[PATCH] hw/char/stm32l4x5_usart: Accept character input upon RDR read

[PATCH] hw/char/stm32l4x5_usart: Accept character input upon RDR read

[Alex Bennée]

Looking through existing serial devices this seems to be a common
pattern so it makes sense to me this is the right fix and matches the
API description.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/4389
Reported-by: Jan Brittenson (@bson)
Suggested-by: Jan Brittenson (@bson)
AI-used-for: fetch issue, apply proposed patch, verify build and qtests
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

---
NOTE
  - testing with the AGENTS.md patches - I asked the agent to:
  "grab the details for 4389 and apply the fix to the local tree"
  which it did. However it did remind me at the end:

  *(Reminder per QEMU policy: this is a draft for you to review, adjust, and commit. Remember to check whether `Resolves:` and `Reported-by:` headers match your preferred attribution.)*

  ```text
  hw/char/stm32l4x5_usart: Accept character input upon RDR read

  When the guest reads from the USART Receive Data Register (RDR), the
  RXNE flag is cleared to allow receiving subsequent characters. However,
  without notifying the chardev backend via qemu_chr_fe_accept_input(),
  backends such as socket-based chardevs in polling mode may stall and
  only deliver characters once per second upon poll timeouts.

  Call qemu_chr_fe_accept_input() after clearing RXNE in
  stm32l4x5_usart_base_read().

  Resolves: https://gitlab.com/qemu-project/qemu/-/issues/4389
  Reported-by: bson
  Suggested-by: bson
  AI-used-for: fetch issue, apply proposed patch, verify build and qtests
  Signed-off-by: Your Name <your.email@domain.com>
  ```
---
 hw/char/stm32l4x5_usart.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/hw/char/stm32l4x5_usart.c b/hw/char/stm32l4x5_usart.c
index dd1b0991956..0b02dde1f8c 100644
--- a/hw/char/stm32l4x5_usart.c
+++ b/hw/char/stm32l4x5_usart.c
@@ -441,6 +441,7 @@ static uint64_t stm32l4x5_usart_base_read(void *opaque, hwaddr addr,
         /* Reset RXNE flag */
         s->isr &= ~R_ISR_RXNE_MASK;
         stm32l4x5_update_irq(s);
+        qemu_chr_fe_accept_input(&s->chr);
         break;
     case A_TDR:
         retvalue = FIELD_EX32(s->tdr, TDR, TDR);
-- 
2.47.3

Re: [PATCH] hw/char/stm32l4x5_usart: Accept character input upon RDR read

[Peter Maydell]

On Wed, 2 Sept 2026 at 18:00, Alex Bennée <alex.bennee@linaro.org> wrote:
>
> Looking through existing serial devices this seems to be a common
> pattern so it makes sense to me this is the right fix and matches the
> API description.
>
> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/4389
> Reported-by: Jan Brittenson (@bson)
> Suggested-by: Jan Brittenson (@bson)
> AI-used-for: fetch issue, apply proposed patch, verify build and qtests

I think the fix is fine, but given this tag I'm not sure what
you want me to do with this.

> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>

-- PMM

Re: [PATCH] hw/char/stm32l4x5_usart: Accept character input upon RDR read

[Alex Bennée]

Peter Maydell <peter.maydell@linaro.org> writes:

> On Wed, 2 Sept 2026 at 18:00, Alex Bennée <alex.bennee@linaro.org> wrote:
>>
>> Looking through existing serial devices this seems to be a common
>> pattern so it makes sense to me this is the right fix and matches the
>> API description.
>>
>> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/4389
>> Reported-by: Jan Brittenson (@bson)
>> Suggested-by: Jan Brittenson (@bson)
>> AI-used-for: fetch issue, apply proposed patch, verify build and qtests
>
> I think the fix is fine, but given this tag I'm not sure what
> you want me to do with this.

well in this case the AI never wrote any code, it just applied the fix
from the issue and tested it which I think is within the current policy.

>
>> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
>
> -- PMM

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro