Re: [PATCH 7/9] PCI: endpoint: pci-epf-test: Advertise reserved BARs

[Niklas Cassel <cassel@kernel.org>]

On Wed, Feb 18, 2026 at 11:00:17AM -0500, Frank Li wrote:
> > Thus, we need to tell pci_endpoint_test that these reserved BARs (which are
> > not disabled) has to be skipped, because if we perfrom READ/WRITE tests
> > against these BARs, bad things will happen. E.g. if you write to a BAR that
> > exposes iATU registers, you will clear/overwrite the current inbound addresss
> > translation configured by the endpoint.
> 
> I see. it actually means skip bar test. is CAP_SKIP_BAR<n>_TEST better?

Yes, for pci-epf-test, a reserved BAR should be skipped, but not only for
the BAR<n>_TEST cases, it should also ignore the BAR in the consecutive
BAR test.

Everywhere in drivers/pci/endpoint, a reserved BAR is a BAR with fixed
MMIO registers (to e.g. eDMA registers, MSI-X table, iATU registers).

E.g. the code for the consecutive BAR test in
drivers/misc/pci_endpoint_test.c is:

@@ -403,7 +426,7 @@ static int pci_endpoint_test_bars(struct pci_endpoint_test *test)
 
 	/* Write all BARs in order (without reading). */
 	for (bar = 0; bar < PCI_STD_NUM_BARS; bar++)
-		if (test->bar[bar])
+		if (test->bar[bar] && !bar_is_reserved(test, bar))
 			pci_endpoint_test_bars_write_bar(test, bar);
 
 	/*


It seems quite clear that pci-epf-test should skip a reserved BAR.

I see your point, but I'm leaning towards that that it is better if we use
the same terminology everywhere.
I'm worried that introducing another terminology (SKIP BAR), might lead to
more confusion than clarity (we already have RESERVED BARs and DISABLED BARs
as enum values in pci_epc_bar_type).


So, while SKIP BAR might be slightly more descriptive for pci-epf-test,
I worry that that using a third term, will cause more harm than good.


Kind regards,
Niklas