[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <874j0ee2ds.ffs@tglx>
Date: Fri, 28 Feb 2025 12:26:39 +0100
From: Thomas Gleixner <tglx@...utronix.de>
To: Hans Zhang <18255117159@....com>, Manivannan Sadhasivam
<manivannan.sadhasivam@...aro.org>
Cc: kw@...ux.com, kwilczynski@...nel.org, bhelgaas@...gle.com,
cassel@...nel.org, linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] genirq/msi: Add the address and data that show MSI/MSIX
On Fri, Feb 28 2025 at 17:04, Hans Zhang wrote:
> Is the following patch OK?
No.
> static void
> irq_debug_show_chip(struct seq_file *m, struct irq_data *data, int ind)
> {
> @@ -178,6 +199,7 @@ static int irq_debug_show(struct seq_file *m, void *p)
> seq_printf(m, "node: %d\n", irq_data_get_node(data));
> irq_debug_show_masks(m, desc);
> irq_debug_show_data(m, data, 0);
> + irq_debug_show_msi_msix(m, data, 0);
> raw_spin_unlock_irq(&desc->lock);
> return 0;
> }
This is just violating the layering and I told you what to do:
"implement a debug_show() callback in the MSI core code and assign
it to domain ops::debug_show() on domain creation, if it does not
provide its own callback."
If you don't understand what I tell you, then please ask instead of
going off and hacking up something completely different.
Here is another hint:
Look at msi_domain_ops_default and at msi_domain_update_dom_ops()
If you still have questions, feel free to ask.
Thanks,
tglx
Powered by blists - more mailing lists