* [PATCH 02/10] via-cuda: Remove redundant temporary variable
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 06/10] via-cuda: Avoid TREQ race condition Finn Thain
` (8 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
There is no possibility that current_req can change during execution of
cuda_start(). This can be confirmed by inspection: cuda_lock is always
held whenever cuda_start() is called or current_req is modified.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 8 ++------
1 file changed, 2 insertions(+), 6 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index 98ced65..4eb8a884 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -414,19 +414,15 @@ cuda_write(struct adb_request *req)
static void
cuda_start(void)
{
- struct adb_request *req;
-
/* assert cuda_state == idle */
- /* get the packet to send */
- req = current_req;
- if (req == 0)
+ if (current_req == NULL)
return;
if ((in_8(&via[B]) & TREQ) == 0)
return; /* a byte is coming in from the CUDA */
/* set the shift register to shift out and send a byte */
out_8(&via[ACR], in_8(&via[ACR]) | SR_OUT);
- out_8(&via[SR], req->data[0]);
+ out_8(&via[SR], current_req->data[0]);
out_8(&via[B], in_8(&via[B]) & ~TIP);
cuda_state = sent_first_byte;
}
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 06/10] via-cuda: Avoid TREQ race condition
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
2016-12-01 6:09 ` [PATCH 02/10] via-cuda: Remove redundant temporary variable Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 05/10] via-cuda: Fix re-initialization of reply_ptr and reading_reply Finn Thain
` (7 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
When a read transaction completes, one of several things will happen:
either a new transfer is started by the driver, a new transfer request
is raised by the Cuda (i.e. TREQ asserted), or both happen at once.
When both happen at once, there is a race condition between the TREQ test
in the read_done state and the same test in cuda_start(). Moreover, the
former test uses a stale TREQ value.
Theoretically, this can result in the undesirable outcome that the
interrupt handler completes with the state machine 'idle' when it should
instead start the next transaction.
Avoid this race by calling cuda_start() first and then confirming that it
succeeded. If not, test the current TREQ value before entering the
'reading' state.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index e65c0b6..ff9062a 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -605,12 +605,11 @@ cuda_interrupt(int irq, void *arg)
memcpy(ibuf, cuda_rbuf, ibuf_len);
}
reply_ptr = cuda_rbuf;
- if (TREQ_asserted(status)) {
+ cuda_state = idle;
+ cuda_start();
+ if (cuda_state == idle && TREQ_asserted(in_8(&via[B]))) {
assert_TIP();
cuda_state = reading;
- } else {
- cuda_state = idle;
- cuda_start();
}
break;
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 05/10] via-cuda: Fix re-initialization of reply_ptr and reading_reply
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
2016-12-01 6:09 ` [PATCH 02/10] via-cuda: Remove redundant temporary variable Finn Thain
2016-12-01 6:09 ` [PATCH 06/10] via-cuda: Avoid TREQ race condition Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 07/10] via-cuda: Use spinlock_irq_save/restore instead of enable/disable_irq Finn Thain
` (6 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
When reading_reply is set, reply_ptr points into an adb_request struct.
Hence, when reply_ptr instead points into the global cuda_rbuf,
it must be the case that reading_reply is not set.
Unfortunately, this rule can be violated because re-initialization
of reply_ptr and reading_reply presently depends on the TREQ input.
Fix this by re-initializing reply_ptr and reading_reply as soon as they
are known to be invalid.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index acf3a95..e65c0b6 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -592,6 +592,7 @@ cuda_interrupt(int irq, void *arg)
}
current_req = req->next;
complete = 1;
+ reading_reply = 0;
} else {
/* This is tricky. We must break the spinlock to call
* cuda_input. However, doing so means we might get
@@ -603,11 +604,10 @@ cuda_interrupt(int irq, void *arg)
ibuf_len = reply_ptr - cuda_rbuf;
memcpy(ibuf, cuda_rbuf, ibuf_len);
}
+ reply_ptr = cuda_rbuf;
if (TREQ_asserted(status)) {
assert_TIP();
cuda_state = reading;
- reply_ptr = cuda_rbuf;
- reading_reply = 0;
} else {
cuda_state = idle;
cuda_start();
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 07/10] via-cuda: Use spinlock_irq_save/restore instead of enable/disable_irq
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (2 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 05/10] via-cuda: Fix re-initialization of reply_ptr and reading_reply Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 04/10] via-cuda: Prevent read buffer overflow Finn Thain
` (5 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
The cuda_start() function uses spinlock_irq_save/restore for mutual
exclusion. Let's have cuda_poll() do the same when polling the VIA
interrupt.
The benefit to disabling local irqs when the interrupt is being polled
is that the interrupt handler now has the same timing properties
regardless of whether it is invoked normally or from cuda_poll().
This driver was written back when local irqs remained enabled during
execution of interrupt handlers and cuda_poll() was probably trying
to achieve the same effect by use of enable/disable_irq.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
I wondered whether the use of enable/disable_irq was an attempt to reduce
interrupt latency given that VIA register accesses are slow. But if the
interrupt has not yet fired, only one VIA access will take place, which
imposes neglible latency anyway.
---
drivers/macintosh/via-cuda.c | 16 +++++-----------
1 file changed, 5 insertions(+), 11 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index ff9062a..b801c67 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -459,14 +459,7 @@ cuda_start(void)
void
cuda_poll(void)
{
- /* cuda_interrupt only takes a normal lock, we disable
- * interrupts here to avoid re-entering and thus deadlocking.
- */
- if (cuda_irq)
- disable_irq(cuda_irq);
- cuda_interrupt(0, NULL);
- if (cuda_irq)
- enable_irq(cuda_irq);
+ cuda_interrupt(0, NULL);
}
EXPORT_SYMBOL(cuda_poll);
@@ -475,13 +468,14 @@ EXPORT_SYMBOL(cuda_poll);
static irqreturn_t
cuda_interrupt(int irq, void *arg)
{
+ unsigned long flags;
u8 status;
struct adb_request *req = NULL;
unsigned char ibuf[16];
int ibuf_len = 0;
int complete = 0;
- spin_lock(&cuda_lock);
+ spin_lock_irqsave(&cuda_lock, flags);
/* On powermacs, this handler is registered for the VIA IRQ. But they use
* just the shift register IRQ -- other VIA interrupt sources are disabled.
@@ -494,7 +488,7 @@ cuda_interrupt(int irq, void *arg)
#endif
{
if ((in_8(&via[IFR]) & SR_INT) == 0) {
- spin_unlock(&cuda_lock);
+ spin_unlock_irqrestore(&cuda_lock, flags);
return IRQ_NONE;
} else {
out_8(&via[IFR], SR_INT);
@@ -616,7 +610,7 @@ cuda_interrupt(int irq, void *arg)
default:
pr_err("cuda_interrupt: unknown cuda_state %d?\n", cuda_state);
}
- spin_unlock(&cuda_lock);
+ spin_unlock_irqrestore(&cuda_lock, flags);
if (complete && req) {
void (*done)(struct adb_request *) = req->done;
mb();
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 04/10] via-cuda: Prevent read buffer overflow
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (3 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 07/10] via-cuda: Use spinlock_irq_save/restore instead of enable/disable_irq Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 03/10] via-cuda: Add TREQ, TIP and TACK signal helpers Finn Thain
` (4 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
If the Cuda driver does not enter the 'read_done' state for some
reason, it may continue in the 'reading' state until the buffer
overflows. Add a bounds check to prevent this.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index 64626d2..acf3a95 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -470,6 +470,8 @@ cuda_poll(void)
}
EXPORT_SYMBOL(cuda_poll);
+#define ARRAY_FULL(a, p) ((p) - (a) == ARRAY_SIZE(a))
+
static irqreturn_t
cuda_interrupt(int irq, void *arg)
{
@@ -558,7 +560,11 @@ cuda_interrupt(int irq, void *arg)
break;
case reading:
- *reply_ptr++ = in_8(&via[SR]);
+ if (reading_reply ? ARRAY_FULL(current_req->reply, reply_ptr)
+ : ARRAY_FULL(cuda_rbuf, reply_ptr))
+ (void)in_8(&via[SR]);
+ else
+ *reply_ptr++ = in_8(&via[SR]);
if (!TREQ_asserted(status)) {
/* that's all folks */
negate_TIP_and_TACK();
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 03/10] via-cuda: Add TREQ, TIP and TACK signal helpers
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (4 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 04/10] via-cuda: Prevent read buffer overflow Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 01/10] via-cuda: Cleanup printk calls Finn Thain
` (3 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
Introduce some helpers for handling the signalling between VIA and
Cuda. This abstraction will be used to add support for Egret devices,
which utilize slightly different signalling.
