Skip to contentSkip to conversation

tty: metag_da: Add metag DA TTY driver

PATCH In review James Hogan 2 participants 5 messages Feb 6, 2013 TTY layer and serial drivers

lore
Share & tools
lore
#0 lore raw

cover tty: metag_da: Add metag DA TTY driver

Greg: is it okay to take this into your tty tree as it depends on Jiri's tty_flip/tty_port changes? It won't get built until the arch/metag tree comes along and defines the METAG_DA Kconfig symbol.

This is the same metag_da driver as before, but rebased on tty-next and omitting all arch/metag bits. Both parts have already been reviewed (and I squashed the buffer flipping change into the first patch).

Sorry for all the resends by the way. I think I should have just aimed for the tty tree in the first place :-) I'll know for future.

James Hogan (2):
tty: metag_da: Add metag DA TTY driver
tty: metag_da: avoid getting tty kref in dashtty_timer()

 drivers/tty/Kconfig    |   13 +
 drivers/tty/Makefile   |    1 +
 drivers/tty/metag_da.c |  677 ++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 691 insertions(+), 0 deletions(-)
 create mode 100644 drivers/tty/metag_da.c
Signature · 1 line
1.7.7.6
#1 lore raw

PATCH 1/2 tty: metag_da: Add metag DA TTY driver

Add a TTY driver for communicating over a Meta DA (Debug Adapter) channel using the bios channel SWITCH operation.

---
The only changes compared to v4 of arch/metag patchset are trivial:
* removal of arch/metag changes
* squashed in tty_prepare_flip_string and tty_flip_buffer_push fix

 drivers/tty/Kconfig    |   13 +
 drivers/tty/Makefile   |    1 +
 drivers/tty/metag_da.c |  680 ++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 694 insertions(+), 0 deletions(-)
 create mode 100644 drivers/tty/metag_da.c
drivers/tty/Kconfig +13 -0
index 29dfc24..978db34 100644
@@ -406,4 +406,17 @@ config GOLDFISH_TTY
 	help
 	  Console and system TTY driver for the Goldfish virtual platform.
 
+config DA_TTY
+	bool "DA TTY"
+	depends on METAG_DA
+	select SERIAL_NONSTANDARD
+	help
+	  This enables a TTY on a Dash channel.
+
+config DA_CONSOLE
+	bool "DA Console"
+	depends on DA_TTY
+	help
+	  This enables a console on a Dash channel.
+
 endif # TTY
drivers/tty/Makefile +1 -0
index 35649d3..6b78399 100644
@@ -28,5 +28,6 @@ obj-$(CONFIG_SYNCLINKMP)	+= synclinkmp.o
 obj-$(CONFIG_SYNCLINK)		+= synclink.o
 obj-$(CONFIG_PPC_EPAPR_HV_BYTECHAN) += ehv_bytechan.o
 obj-$(CONFIG_GOLDFISH_TTY)	+= goldfish.o
+obj-$(CONFIG_DA_TTY)		+= metag_da.o
 
 obj-y += ipwireless/
drivers/tty/metag_da.c new file +680 -0

680 lines not on this page · raw message

Signature · 1 line
1.7.7.6
#2 lore raw

PATCH 2/2 tty: metag_da: avoid getting tty kref in dashtty_timer()

In reply to James Hogan

Getting the tty kref in dashtty_timer() is no longer necessary since it isn't needed in fetch_data() any longer (due to changes which make the tty flip functions refer to tty_ports instead of tty_structs), so just pass around a channel number instead.

---
 drivers/tty/metag_da.c |   31 ++++++++++++++-----------------
 1 files changed, 14 insertions(+), 17 deletions(-)
drivers/tty/metag_da.c +14 -17
index fc2a36b..0e88862 100644
@@ -147,9 +147,8 @@ static int chancall(int in_bios_function, int in_channel,
 /*
  * Attempts to fetch count bytes from channel and returns actual count.
  */
-static int fetch_data(struct tty_struct *tty)
+static int fetch_data(unsigned int channel)
 {
-	unsigned int channel = tty->index;
 	struct dashtty_port *dport = &dashtty_ports[channel];
 	int received = 0;
 
@@ -180,31 +179,31 @@ unlock:
 }
 
 /**
- * find_channel_to_poll() - Returns kref to the next channel tty to poll.
- * Returns:	The TTY of the next channel to poll, or NULL if no TTY needs
- *		polling. Release with tty_kref_put().
+ * find_channel_to_poll() - Returns number of the next channel to poll.
+ * Returns:	The number of the next channel to poll, or -1 if none need
+ *		polling.
  */
-static struct tty_struct *find_channel_to_poll(void)
+static int find_channel_to_poll(void)
 {
 	static int last_polled_channel;
 	int last = last_polled_channel;
 	int chan;
-	struct tty_struct *tty = NULL;
+	struct dashtty_port *dport;
 
 	for (chan = last + 1; ; ++chan) {
 		if (chan >= NUM_TTY_CHANNELS)
 			chan = 0;
 
-		tty = tty_port_tty_get(&dashtty_ports[chan].port);
-		if (tty) {
+		dport = &dashtty_ports[chan];
+		if (dport->rx_buf) {
 			last_polled_channel = chan;
-			return tty;
+			return chan;
 		}
 
 		if (chan == last)
 			break;
 	}
-	return tty;
+	return -1;
 }
 
 /**
@@ -302,19 +301,17 @@ static int put_data(void *arg)
  */
 static void dashtty_timer(unsigned long ignored)
 {
-	struct tty_struct *tty;
+	int channel;
 
 	/* If there are no ports open do nothing and don't poll again. */
 	if (!atomic_read(&num_channels_need_poll))
 		return;
 
-	tty = find_channel_to_poll();
+	channel = find_channel_to_poll();
 
 	/* Did we find a channel to poll? */
-	if (tty) {
-		fetch_data(tty);
-		tty_kref_put(tty);
-	}
+	if (channel >= 0)
+		fetch_data(channel);
 
 	mod_timer_pinned(&poll_timer, jiffies + DA_TTY_POLL);
 }
Signature · 1 line
1.7.7.6
#3 lore raw
In reply to James Hogan
Quoted 3 lines · James Hogan

Greg: is it okay to take this into your tty tree as it depends on Jiri's tty_flip/tty_port changes? It won't get built until the arch/metag tree comes along and defines the METAG_DA Kconfig symbol.

Sure, I'll pick it up and push it to Linus for 3.9-rc1.

thanks,

greg k-h

#4 lore raw
In reply to James Hogan
Quoted 10 lines · James Hogan

Greg: is it okay to take this into your tty tree as it depends on Jiri's tty_flip/tty_port changes? It won't get built until the arch/metag tree comes along and defines the METAG_DA Kconfig symbol.

This is the same metag_da driver as before, but rebased on tty-next and omitting all arch/metag bits. Both parts have already been reviewed (and I squashed the buffer flipping change into the first patch).

Sorry for all the resends by the way. I think I should have just aimed for the tty tree in the first place :-) I'll know for future.

Ah, no problem with that, it's tough to figure out how to get a whole new arch merged at times, you've done a great job here, nice work.

greg k-h