Re: [PATCH v2 3/3] xen/arm: handle irq_set_type() failures

Hi Jan,

Thank you for the review.

On Wed, Aug 12, 2026 at 09:39:55AM +0200, Jan Beulich wrote:
> On 10.08.2026 20:38, Mykola Kvach wrote:
> > --- a/xen/drivers/char/ns16550.c
> > +++ b/xen/drivers/char/ns16550.c
> > @@ -1928,6 +1928,7 @@ static int __init ns16550_acpi_uart_init(const void *data)
> >      struct acpi_table_header *table;
> >      struct acpi_table_spcr *spcr;
> >      acpi_status status;
> > +    int rc;
> >      /*
> >       * Same as the DT part.
> >       * Only support one UART on ARM which happen to be ns16550_com[0].
> > @@ -1976,7 +1977,9 @@ static int __init ns16550_acpi_uart_init(const void *data)
> >      uart->reg_width = spcr->serial_port.access_width;
> >  
> >      /* The trigger/polarity information is not available in spcr. */
> > -    irq_set_type(spcr->interrupt, IRQ_TYPE_LEVEL_HIGH);
> > +    rc = irq_set_type(spcr->interrupt, IRQ_TYPE_LEVEL_HIGH);
> > +    if ( rc )
> > +        return rc;
> 
> Is erroring out still appropriate when part of ns16550_com[] was already
> modified? I.e. doesn't the call need to move up then?

Good point. Returning there can leave ns16550_com[0] partially
initialized. I'll move irq_set_type() before ns16550_init_common()
and before modifying the UART state.

Best regards,
Mykola