Don't invert the sense of the Cuda's active-low signals when storing
them in the 'status' variable. Just assert, negate and test those
signals using the helpers.
The state machine does not need to test its own output signals to
figure out what to do next: the next state depends on the Cuda's TREQ
output. Just call the TREQ_asserted() helper function to test for that.
Similarly, there is no need to store pin directions in the 'status'
variable. That was only useful for debugging messages.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 76 ++++++++++++++++++++++++++++++--------------
1 file changed, 53 insertions(+), 23 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index 4eb8a884..64626d2 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -65,6 +65,36 @@ static DEFINE_SPINLOCK(cuda_lock);
#define IER_CLR 0 /* clear bits in IER */
#define SR_INT 0x04 /* Shift register full/empty */
+static inline bool TREQ_asserted(u8 portb)
+{
+ return !(portb & TREQ);
+}
+
+static inline void assert_TIP(void)
+{
+ out_8(&via[B], in_8(&via[B]) & ~TIP);
+}
+
+static inline void assert_TACK(void)
+{
+ out_8(&via[B], in_8(&via[B]) & ~TACK);
+}
+
+static inline void toggle_TACK(void)
+{
+ out_8(&via[B], in_8(&via[B]) ^ TACK);
+}
+
+static inline void negate_TACK(void)
+{
+ out_8(&via[B], in_8(&via[B]) | TACK);
+}
+
+static inline void negate_TIP_and_TACK(void)
+{
+ out_8(&via[B], in_8(&via[B]) | TIP | TACK);
+}
+
static enum cuda_state {
idle,
sent_first_byte,
@@ -262,7 +292,7 @@ static int
__init cuda_init_via(void)
{
out_8(&via[DIRB], (in_8(&via[DIRB]) | TACK | TIP) & ~TREQ); /* TACK & TIP out */
- out_8(&via[B], in_8(&via[B]) | TACK | TIP); /* negate them */
+ negate_TIP_and_TACK();
out_8(&via[ACR] ,(in_8(&via[ACR]) & ~SR_CTRL) | SR_EXT); /* SR data in */
(void)in_8(&via[SR]); /* clear any left-over data */
#ifdef CONFIG_PPC
@@ -278,10 +308,10 @@ __init cuda_init_via(void)
out_8(&via[IFR], SR_INT);
/* sync with the CUDA - assert TACK without TIP */
- out_8(&via[B], in_8(&via[B]) & ~TACK);
+ assert_TACK();
/* wait for the CUDA to assert TREQ in response */
- WAIT_FOR((in_8(&via[B]) & TREQ) == 0, "CUDA response to sync");
+ WAIT_FOR(TREQ_asserted(in_8(&via[B])), "CUDA response to sync");
/* wait for the interrupt and then clear it */
WAIT_FOR(in_8(&via[IFR]) & SR_INT, "CUDA response to sync (2)");
@@ -289,14 +319,13 @@ __init cuda_init_via(void)
out_8(&via[IFR], SR_INT);
/* finish the sync by negating TACK */
- out_8(&via[B], in_8(&via[B]) | TACK);
+ negate_TACK();
/* wait for the CUDA to negate TREQ and the corresponding interrupt */
- WAIT_FOR(in_8(&via[B]) & TREQ, "CUDA response to sync (3)");
+ WAIT_FOR(!TREQ_asserted(in_8(&via[B])), "CUDA response to sync (3)");
WAIT_FOR(in_8(&via[IFR]) & SR_INT, "CUDA response to sync (4)");
(void)in_8(&via[SR]);
out_8(&via[IFR], SR_INT);
- out_8(&via[B], in_8(&via[B]) | TIP); /* should be unnecessary */
return 0;
}
@@ -417,13 +446,13 @@ cuda_start(void)
/* assert cuda_state == idle */
if (current_req == NULL)
return;
- if ((in_8(&via[B]) & TREQ) == 0)
+ if (TREQ_asserted(in_8(&via[B])))
return; /* a byte is coming in from the CUDA */
/* set the shift register to shift out and send a byte */
out_8(&via[ACR], in_8(&via[ACR]) | SR_OUT);
out_8(&via[SR], current_req->data[0]);
- out_8(&via[B], in_8(&via[B]) & ~TIP);
+ assert_TIP();
cuda_state = sent_first_byte;
}
@@ -444,7 +473,7 @@ EXPORT_SYMBOL(cuda_poll);
static irqreturn_t
cuda_interrupt(int irq, void *arg)
{
- int status;
+ u8 status;
struct adb_request *req = NULL;
unsigned char ibuf[16];
int ibuf_len = 0;
@@ -469,13 +498,14 @@ cuda_interrupt(int irq, void *arg)
out_8(&via[IFR], SR_INT);
}
}
-
- status = (~in_8(&via[B]) & (TIP|TREQ)) | (in_8(&via[ACR]) & SR_OUT);
+
+ status = in_8(&via[B]) & (TIP | TACK | TREQ);
+
switch (cuda_state) {
case idle:
/* CUDA has sent us the first byte of data - unsolicited */
(void)in_8(&via[SR]);
- out_8(&via[B], in_8(&via[B]) & ~TIP);
+ assert_TIP();
cuda_state = reading;
reply_ptr = cuda_rbuf;
reading_reply = 0;
@@ -484,22 +514,22 @@ cuda_interrupt(int irq, void *arg)
case awaiting_reply:
/* CUDA has sent us the first byte of data of a reply */
(void)in_8(&via[SR]);
- out_8(&via[B], in_8(&via[B]) & ~TIP);
+ assert_TIP();
cuda_state = reading;
reply_ptr = current_req->reply;
reading_reply = 1;
break;
case sent_first_byte:
- if (status == TREQ + TIP + SR_OUT) {
+ if (TREQ_asserted(status)) {
/* collision */
out_8(&via[ACR], in_8(&via[ACR]) & ~SR_OUT);
(void)in_8(&via[SR]);
- out_8(&via[B], in_8(&via[B]) | TIP | TACK);
+ negate_TIP_and_TACK();
cuda_state = idle;
} else {
out_8(&via[SR], current_req->data[1]);
- out_8(&via[B], in_8(&via[B]) ^ TACK);
+ toggle_TACK();
data_index = 2;
cuda_state = sending;
}
@@ -510,7 +540,7 @@ cuda_interrupt(int irq, void *arg)
if (data_index >= req->nbytes) {
out_8(&via[ACR], in_8(&via[ACR]) & ~SR_OUT);
(void)in_8(&via[SR]);
- out_8(&via[B], in_8(&via[B]) | TACK | TIP);
+ negate_TIP_and_TACK();
req->sent = 1;
if (req->reply_expected) {
cuda_state = awaiting_reply;
@@ -523,18 +553,18 @@ cuda_interrupt(int irq, void *arg)
}
} else {
out_8(&via[SR], req->data[data_index++]);
- out_8(&via[B], in_8(&via[B]) ^ TACK);
+ toggle_TACK();
}
break;
case reading:
*reply_ptr++ = in_8(&via[SR]);
- if (status == TIP) {
+ if (!TREQ_asserted(status)) {
/* that's all folks */
- out_8(&via[B], in_8(&via[B]) | TACK | TIP);
+ negate_TIP_and_TACK();
cuda_state = read_done;
} else {
- out_8(&via[B], in_8(&via[B]) ^ TACK);
+ toggle_TACK();
}
break;
@@ -567,8 +597,8 @@ cuda_interrupt(int irq, void *arg)
ibuf_len = reply_ptr - cuda_rbuf;
memcpy(ibuf, cuda_rbuf, ibuf_len);
}
- if (status == TREQ) {
- out_8(&via[B], in_8(&via[B]) & ~TIP);
+ if (TREQ_asserted(status)) {
+ assert_TIP();
cuda_state = reading;
reply_ptr = cuda_rbuf;
reading_reply = 0;
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 01/10] via-cuda: Cleanup printk calls
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (5 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 03/10] via-cuda: Add TREQ, TIP and TACK signal helpers Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 10/10] m68k/mac: Replace via-maciisi driver with via-cuda driver Finn Thain
` (2 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
Add missing log message severity, remove old debug messages and
replace printk() loop with print_hex_dump() call.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 26 ++++++--------------------
1 file changed, 6 insertions(+), 20 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index bad1813..98ced65 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -221,7 +221,7 @@ static int __init via_cuda_start(void)
return -EAGAIN;
}
- printk("Macintosh CUDA driver v0.5 for Unified ADB.\n");
+ pr_info("Macintosh CUDA driver v0.5 for Unified ADB.\n");
cuda_fully_inited = 1;
return 0;
@@ -251,7 +251,7 @@ cuda_probe(void)
int x; \
for (x = 1000; !(cond); --x) { \
if (x == 0) { \
- printk("Timeout waiting for " what "\n"); \
+ pr_err("Timeout waiting for " what "\n"); \
return -ENXIO; \
} \
udelay(100); \
@@ -357,6 +357,7 @@ cuda_reset_adb_bus(void)
return 0;
}
#endif /* CONFIG_ADB */
+
/* Construct and send a cuda request */
int
cuda_request(struct adb_request *req, void (*done)(struct adb_request *),
@@ -474,12 +475,9 @@ cuda_interrupt(int irq, void *arg)
}
status = (~in_8(&via[B]) & (TIP|TREQ)) | (in_8(&via[ACR]) & SR_OUT);
- /* printk("cuda_interrupt: state=%d status=%x\n", cuda_state, status); */
switch (cuda_state) {
case idle:
/* CUDA has sent us the first byte of data - unsolicited */
- if (status != TREQ)
- printk("cuda: state=idle, status=%x\n", status);
(void)in_8(&via[SR]);
out_8(&via[B], in_8(&via[B]) & ~TIP);
cuda_state = reading;
@@ -489,8 +487,6 @@ cuda_interrupt(int irq, void *arg)
case awaiting_reply:
/* CUDA has sent us the first byte of data of a reply */
- if (status != TREQ)
- printk("cuda: state=awaiting_reply, status=%x\n", status);
(void)in_8(&via[SR]);
out_8(&via[B], in_8(&via[B]) & ~TIP);
cuda_state = reading;
@@ -506,9 +502,6 @@ cuda_interrupt(int irq, void *arg)
out_8(&via[B], in_8(&via[B]) | TIP | TACK);
cuda_state = idle;
} else {
- /* assert status == TIP + SR_OUT */
- if (status != TIP + SR_OUT)
- printk("cuda: state=sent_first_byte status=%x\n", status);
out_8(&via[SR], current_req->data[1]);
out_8(&via[B], in_8(&via[B]) ^ TACK);
data_index = 2;
@@ -545,9 +538,6 @@ cuda_interrupt(int irq, void *arg)
out_8(&via[B], in_8(&via[B]) | TACK | TIP);
cuda_state = read_done;
} else {
- /* assert status == TIP | TREQ */
- if (status != TIP + TREQ)
- printk("cuda: state=reading status=%x\n", status);
out_8(&via[B], in_8(&via[B]) ^ TACK);
}
break;
@@ -593,7 +583,7 @@ cuda_interrupt(int irq, void *arg)
break;
default:
- printk("cuda_interrupt: unknown cuda_state %d?\n", cuda_state);
+ pr_err("cuda_interrupt: unknown cuda_state %d?\n", cuda_state);
}
spin_unlock(&cuda_lock);
if (complete && req) {
@@ -614,8 +604,6 @@ cuda_interrupt(int irq, void *arg)
static void
cuda_input(unsigned char *buf, int nb)
{
- int i;
-
switch (buf[0]) {
case ADB_PACKET:
#ifdef CONFIG_XMON
@@ -633,9 +621,7 @@ cuda_input(unsigned char *buf, int nb)
break;
default:
- printk("data from cuda (%d bytes):", nb);
- for (i = 0; i < nb; ++i)
- printk(" %.2x", buf[i]);
- printk("\n");
+ print_hex_dump(KERN_INFO, "cuda_input: ", DUMP_PREFIX_NONE, 32, 1,
+ buf, nb, false);
}
}
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 10/10] m68k/mac: Replace via-maciisi driver with via-cuda driver
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (6 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 01/10] via-cuda: Cleanup printk calls Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 08/10] via-cuda: Initialize data_index early and increment consistently Finn Thain
2016-12-01 6:09 ` [PATCH 09/10] via-cuda: Add support for Egret system controller Finn Thain
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
Change the device probe test in the via-cuda.c driver so it will load on
Egret-based machines too. Remove the now redundant via-maciisi.c driver.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
arch/m68k/include/asm/macintosh.h | 2 +-
arch/m68k/mac/config.c | 18 +-
arch/m68k/mac/misc.c | 72 +---
drivers/macintosh/Kconfig | 24 +-
drivers/macintosh/Makefile | 1 -
drivers/macintosh/adb.c | 4 -
drivers/macintosh/via-cuda.c | 8 +-
drivers/macintosh/via-maciisi.c | 677 --------------------------------------
8 files changed, 30 insertions(+), 776 deletions(-)
delete mode 100644 drivers/macintosh/via-maciisi.c
diff --git a/arch/m68k/include/asm/macintosh.h b/arch/m68k/include/asm/macintosh.h
index 42235e7..5b81ab1 100644
--- a/arch/m68k/include/asm/macintosh.h
+++ b/arch/m68k/include/asm/macintosh.h
@@ -38,7 +38,7 @@ struct mac_model
#define MAC_ADB_NONE 0
#define MAC_ADB_II 1
-#define MAC_ADB_IISI 2
+#define MAC_ADB_EGRET 2
#define MAC_ADB_CUDA 3
#define MAC_ADB_PB1 4
#define MAC_ADB_PB2 5
diff --git a/arch/m68k/mac/config.c b/arch/m68k/mac/config.c
index 689b47d..91b58d0 100644
--- a/arch/m68k/mac/config.c
+++ b/arch/m68k/mac/config.c
@@ -285,7 +285,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_IISI,
.name = "IIsi",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_OLD,
.scc_type = MAC_SCC_II,
@@ -294,7 +294,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_IIVI,
.name = "IIvi",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -303,7 +303,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_IIVX,
.name = "IIvx",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -318,7 +318,7 @@ static struct mac_model mac_data_table[] = {
{
.ident = MAC_MODEL_CLII,
.name = "Classic II",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -351,7 +351,7 @@ static struct mac_model mac_data_table[] = {
{
.ident = MAC_MODEL_LC,
.name = "LC",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -360,7 +360,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_LCII,
.name = "LC II",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -369,7 +369,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_LCIII,
.name = "LC III",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -497,7 +497,7 @@ static struct mac_model mac_data_table[] = {
{
.ident = MAC_MODEL_P460,
.name = "Performa 460",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
@@ -574,7 +574,7 @@ static struct mac_model mac_data_table[] = {
}, {
.ident = MAC_MODEL_P600,
.name = "Performa 600",
- .adb_type = MAC_ADB_IISI,
+ .adb_type = MAC_ADB_EGRET,
.via_type = MAC_VIA_IICI,
.scsi_type = MAC_SCSI_LC,
.scc_type = MAC_SCC_II,
diff --git a/arch/m68k/mac/misc.c b/arch/m68k/mac/misc.c
index 707b61a..c79570a 100644
--- a/arch/m68k/mac/misc.c
+++ b/arch/m68k/mac/misc.c
@@ -143,54 +143,6 @@ static void pmu_write_pram(int offset, __u8 data)
#define pmu_write_pram NULL
#endif
-#if 0 /* def CONFIG_ADB_MACIISI */
-extern int maciisi_request(struct adb_request *req,
- void (*done)(struct adb_request *), int nbytes, ...);
-
-static long maciisi_read_time(void)
-{
- struct adb_request req;
- long time;
-
- if (maciisi_request(&req, NULL, 2, CUDA_PACKET, CUDA_GET_TIME))
- return 0;
-
- time = (req.reply[3] << 24) | (req.reply[4] << 16)
- | (req.reply[5] << 8) | req.reply[6];
- return time - RTC_OFFSET;
-}
-
-static void maciisi_write_time(long data)
-{
- struct adb_request req;
- data += RTC_OFFSET;
- maciisi_request(&req, NULL, 6, CUDA_PACKET, CUDA_SET_TIME,
- (data >> 24) & 0xFF, (data >> 16) & 0xFF,
- (data >> 8) & 0xFF, data & 0xFF);
-}
-
-static __u8 maciisi_read_pram(int offset)
-{
- struct adb_request req;
- if (maciisi_request(&req, NULL, 4, CUDA_PACKET, CUDA_GET_PRAM,
- (offset >> 8) & 0xFF, offset & 0xFF))
- return 0;
- return req.reply[3];
-}
-
-static void maciisi_write_pram(int offset, __u8 data)
