Thread (12 messages) 12 messages, 4 authors, 2019-08-09

Re: [PATCH v3 5/7] mfd: ioc3: Add driver for SGI IOC3 chip

From: Thomas Bogendoerfer <hidden>
Date: 2019-07-29 18:46:02
Also in: linux-input, linux-mips, linux-rtc, linux-serial, lkml

On Thu, 25 Jul 2019 12:47:16 +0100
Lee Jones [off-list ref] wrote:
On Thu, 13 Jun 2019, Thomas Bogendoerfer wrote:
quoted
+/*
+ * On IP30 the RTC (a DS1687) is behind the IOC3 on the generic
+ * ByteBus regions. We have to write the RTC address of interest to
+ * IOC3_BYTEBUS_DEV1, then read the data from IOC3_BYTEBUS_DEV2.
+ * rtc->regs already points to IOC3_BYTEBUS_DEV1.
+ */
+#define IP30_RTC_ADDR(rtc) (rtc->regs)
+#define IP30_RTC_DATA(rtc) ((rtc->regs) + IOC3_BYTEBUS_DEV2 - IOC3_BYTEBUS_DEV1)
+
+static u8 ip30_rtc_read(struct ds1685_priv *rtc, int reg)
+{
+	writeb((reg & 0x7f), IP30_RTC_ADDR(rtc));
+	return readb(IP30_RTC_DATA(rtc));
+}
+
+static void ip30_rtc_write(struct ds1685_priv *rtc, int reg, u8 value)
+{
+	writeb((reg & 0x7f), IP30_RTC_ADDR(rtc));
+	writeb(value, IP30_RTC_DATA(rtc));
+}
Why is this not in the RTC driver?
because rtc1685 is used in different systems and accessing the chip
differs between those systems. 
quoted
+static struct ds1685_rtc_platform_data ip30_rtc_platform_data = {
+	.bcd_mode = false,
+	.no_irq = false,
+	.uie_unsupported = true,
+	.alloc_io_resources = true,
quoted
+	.plat_read = ip30_rtc_read,
+	.plat_write = ip30_rtc_write,
Call-backs in a non-subsystem API is pretty ugly IMHO.
I agree
Where are these called from?
drivers/rtc/rtc-ds1685.c

I could do the same as done for serial8250 and add an additional .c file
in  drivers/rtc which handles this for SGI-IP30. Alexandre would this work
for you as well ?
quoted
+#define IOC3_SID(_name, _sid, _setup) \
+	{								   \
+		.name = _name,						   \
+		.sid = (PCI_VENDOR_ID_SGI << 16) | IOC3_SUBSYS_ ## _sid,   \
+		.setup = _setup,					   \
+	}
+
+static struct {
+	const char *name;
+	u32 sid;
+	int (*setup)(struct ioc3_priv_data *ipd);
+} ioc3_infos[] = {
IMHO it's neater if you separate the definition and static data part.
I don't quite understand what you mean here. Should I move the #define at
the beginning of the file ? Why is it neater ?

Thomas.

-- 
SUSE Linux GmbH
GF: Felix Imendörffer, Mary Higgins, Sri Rasiah
HRB 21284 (AG Nürnberg)
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help