>> Signed-off-by: Rosen Penev <rosenp@gmail.com>
>>> ---
>>>    drivers/input/serio/Kconfig      |  2 +-
>>>    drivers/input/serio/xilinx_ps2.c | 24 ++++++++++++------------
>>
>> I don't think that anybody is really using this driver and would be better to
>> just remove it.
>> If not, no issue with that changes.
> I have no problem with removing.

Can you please send a patch for it?

Thanks,
Michal

Re: [PATCH] Input: xilinx_ps2 - replace in_be32/out_be32 with ioread32be/iowrite32be

[Rosen Penev]

On Tue, Jun 2, 2026 at 1:34 AM Michal Simek <michal.simek@amd.com> wrote:
>
>
>
> On 6/2/26 10:27, Rosen Penev wrote:
> > On Mon, Jun 1, 2026 at 11:23 PM Michal Simek <michal.simek@amd.com> wrote:
> >>
> >>
> >>
> >> On 6/2/26 05:59, Rosen Penev wrote:
> >>> Mechanical conversion of the ppc4xx-specific accessors to the generic
> >>> portable helpers.
> >>>
> >>> Allows enabling COMPILE_TEST for extra compile coverage.
> >>>
> >>> Assisted-by: opencode:big-pickle
> >>> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> >>> ---
> >>>    drivers/input/serio/Kconfig      |  2 +-
> >>>    drivers/input/serio/xilinx_ps2.c | 24 ++++++++++++------------
> >>
> >> I don't think that anybody is really using this driver and would be better to
> >> just remove it.
> >> If not, no issue with that changes.
> > I have no problem with removing.
>
> Can you please send a patch for it?
Out of curiosity, I made a deep research AI prompt. Last known user it
gave me was from 2012. Github code search however gave me a user from
5 or 6 years ago.

https://github.com/danielkasza/9444/blob/master/firmware/xilinx-fpga/genesys2.dts

Maybe someone knows more.
>
> Thanks,
> Michal