-{
- struct adb_request req;
- maciisi_request(&req, NULL, 5, CUDA_PACKET, CUDA_SET_PRAM,
- (offset >> 8) & 0xFF, offset & 0xFF, data);
-}
-#else
-#define maciisi_read_time() 0
-#define maciisi_write_time(n)
-#define maciisi_read_pram NULL
-#define maciisi_write_pram NULL
-#endif
-
/*
* VIA PRAM/RTC access routines
*
@@ -459,11 +411,10 @@ void mac_pram_read(int offset, __u8 *buffer, int len)
int i;
switch(macintosh_config->adb_type) {
- case MAC_ADB_IISI:
- func = maciisi_read_pram; break;
case MAC_ADB_PB1:
case MAC_ADB_PB2:
func = pmu_read_pram; break;
+ case MAC_ADB_EGRET:
case MAC_ADB_CUDA:
func = cuda_read_pram; break;
default:
@@ -482,11 +433,10 @@ void mac_pram_write(int offset, __u8 *buffer, int len)
int i;
switch(macintosh_config->adb_type) {
- case MAC_ADB_IISI:
- func = maciisi_write_pram; break;
case MAC_ADB_PB1:
case MAC_ADB_PB2:
func = pmu_write_pram; break;
+ case MAC_ADB_EGRET:
case MAC_ADB_CUDA:
func = cuda_write_pram; break;
default:
@@ -501,17 +451,13 @@ void mac_pram_write(int offset, __u8 *buffer, int len)
void mac_poweroff(void)
{
- /*
- * MAC_ADB_IISI may need to be moved up here if it doesn't actually
- * work using the ADB packet method. --David Kilzer
- */
-
if (oss_present) {
oss_shutdown();
} else if (macintosh_config->adb_type == MAC_ADB_II) {
via_shutdown();
#ifdef CONFIG_ADB_CUDA
- } else if (macintosh_config->adb_type == MAC_ADB_CUDA) {
+ } else if (macintosh_config->adb_type == MAC_ADB_EGRET ||
+ macintosh_config->adb_type == MAC_ADB_CUDA) {
cuda_shutdown();
#endif
#ifdef CONFIG_ADB_PMU68K
@@ -551,7 +497,8 @@ void mac_reset(void)
local_irq_restore(flags);
}
#ifdef CONFIG_ADB_CUDA
- } else if (macintosh_config->adb_type == MAC_ADB_CUDA) {
+ } else if (macintosh_config->adb_type == MAC_ADB_EGRET ||
+ macintosh_config->adb_type == MAC_ADB_CUDA) {
cuda_restart();
#endif
#ifdef CONFIG_ADB_PMU68K
@@ -700,13 +647,11 @@ int mac_hwclk(int op, struct rtc_time *t)
case MAC_ADB_IOP:
now = via_read_time();
break;
- case MAC_ADB_IISI:
- now = maciisi_read_time();
- break;
case MAC_ADB_PB1:
case MAC_ADB_PB2:
now = pmu_read_time();
break;
+ case MAC_ADB_EGRET:
case MAC_ADB_CUDA:
now = cuda_read_time();
break;
@@ -738,6 +683,7 @@ int mac_hwclk(int op, struct rtc_time *t)
case MAC_ADB_IOP:
via_write_time(now);
break;
+ case MAC_ADB_EGRET:
case MAC_ADB_CUDA:
cuda_write_time(now);
break;
@@ -745,8 +691,6 @@ int mac_hwclk(int op, struct rtc_time *t)
case MAC_ADB_PB2:
pmu_write_time(now);
break;
- case MAC_ADB_IISI:
- maciisi_write_time(now);
}
}
return 0;
diff --git a/drivers/macintosh/Kconfig b/drivers/macintosh/Kconfig
index 5844b80..b2f3588 100644
--- a/drivers/macintosh/Kconfig
+++ b/drivers/macintosh/Kconfig
@@ -30,14 +30,6 @@ config ADB_MACII
Quadra 610, Quadra 650, Quadra 700, Quadra 800, Centris 610 and
Centris 650.
-config ADB_MACIISI
- bool "Include Mac IIsi ADB driver"
- depends on ADB && MAC && BROKEN
- help
- Say Y here if want your kernel to support Macintosh systems that use
- the Mac IIsi style ADB. This includes the IIsi, IIvi, IIvx, Classic
- II, LC, LC II, LC III, Performa 460, and the Performa 600.
-
config ADB_IOP
bool "Include IOP (IIfx/Quadra 9x0) ADB driver"
depends on ADB && MAC
@@ -60,17 +52,15 @@ config ADB_PMU68K
# we want to change this to something like CONFIG_SYSCTRL_CUDA/PMU
config ADB_CUDA
- bool "Support for CUDA based Macs and PowerMacs"
+ bool "Support for Cuda/Egret based Macs and PowerMacs"
depends on (ADB || PPC_PMAC) && !PPC_PMAC64
help
- This provides support for CUDA based Macintosh and Power Macintosh
- systems. This includes many m68k based Macs (Color Classic, Mac TV,
- Performa 475, Performa 520, Performa 550, Performa 575,
- Performa 588, Quadra 605, Quadra 630, Quadra/Centris 660AV, and
- Quadra 840AV), most OldWorld PowerMacs, the first generation iMacs,
- the Blue&White G3 and the "Yikes" G4 (PCI Graphics). All later
- models should use CONFIG_ADB_PMU instead. It is safe to say Y here
- even if your machine doesn't have a CUDA.
+ This provides support for Cuda/Egret based Macintosh and
+ Power Macintosh systems. This includes most m68k based Macs,
+ most Old World PowerMacs, the first generation iMacs, the
+ Blue & White G3 and the "Yikes" G4 (PCI Graphics). All later
+ models should use CONFIG_ADB_PMU instead. It is safe to say Y
+ here even if your machine doesn't have a Cuda or Egret device.
If unsure say Y.
diff --git a/drivers/macintosh/Makefile b/drivers/macintosh/Makefile
index 383ba92..516eb65b 100644
--- a/drivers/macintosh/Makefile
+++ b/drivers/macintosh/Makefile
@@ -20,7 +20,6 @@ obj-$(CONFIG_PMAC_SMU) += smu.o
obj-$(CONFIG_ADB) += adb.o
obj-$(CONFIG_ADB_MACII) += via-macii.o
-obj-$(CONFIG_ADB_MACIISI) += via-maciisi.o
obj-$(CONFIG_ADB_IOP) += adb-iop.o
obj-$(CONFIG_ADB_PMU68K) += via-pmu68k.o
obj-$(CONFIG_ADB_MACIO) += macio-adb.o
diff --git a/drivers/macintosh/adb.c b/drivers/macintosh/adb.c
index 226179b..152414e 100644
--- a/drivers/macintosh/adb.c
+++ b/drivers/macintosh/adb.c
@@ -48,7 +48,6 @@
EXPORT_SYMBOL(adb_client_list);
extern struct adb_driver via_macii_driver;
-extern struct adb_driver via_maciisi_driver;
extern struct adb_driver via_cuda_driver;
extern struct adb_driver adb_iop_driver;
extern struct adb_driver via_pmu_driver;
@@ -59,9 +58,6 @@ static struct adb_driver *adb_driver_list[] = {
#ifdef CONFIG_ADB_MACII
&via_macii_driver,
#endif
-#ifdef CONFIG_ADB_MACIISI
- &via_maciisi_driver,
-#endif
#ifdef CONFIG_ADB_CUDA
&via_cuda_driver,
#endif
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index 3254070..20c81e7 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -205,12 +205,13 @@ int __init find_via_cuda(void)
struct adb_request req;
int err;
- if (macintosh_config->adb_type != MAC_ADB_CUDA)
+ if (macintosh_config->adb_type != MAC_ADB_CUDA &&
+ macintosh_config->adb_type != MAC_ADB_EGRET)
return 0;
via = via1;
cuda_state = idle;
- mcu_is_egret = false;
+ mcu_is_egret = macintosh_config->adb_type == MAC_ADB_EGRET;
err = cuda_init_via();
if (err) {
@@ -323,7 +324,8 @@ cuda_probe(void)
if (sys_ctrler != SYS_CTRLER_CUDA)
return -ENODEV;
#else
- if (macintosh_config->adb_type != MAC_ADB_CUDA)
+ if (macintosh_config->adb_type != MAC_ADB_CUDA &&
+ macintosh_config->adb_type != MAC_ADB_EGRET)
return -ENODEV;
#endif
if (via == NULL)
diff --git a/drivers/macintosh/via-maciisi.c b/drivers/macintosh/via-maciisi.c
deleted file mode 100644
index 34d02a9..0000000
--- a/drivers/macintosh/via-maciisi.c
+++ /dev/null
@@ -1,677 +0,0 @@
-/*
- * Device driver for the IIsi-style ADB on some Mac LC and II-class machines
- *
- * Based on via-cuda.c and via-macii.c, as well as the original
- * adb-bus.c, which in turn is somewhat influenced by (but uses no
- * code from) the NetBSD HWDIRECT ADB code. Original IIsi driver work
- * was done by Robert Thompson and integrated into the old style
- * driver by Michael Schmitz.
