Message ID | 20211227164317.4146918-28-schnelle@linux.ibm.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | Kconfig: Introduce HAS_IOPORT and LEGACY_PCI options | expand |
Make the subject match historical convention (capitalize "Make"). On Mon, Dec 27, 2021 at 05:43:12PM +0100, Niklas Schnelle wrote: > Exporting I/O resources only makes sense if legacy I/O spaces are > supported so conditionally add them only if HAS_IOPORT is set. IIUC, the effect of this is that the "resource%d" file for an I/O BAR still appears in /sys, but reads or writes will fail with ENXIO. Worth mentioning that in the commit log, since one could interpret the above as meaning that the "resource%d" file exists only if HAS_IOPORT is set. I think I will *exist* but not be very useful. I also wonder what this looks like in the sysfs "resource" file and via lspci. I suppose it's useful if lspci shows the fact that the BAR exists and is an I/O BAR, even if the arch doesn't set HAS_IOPORT. > Co-developed-by: Arnd Bergmann <arnd@kernel.org> > Signed-off-by: Arnd Bergmann <arnd@kernel.org> > Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com> > --- > drivers/pci/pci-sysfs.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c > index cfe2f85af09e..a59a85593972 100644 > --- a/drivers/pci/pci-sysfs.c > +++ b/drivers/pci/pci-sysfs.c > @@ -1099,6 +1099,7 @@ static int pci_mmap_resource_wc(struct file *filp, struct kobject *kobj, > return pci_mmap_resource(kobj, attr, vma, 1); > } > > +#ifdef CONFIG_HAS_IOPORT > static ssize_t pci_resource_io(struct file *filp, struct kobject *kobj, > struct bin_attribute *attr, char *buf, > loff_t off, size_t count, bool write) > @@ -1157,6 +1158,21 @@ static ssize_t pci_write_resource_io(struct file *filp, struct kobject *kobj, > > return pci_resource_io(filp, kobj, attr, buf, off, count, true); > } > +#else > +static ssize_t pci_read_resource_io(struct file *filp, struct kobject *kobj, > + struct bin_attribute *attr, char *buf, > + loff_t off, size_t count) > +{ > + return -ENXIO; > +} I assume the sysfs infrastructure prevents or fails reads/write if res_attr->read and res_attr->write are not set at all, so maybe we wouldn't need the stubs if we did something like this? if (pci_resource_flags(pdev, num) & IORESOURCE_IO) { #ifdef CONFIG_HAS_IOPORT res_attr->read = pci_read_resource_io; res_attr->write = pci_write_resource_io; ... #endif } else { > +static ssize_t pci_write_resource_io(struct file *filp, struct kobject *kobj, > + struct bin_attribute *attr, char *buf, > + loff_t off, size_t count) > +{ > + return -ENXIO; > +} > +#endif > > /** > * pci_remove_resource_files - cleanup resource files > -- > 2.32.0 >
diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index cfe2f85af09e..a59a85593972 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -1099,6 +1099,7 @@ static int pci_mmap_resource_wc(struct file *filp, struct kobject *kobj, return pci_mmap_resource(kobj, attr, vma, 1); } +#ifdef CONFIG_HAS_IOPORT static ssize_t pci_resource_io(struct file *filp, struct kobject *kobj, struct bin_attribute *attr, char *buf, loff_t off, size_t count, bool write) @@ -1157,6 +1158,21 @@ static ssize_t pci_write_resource_io(struct file *filp, struct kobject *kobj, return pci_resource_io(filp, kobj, attr, buf, off, count, true); } +#else +static ssize_t pci_read_resource_io(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + return -ENXIO; +} + +static ssize_t pci_write_resource_io(struct file *filp, struct kobject *kobj, + struct bin_attribute *attr, char *buf, + loff_t off, size_t count) +{ + return -ENXIO; +} +#endif /** * pci_remove_resource_files - cleanup resource files