Re: [PATCH] nfc: pn533: add error handling for i2c_master_send() in pn533_i2c_send_ack()

[Simon Horman]

On Thu, Mar 19, 2026 at 02:50:05PM +0800, Wenyuan Li wrote:
> In pn533_i2c_send_ack(), the return value of i2c_master_send() is not
> checked. If the I2C transfer fails, the driver continues execution
> without detecting the error, which may lead to incorrect device state
> or silent failures.

I would like to clarify that this patch addresses the silent failures
aspect, but not the potentially incorrect device state aspect. Although,
with notification users may be better able to diagnose that problem.

If so, I agree this patch should be treated as an enhancement
and routed via the net-next tree (the default for Networking patches)
without a fixes tag.

But for reference, it would be better if the tree was targeted like this:

Subject: [PATCH net-next]: ...

> 
> Specifically:
> 1. The ACK may not be sent to the device, leaving it in an undefined state
> 2. The caller (pn533_i2c_abort_cmd()) ignores the return value entirely
> 3. No error logging is provided for debugging
> 
> Fix this by:
> - Adding proper return value check for i2c_master_send()
> - Converting partial sends to -EIO and preserving kernel error codes
> - Adding nfc_err() logging with %pe format for human-readable errors
> - Propagating the error to the caller pn533_i2c_abort_cmd()
> 
> Even if ACK sending fails, the abort procedure continues by scheduling
> cmd_complete_work to ensure userspace is notified, but the error is
> properly logged for debugging.
> 
> Signed-off-by: Wenyuan Li <2063309626@xxxxxx>

...