- *
- * Original sources (c) Alan Cox, Paul Mackerras, and others.
- *
- * Rewritten for Unified ADB by David Huggins-Daines <dhd@debian.org>
- *
- * 7/13/2000- extensive changes by Andrew McPherson <andrew@macduff.dhs.org>
- * Works about 30% of the time now.
- */
-
-#include <linux/types.h>
-#include <linux/errno.h>
-#include <linux/kernel.h>
-#include <linux/adb.h>
-#include <linux/cuda.h>
-#include <linux/delay.h>
-#include <linux/interrupt.h>
-#include <asm/macintosh.h>
-#include <asm/macints.h>
-#include <asm/mac_via.h>
-
-static volatile unsigned char *via;
-
-/* VIA registers - spaced 0x200 bytes apart - only the ones we actually use */
-#define RS 0x200 /* skip between registers */
-#define B 0 /* B-side data */
-#define A RS /* A-side data */
-#define DIRB (2*RS) /* B-side direction (1=output) */
-#define DIRA (3*RS) /* A-side direction (1=output) */
-#define SR (10*RS) /* Shift register */
-#define ACR (11*RS) /* Auxiliary control register */
-#define IFR (13*RS) /* Interrupt flag register */
-#define IER (14*RS) /* Interrupt enable register */
-
-/* Bits in B data register: all active low */
-#define TREQ 0x08 /* Transfer request (input) */
-#define TACK 0x10 /* Transfer acknowledge (output) */
-#define TIP 0x20 /* Transfer in progress (output) */
-#define ST_MASK 0x30 /* mask for selecting ADB state bits */
-
-/* Bits in ACR */
-#define SR_CTRL 0x1c /* Shift register control bits */
-#define SR_EXT 0x0c /* Shift on external clock */
-#define SR_OUT 0x10 /* Shift out if 1 */
-
-/* Bits in IFR and IER */
-#define IER_SET 0x80 /* set bits in IER */
-#define IER_CLR 0 /* clear bits in IER */
-#define SR_INT 0x04 /* Shift register full/empty */
-#define SR_DATA 0x08 /* Shift register data */
-#define SR_CLOCK 0x10 /* Shift register clock */
-
-#define ADB_DELAY 150
-
-#undef DEBUG_MACIISI_ADB
-
-static struct adb_request* current_req;
-static struct adb_request* last_req;
-static unsigned char maciisi_rbuf[16];
-static unsigned char *reply_ptr;
-static int data_index;
-static int reading_reply;
-static int reply_len;
-static int tmp;
-static int need_sync;
-
-static enum maciisi_state {
- idle,
- sending,
- reading,
-} maciisi_state;
-
-static int maciisi_probe(void);
-static int maciisi_init(void);
-static int maciisi_send_request(struct adb_request* req, int sync);
-static void maciisi_sync(struct adb_request *req);
-static int maciisi_write(struct adb_request* req);
-static irqreturn_t maciisi_interrupt(int irq, void* arg);
-static void maciisi_input(unsigned char *buf, int nb);
-static int maciisi_init_via(void);
-static void maciisi_poll(void);
-static int maciisi_start(void);
-
-struct adb_driver via_maciisi_driver = {
- "Mac IIsi",
- maciisi_probe,
- maciisi_init,
- maciisi_send_request,
- NULL, /* maciisi_adb_autopoll, */
- maciisi_poll,
- NULL /* maciisi_reset_adb_bus */
-};
-
-static int
-maciisi_probe(void)
-{
- if (macintosh_config->adb_type != MAC_ADB_IISI)
- return -ENODEV;
-
- via = via1;
- return 0;
-}
-
-static int
-maciisi_init(void)
-{
- int err;
-
- if (via == NULL)
- return -ENODEV;
-
- if ((err = maciisi_init_via())) {
- printk(KERN_ERR "maciisi_init: maciisi_init_via() failed, code %d\n", err);
- via = NULL;
- return err;
- }
-
- if (request_irq(IRQ_MAC_ADB, maciisi_interrupt, 0, "ADB",
- maciisi_interrupt)) {
- printk(KERN_ERR "maciisi_init: can't get irq %d\n", IRQ_MAC_ADB);
- return -EAGAIN;
- }
-
- printk("adb: Mac IIsi driver v0.2 for Unified ADB.\n");
- return 0;
-}
-
-/* Flush data from the ADB controller */
-static void
-maciisi_stfu(void)
-{
- int status = via[B] & (TIP|TREQ);
-
- if (status & TREQ) {
-#ifdef DEBUG_MACIISI_ADB
- printk (KERN_DEBUG "maciisi_stfu called with TREQ high!\n");
-#endif
- return;
- }
-
- udelay(ADB_DELAY);
- via[ACR] &= ~SR_OUT;
- via[IER] = IER_CLR | SR_INT;
-
- udelay(ADB_DELAY);
-
- status = via[B] & (TIP|TREQ);
-
- if (!(status & TREQ))
- {
- via[B] |= TIP;
-
- while(1)
- {
- int poll_timeout = ADB_DELAY * 5;
- /* Poll for SR interrupt */
- while (!(via[IFR] & SR_INT) && poll_timeout-- > 0)
- status = via[B] & (TIP|TREQ);
-
- tmp = via[SR]; /* Clear shift register */
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_stfu: status %x timeout %d data %x\n",
- status, poll_timeout, tmp);
-#endif
- if(via[B] & TREQ)
- break;
-
- /* ACK on-off */
- via[B] |= TACK;
- udelay(ADB_DELAY);
- via[B] &= ~TACK;
- }
-
- /* end frame */
- via[B] &= ~TIP;
- udelay(ADB_DELAY);
- }
-
- via[IER] = IER_SET | SR_INT;
-}
-
-/* All specifically VIA-related initialization goes here */
-static int
-maciisi_init_via(void)
-{
- int i;
-
- /* Set the lines up. We want TREQ as input TACK|TIP as output */
- via[DIRB] = (via[DIRB] | TACK | TIP) & ~TREQ;
- /* Shift register on input */
- via[ACR] = (via[ACR] & ~SR_CTRL) | SR_EXT;
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_init_via: initial status %x\n", via[B] & (TIP|TREQ));
-#endif
- /* Wipe any pending data and int */
- tmp = via[SR];
- /* Enable keyboard interrupts */
- via[IER] = IER_SET | SR_INT;
- /* Set initial state: idle */
- via[B] &= ~(TACK|TIP);
- /* Clear interrupt bit */
- via[IFR] = SR_INT;
-
- for(i = 0; i < 60; i++) {
- udelay(ADB_DELAY);
- maciisi_stfu();
- udelay(ADB_DELAY);
- if(via[B] & TREQ)
- break;
- }
- if (i == 60)
- printk(KERN_ERR "maciisi_init_via: bus jam?\n");
-
- maciisi_state = idle;
- need_sync = 0;
-
- return 0;
-}
-
-/* Send a request, possibly waiting for a reply */
-static int
-maciisi_send_request(struct adb_request* req, int sync)
-{
- int i;
-
-#ifdef DEBUG_MACIISI_ADB
- static int dump_packet = 0;
-#endif
-
- if (via == NULL) {
- req->complete = 1;
- return -ENXIO;
- }
-
-#ifdef DEBUG_MACIISI_ADB
- if (dump_packet) {
- printk(KERN_DEBUG "maciisi_send_request:");
- for (i = 0; i < req->nbytes; i++) {
- printk(" %.2x", req->data[i]);
- }
- printk(" sync %d\n", sync);
- }
-#endif
-
- req->reply_expected = 1;
-
- i = maciisi_write(req);
- if (i)
- {
- /* Normally, if a packet requires syncing, that happens at the end of
- * maciisi_send_request. But if the transfer fails, it will be restarted
- * by maciisi_interrupt(). We use need_sync to tell maciisi_interrupt
- * when to sync a packet that it sends out.
