Re: [RFC 1/8] pci: adding new irq api to pci-designware

[Lucas Stach <l.stach@pengutronix.de>]

Am Freitag, den 12.05.2017, 17:56 +0100 schrieb Joao Pinto:
> This patch adds the new interrupt api to pecie-designware, keeping the old
> one. A define was added to toggle between the two apis (for testing purposes).
> 
> Signed-off-by: Joao Pinto <jpinto@synopsys.com>
> ---
>  drivers/pci/dwc/pcie-designware-host.c | 241 ++++++++++++++++++++++++++++++++-
>  drivers/pci/dwc/pcie-designware.h      |   5 +
>  2 files changed, 244 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/dwc/pcie-designware-host.c b/drivers/pci/dwc/pcie-designware-host.c
> index 28ed32b..6810ea10b 100644
> --- a/drivers/pci/dwc/pcie-designware-host.c
> +++ b/drivers/pci/dwc/pcie-designware-host.c

[...]

> +static int dw_pcie_irq_domain_alloc(struct irq_domain *domain,
> +				    unsigned int virq, unsigned int nr_irqs,
> +				    void *args)
> +{
> +	struct dw_pcie *pci = domain->host_data;
> +	struct pcie_port *pp = &pci->pp;
> +	unsigned long bit;
> +
> +	mutex_lock(&pp->lock);
> +
> +	WARN_ON(nr_irqs != 1);

So the new API doesn't support multi MSI, where we need more than a
single contiguous MSI in a range? If this is the case then this change
is breaking real use-cases that work today.

Regards,
Luvas