- *
- * Suggestions on a better way to do this are welcome.
- */
- if(i == -EBUSY && sync)
- need_sync = 1;
- else
- need_sync = 0;
- return i;
- }
- if(sync)
- maciisi_sync(req);
-
- return 0;
-}
-
-/* Poll the ADB chip until the request completes */
-static void maciisi_sync(struct adb_request *req)
-{
- int count = 0;
-
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_sync called\n");
-#endif
-
- /* If for some reason the ADB chip shuts up on us, we want to avoid an endless loop. */
- while (!req->complete && count++ < 50) {
- maciisi_poll();
- }
- /* This could be BAD... when the ADB controller doesn't respond
- * for this long, it's probably not coming back :-( */
- if (count > 50) /* Hopefully shouldn't happen */
- printk(KERN_ERR "maciisi_send_request: poll timed out!\n");
-}
-
-int
-maciisi_request(struct adb_request *req, void (*done)(struct adb_request *),
- int nbytes, ...)
-{
- va_list list;
- int i;
-
- req->nbytes = nbytes;
- req->done = done;
- req->reply_expected = 0;
- va_start(list, nbytes);
- for (i = 0; i < nbytes; i++)
- req->data[i++] = va_arg(list, int);
- va_end(list);
-
- return maciisi_send_request(req, 1);
-}
-
-/* Enqueue a request, and run the queue if possible */
-static int
-maciisi_write(struct adb_request* req)
-{
- unsigned long flags;
- int i;
-
- /* We will accept CUDA packets - the VIA sends them to us, so
- it figures that we should be able to send them to it */
- if (req->nbytes < 2 || req->data[0] > CUDA_PACKET) {
- printk(KERN_ERR "maciisi_write: packet too small or not an ADB or CUDA packet\n");
- req->complete = 1;
- return -EINVAL;
- }
- req->next = NULL;
- req->sent = 0;
- req->complete = 0;
- req->reply_len = 0;
-
- local_irq_save(flags);
-
- if (current_req) {
- last_req->next = req;
- last_req = req;
- } else {
- current_req = req;
- last_req = req;
- }
- if (maciisi_state == idle)
- {
- i = maciisi_start();
- if(i != 0)
- {
- local_irq_restore(flags);
- return i;
- }
- }
- else
- {
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_write: would start, but state is %d\n", maciisi_state);
-#endif
- local_irq_restore(flags);
- return -EBUSY;
- }
-
- local_irq_restore(flags);
-
- return 0;
-}
-
-static int
-maciisi_start(void)
-{
- struct adb_request* req;
- int status;
-
-#ifdef DEBUG_MACIISI_ADB
- status = via[B] & (TIP | TREQ);
-
- printk(KERN_DEBUG "maciisi_start called, state=%d, status=%x, ifr=%x\n", maciisi_state, status, via[IFR]);
-#endif
-
- if (maciisi_state != idle) {
- /* shouldn't happen */
- printk(KERN_ERR "maciisi_start: maciisi_start called when driver busy!\n");
- return -EBUSY;
- }
-
- req = current_req;
- if (req == NULL)
- return -EINVAL;
-
- status = via[B] & (TIP|TREQ);
- if (!(status & TREQ)) {
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_start: bus busy - aborting\n");
-#endif
- return -EBUSY;
- }
-
- /* Okay, send */
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "maciisi_start: sending\n");
-#endif
- /* Set state to active */
- via[B] |= TIP;
- /* ACK off */
- via[B] &= ~TACK;
- /* Delay */
- udelay(ADB_DELAY);
- /* Shift out and send */
- via[ACR] |= SR_OUT;
- via[SR] = req->data[0];
- data_index = 1;
- /* ACK on */
- via[B] |= TACK;
- maciisi_state = sending;
-
- return 0;
-}
-
-void
-maciisi_poll(void)
-{
- unsigned long flags;
-
- local_irq_save(flags);
- if (via[IFR] & SR_INT) {
- maciisi_interrupt(0, NULL);
- }
- else /* avoid calling this function too quickly in a loop */
- udelay(ADB_DELAY);
-
- local_irq_restore(flags);
-}
-
-/* Shift register interrupt - this is *supposed* to mean that the
- register is either full or empty. In practice, I have no idea what
- it means :( */
-static irqreturn_t
-maciisi_interrupt(int irq, void* arg)
-{
- int status;
- struct adb_request *req;
-#ifdef DEBUG_MACIISI_ADB
- static int dump_reply = 0;
-#endif
- int i;
- unsigned long flags;
-
- local_irq_save(flags);
-
- status = via[B] & (TIP|TREQ);
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "state %d status %x ifr %x\n", maciisi_state, status, via[IFR]);
-#endif
-
- if (!(via[IFR] & SR_INT)) {
- /* Shouldn't happen, we hope */
- printk(KERN_ERR "maciisi_interrupt: called without interrupt flag set\n");
- local_irq_restore(flags);
- return IRQ_NONE;
- }
-
- /* Clear the interrupt */
- /* via[IFR] = SR_INT; */
-
- switch_start:
- switch (maciisi_state) {
- case idle:
- if (status & TIP)
- printk(KERN_ERR "maciisi_interrupt: state is idle but TIP asserted!\n");
-
- if(!reading_reply)
- udelay(ADB_DELAY);
- /* Shift in */
- via[ACR] &= ~SR_OUT;
- /* Signal start of frame */
- via[B] |= TIP;
- /* Clear the interrupt (throw this value on the floor, it's useless) */
- tmp = via[SR];
- /* ACK adb chip, high-low */
- via[B] |= TACK;
- udelay(ADB_DELAY);
- via[B] &= ~TACK;
- reply_len = 0;
- maciisi_state = reading;
- if (reading_reply) {
- reply_ptr = current_req->reply;
- } else {
- reply_ptr = maciisi_rbuf;
- }
- break;
-
- case sending:
- /* via[SR]; */
- /* Set ACK off */
- via[B] &= ~TACK;
- req = current_req;
-
- if (!(status & TREQ)) {
- /* collision */
- printk(KERN_ERR "maciisi_interrupt: send collision\n");
- /* Set idle and input */
- via[ACR] &= ~SR_OUT;
- tmp = via[SR];
- via[B] &= ~TIP;
- /* Must re-send */
- reading_reply = 0;
- reply_len = 0;
- maciisi_state = idle;
- udelay(ADB_DELAY);
- /* process this now, because the IFR has been cleared */
- goto switch_start;
- }
-
- udelay(ADB_DELAY);
-
- if (data_index >= req->nbytes) {
- /* Sent the whole packet, put the bus back in idle state */
- /* Shift in, we are about to read a reply (hopefully) */
- via[ACR] &= ~SR_OUT;
- tmp = via[SR];
- /* End of frame */
- via[B] &= ~TIP;
- req->sent = 1;
- maciisi_state = idle;
- if (req->reply_expected) {
- /* Note: only set this once we've
- successfully sent the packet */
- reading_reply = 1;
- } else {
- current_req = req->next;
- if (req->done)
- (*req->done)(req);
- /* Do any queued requests now */
- i = maciisi_start();
- if(i == 0 && need_sync) {
- /* Packet needs to be synced */
- maciisi_sync(current_req);
- }
- if(i != -EBUSY)
- need_sync = 0;
- }
- } else {
- /* Sending more stuff */
- /* Shift out */
- via[ACR] |= SR_OUT;
- /* Write */
- via[SR] = req->data[data_index++];
- /* Signal 'byte ready' */
- via[B] |= TACK;
- }
- break;
-
- case reading:
- /* Shift in */
- /* via[ACR] &= ~SR_OUT; */ /* Not in 2.2 */
- if (reply_len++ > 16) {
- printk(KERN_ERR "maciisi_interrupt: reply too long, aborting read\n");
- via[B] |= TACK;
- udelay(ADB_DELAY);
- via[B] &= ~(TACK|TIP);
- maciisi_state = idle;
- i = maciisi_start();
- if(i == 0 && need_sync) {
- /* Packet needs to be synced */
- maciisi_sync(current_req);
- }
- if(i != -EBUSY)
- need_sync = 0;
- break;
- }
- /* Read data */
- *reply_ptr++ = via[SR];
- status = via[B] & (TIP|TREQ);
- /* ACK on/off */
- via[B] |= TACK;
- udelay(ADB_DELAY);
- via[B] &= ~TACK;
- if (!(status & TREQ))
- break; /* more stuff to deal with */
-
- /* end of frame */
- via[B] &= ~TIP;
- tmp = via[SR]; /* That's what happens in 2.2 */
- udelay(ADB_DELAY); /* Give controller time to recover */
-
- /* end of packet, deal with it */
- if (reading_reply) {
- req = current_req;
- req->reply_len = reply_ptr - req->reply;
- if (req->data[0] == ADB_PACKET) {
- /* Have to adjust the reply from ADB commands */
- if (req->reply_len <= 2 || (req->reply[1] & 2) != 0) {
- /* the 0x2 bit indicates no response */
- req->reply_len = 0;
- } else {
- /* leave just the command and result bytes in the reply */
- req->reply_len -= 2;
- memmove(req->reply, req->reply + 2, req->reply_len);
- }
- }
-#ifdef DEBUG_MACIISI_ADB
- if (dump_reply) {
- int i;
- printk(KERN_DEBUG "maciisi_interrupt: reply is ");
- for (i = 0; i < req->reply_len; ++i)
- printk(" %.2x", req->reply[i]);
- printk("\n");
- }
-#endif
- req->complete = 1;
- current_req = req->next;
- if (req->done)
- (*req->done)(req);
- /* Obviously, we got it */
- reading_reply = 0;
- } else {
- maciisi_input(maciisi_rbuf, reply_ptr - maciisi_rbuf);
- }
- maciisi_state = idle;
- status = via[B] & (TIP|TREQ);
- if (!(status & TREQ)) {
- /* Timeout?! More likely, another packet coming in already */
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "extra data after packet: status %x ifr %x\n",
- status, via[IFR]);
-#endif
-#if 0
- udelay(ADB_DELAY);
- via[B] |= TIP;
-
- maciisi_state = reading;
- reading_reply = 0;
- reply_ptr = maciisi_rbuf;
-#else
- /* Process the packet now */
- reading_reply = 0;
- goto switch_start;
-#endif
- /* We used to do this... but the controller might actually have data for us */
- /* maciisi_stfu(); */
- }
- else {
- /* Do any queued requests now if possible */
- i = maciisi_start();
- if(i == 0 && need_sync) {
- /* Packet needs to be synced */
- maciisi_sync(current_req);
- }
- if(i != -EBUSY)
- need_sync = 0;
- }
- break;
-
- default:
- printk("maciisi_interrupt: unknown maciisi_state %d?\n", maciisi_state);
- }
- local_irq_restore(flags);
- return IRQ_HANDLED;
-}
-
-static void
-maciisi_input(unsigned char *buf, int nb)
-{
-#ifdef DEBUG_MACIISI_ADB
- int i;
-#endif
-
- switch (buf[0]) {
- case ADB_PACKET:
- adb_input(buf+2, nb-2, buf[1] & 0x40);
- break;
- default:
-#ifdef DEBUG_MACIISI_ADB
- printk(KERN_DEBUG "data from IIsi ADB (%d bytes):", nb);
- for (i = 0; i < nb; ++i)
- printk(" %.2x", buf[i]);
- printk("\n");
-#endif
- break;
- }
-}
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 08/10] via-cuda: Initialize data_index early and increment consistently
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (7 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 10/10] m68k/mac: Replace via-maciisi driver with via-cuda driver Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
2016-12-01 6:09 ` [PATCH 09/10] via-cuda: Add support for Egret system controller Finn Thain
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
Initialize data_index where appropriate to improve readability and
assist debugging. This change doesn't affect driver behaviour.
I prefer to see
current_req->data[data_index++]
in place of
current_req->data[0]
or
current_req->data[1]
inasmuchas it becomes obvious what the data_index variable does.
Moreover, the actual value of data_index when examined at any given moment
tells me something about prior events, which did prove helpful.
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index b801c67..f1b2c5b 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -446,12 +446,13 @@ cuda_start(void)
/* assert cuda_state == idle */
if (current_req == NULL)
return;
+ data_index = 0;
if (TREQ_asserted(in_8(&via[B])))
return; /* a byte is coming in from the CUDA */
/* set the shift register to shift out and send a byte */
out_8(&via[ACR], in_8(&via[ACR]) | SR_OUT);
- out_8(&via[SR], current_req->data[0]);
+ out_8(&via[SR], current_req->data[data_index++]);
assert_TIP();
cuda_state = sent_first_byte;
}
@@ -524,9 +525,8 @@ cuda_interrupt(int irq, void *arg)
negate_TIP_and_TACK();
cuda_state = idle;
} else {
- out_8(&via[SR], current_req->data[1]);
+ out_8(&via[SR], current_req->data[data_index++]);
toggle_TACK();
- data_index = 2;
cuda_state = sending;
}
break;
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread* [PATCH 09/10] via-cuda: Add support for Egret system controller
2016-12-01 6:09 [PATCH 00/10] Replace via-maciisi with via-cuda driver Finn Thain
` (8 preceding siblings ...)
2016-12-01 6:09 ` [PATCH 08/10] via-cuda: Initialize data_index early and increment consistently Finn Thain
@ 2016-12-01 6:09 ` Finn Thain
9 siblings, 0 replies; 11+ messages in thread
From: Finn Thain @ 2016-12-01 6:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt, Geert Uytterhoeven
Cc: linux-m68k, linuxppc-dev, linux-kernel
The Egret system controller was the predecessor to the Cuda and the
differences are minor.
On Cuda, byte acknowledgement requires one transition of the TACK
signal; on Egret two are needed. On Cuda, TIP is active low; on Egret
it is active high. And Cuda raises certain interrupts that Egret omits.
Accomodating these differences complicates the Cuda driver slightly
but avoids a lot of duplication (see next patch).
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
---
drivers/macintosh/via-cuda.c | 155 +++++++++++++++++++++++++++++++++++++------
1 file changed, 134 insertions(+), 21 deletions(-)
diff --git a/drivers/macintosh/via-cuda.c b/drivers/macintosh/via-cuda.c
index f1b2c5b..3254070 100644
--- a/drivers/macintosh/via-cuda.c
+++ b/drivers/macintosh/via-cuda.c
@@ -1,10 +1,10 @@
/*
- * Device driver for the via-cuda on Apple Powermacs.
+ * Device driver for the Cuda and Egret system controllers found on PowerMacs
+ * and 68k Macs.
*
- * The VIA (versatile interface adapter) interfaces to the CUDA,
- * a 6805 microprocessor core which controls the ADB (Apple Desktop
- * Bus) which connects to the keyboard and mouse. The CUDA also
- * controls system power and the RTC (real time clock) chip.
+ * The Cuda or Egret is a 6805 microcontroller interfaced to the 6522 VIA.
+ * This MCU controls system power, Parameter RAM, Real Time Clock and the
+ * Apple Desktop Bus (ADB) that connects to the keyboard and mouse.
*
* Copyright (C) 1996 Paul Mackerras.
*/
@@ -50,10 +50,27 @@ static DEFINE_SPINLOCK(cuda_lock);
#define IER (14*RS) /* Interrupt enable register */
#define ANH (15*RS) /* A-side data, no handshake */
-/* Bits in B data register: all active low */
-#define TREQ 0x08 /* Transfer request (input) */
-#define TACK 0x10 /* Transfer acknowledge (output) */
-#define TIP 0x20 /* Transfer in progress (output) */
+/*
+ * When the Cuda design replaced the Egret, some signal names and
+ * logic sense changed. They all serve the same purposes, however.
+ *
+ * VIA pin | Egret pin
+ * ----------------+------------------------------------------
+ * PB3 (input) | Transceiver session (active low)
+ * PB4 (output) | VIA full (active high)
+ * PB5 (output) | System session (active high)
+ *
+ * VIA pin | Cuda pin
+ * ----------------+------------------------------------------
+ * PB3 (input) | Transfer request (active low)
+ * PB4 (output) | Byte acknowledge (active low)
+ * PB5 (output) | Transfer in progress (active low)
+ */
+
+/* Bits in Port B data register */
+#define TREQ 0x08 /* Transfer request */
+#define TACK 0x10 /* Transfer acknowledge */
+#define TIP 0x20 /* Transfer in progress */
/* Bits in ACR */
#define SR_CTRL 0x1c /* Shift register control bits */
@@ -65,6 +82,19 @@ static DEFINE_SPINLOCK(cuda_lock);
#define IER_CLR 0 /* clear bits in IER */
#define SR_INT 0x04 /* Shift register full/empty */
+/* Duration of byte acknowledgement pulse (us) */
+#define EGRET_TACK_ASSERTED_DELAY 300
+#define EGRET_TACK_NEGATED_DELAY 400
+
+/* Interval from interrupt to start of session (us) */
+#define EGRET_SESSION_DELAY 450
+
+#ifdef CONFIG_PPC
+#define mcu_is_egret false
+#else
+static bool mcu_is_egret;
+#endif
+
static inline bool TREQ_asserted(u8 portb)
{
return !(portb & TREQ);
@@ -72,12 +102,29 @@ static inline bool TREQ_asserted(u8 portb)
static inline void assert_TIP(void)
{
- out_8(&via[B], in_8(&via[B]) & ~TIP);
+ if (mcu_is_egret) {
+ udelay(EGRET_SESSION_DELAY);
+ out_8(&via[B], in_8(&via[B]) | TIP);
+ } else
+ out_8(&via[B], in_8(&via[B]) & ~TIP);
+}
+
+static inline void assert_TIP_and_TACK(void)
+{
+ if (mcu_is_egret) {
+ udelay(EGRET_SESSION_DELAY);
+ out_8(&via[B], in_8(&via[B]) | TIP | TACK);
+ } else
+ out_8(&via[B], in_8(&via[B]) & ~(TIP | TACK));
}
static inline void assert_TACK(void)
{
- out_8(&via[B], in_8(&via[B]) & ~TACK);
+ if (mcu_is_egret) {
+ udelay(EGRET_TACK_NEGATED_DELAY);
+ out_8(&via[B], in_8(&via[B]) | TACK);
+ } else
+ out_8(&via[B], in_8(&via[B]) & ~TACK);
}
static inline void toggle_TACK(void)
@@ -87,12 +134,20 @@ static inline void toggle_TACK(void)
static inline void negate_TACK(void)
{
- out_8(&via[B], in_8(&via[B]) | TACK);
+ if (mcu_is_egret) {
+ udelay(EGRET_TACK_ASSERTED_DELAY);
+ out_8(&via[B], in_8(&via[B]) & ~TACK);
+ } else
+ out_8(&via[B], in_8(&via[B]) | TACK);
}
static inline void negate_TIP_and_TACK(void)
{
- out_8(&via[B], in_8(&via[B]) | TIP | TACK);
+ if (mcu_is_egret) {
+ udelay(EGRET_TACK_ASSERTED_DELAY);
+ out_8(&via[B], in_8(&via[B]) & ~(TIP | TACK));
+ } else
+ out_8(&via[B], in_8(&via[B]) | TIP | TACK);
}
static enum cuda_state {
@@ -155,6 +210,7 @@ int __init find_via_cuda(void)
via = via1;
cuda_state = idle;
+ mcu_is_egret = false;
err = cuda_init_via();
if (err) {
@@ -251,7 +307,7 @@ static int __init via_cuda_start(void)
return -EAGAIN;
}
- pr_info("Macintosh CUDA driver v0.5 for Unified ADB.\n");
+ pr_info("Macintosh Cuda and Egret driver.\n");
cuda_fully_inited = 1;
return 0;
@@ -276,6 +332,33 @@ cuda_probe(void)
}
#endif /* CONFIG_ADB */
+static int __init sync_egret(void)
+{
+ if (TREQ_asserted(in_8(&via[B]))) {
+ /* Complete the inbound transfer */
+ assert_TIP_and_TACK();
+ while (1) {
+ negate_TACK();
+ mdelay(1);
+ (void)in_8(&via[SR]);
+ assert_TACK();
+ if (!TREQ_asserted(in_8(&via[B])))
+ break;
+ }
+ negate_TIP_and_TACK();
+ } else if (in_8(&via[B]) & TIP) {
+ /* Terminate the outbound transfer */
+ negate_TACK();
+ assert_TACK();
+ mdelay(1);
+ negate_TIP_and_TACK();
+ }
+ /* Clear shift register interrupt */
+ if (in_8(&via[IFR]) & SR_INT)
+ (void)in_8(&via[SR]);
+ return 0;
+}
+
#define WAIT_FOR(cond, what) \
do { \
int x; \
@@ -291,10 +374,6 @@ cuda_probe(void)
static int
__init cuda_init_via(void)
{
- out_8(&via[DIRB], (in_8(&via[DIRB]) | TACK | TIP) & ~TREQ); /* TACK & TIP out */
- negate_TIP_and_TACK();
- out_8(&via[ACR] ,(in_8(&via[ACR]) & ~SR_CTRL) | SR_EXT); /* SR data in */
- (void)in_8(&via[SR]); /* clear any left-over data */
#ifdef CONFIG_PPC
out_8(&via[IER], 0x7f); /* disable interrupts from VIA */
(void)in_8(&via[IER]);
@@ -302,6 +381,15 @@ __init cuda_init_via(void)
out_8(&via[IER], SR_INT); /* disable SR interrupt from VIA */
#endif
+ out_8(&via[DIRB], (in_8(&via[DIRB]) | TACK | TIP) & ~TREQ); /* TACK & TIP out */
+ out_8(&via[ACR], (in_8(&via[ACR]) & ~SR_CTRL) | SR_EXT); /* SR data in */
+ (void)in_8(&via[SR]); /* clear any left-over data */
+
+ if (mcu_is_egret)
+ return sync_egret();
+
+ negate_TIP_and_TACK();
+
/* delay 4ms and then clear any pending interrupt */
mdelay(4);
(void)in_8(&via[SR]);
@@ -453,7 +541,10 @@ cuda_start(void)
/* set the shift register to shift out and send a byte */
out_8(&via[ACR], in_8(&via[ACR]) | SR_OUT);
out_8(&via[SR], current_req->data[data_index++]);
- assert_TIP();
+ if (mcu_is_egret)
+ assert_TIP_and_TACK();
+ else
+ assert_TIP();
cuda_state = sent_first_byte;
}
@@ -500,8 +591,9 @@ cuda_interrupt(int irq, void *arg)
switch (cuda_state) {
case idle:
- /* CUDA has sent us the first byte of data - unsolicited */
+ /* System controller has unsolicited data for us */
(void)in_8(&via[SR]);
+idle_state:
assert_TIP();
cuda_state = reading;
reply_ptr = cuda_rbuf;
@@ -509,7 +601,7 @@ cuda_interrupt(int irq, void *arg)
break;
case awaiting_reply:
- /* CUDA has sent us the first byte of data of a reply */
+ /* System controller has reply data for us */
(void)in_8(&via[SR]);
assert_TIP();
cuda_state = reading;
@@ -524,9 +616,14 @@ cuda_interrupt(int irq, void *arg)
(void)in_8(&via[SR]);
negate_TIP_and_TACK();
cuda_state = idle;
+ /* Egret does not raise an "aborted" interrupt */
+ if (mcu_is_egret)
+ goto idle_state;
} else {
out_8(&via[SR], current_req->data[data_index++]);
toggle_TACK();
+ if (mcu_is_egret)
+ assert_TACK();
cuda_state = sending;
}
break;
@@ -550,6 +647,8 @@ cuda_interrupt(int irq, void *arg)
} else {
out_8(&via[SR], req->data[data_index++]);
toggle_TACK();
+ if (mcu_is_egret)
+ assert_TACK();
}
break;
@@ -560,16 +659,24 @@ cuda_interrupt(int irq, void *arg)
else
*reply_ptr++ = in_8(&via[SR]);
if (!TREQ_asserted(status)) {
+ if (mcu_is_egret)
+ assert_TACK();
/* that's all folks */
negate_TIP_and_TACK();
cuda_state = read_done;
+ /* Egret does not raise a "read done" interrupt */
+ if (mcu_is_egret)
+ goto read_done_state;
} else {
toggle_TACK();
+ if (mcu_is_egret)
+ negate_TACK();
}
break;
case read_done:
(void)in_8(&via[SR]);
+read_done_state:
if (reading_reply) {
req = current_req;
req->reply_len = reply_ptr - req->reply;
@@ -645,6 +752,12 @@ cuda_input(unsigned char *buf, int nb)
#endif /* CONFIG_ADB */
break;
+ case TIMER_PACKET:
+ /* Egret sends these periodically. Might be useful as a 'heartbeat'
+ * to trigger a recovery for the VIA shift register errata.
+ */
+ break;
+
default:
print_hex_dump(KERN_INFO, "cuda_input: ", DUMP_PREFIX_NONE, 32, 1,
buf, nb, false);
--
2.7.3
^ permalink raw reply related [flat|nested] 11+ messages in thread