patch-2.4.20 linux-2.4.20/drivers/hil/hil_mlc.c
Next file: linux-2.4.20/drivers/hil/hil_ptr.c
Previous file: linux-2.4.20/drivers/hil/hil_kbd.c
Back to the patch index
Back to the overall index
- Lines: 952
- Date:
Thu Nov 28 15:53:12 2002
- Orig file:
linux-2.4.19/drivers/hil/hil_mlc.c
- Orig date:
Wed Dec 31 16:00:00 1969
diff -urN linux-2.4.19/drivers/hil/hil_mlc.c linux-2.4.20/drivers/hil/hil_mlc.c
@@ -0,0 +1,951 @@
+/*
+ * HIL MLC state machine and serio interface driver
+ *
+ * Copyright (c) 2001 Brian S. Julin
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions, and the following disclaimer,
+ * without modification.
+ * 2. The name of the author may not be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL").
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ *
+ * References:
+ * HP-HIL Technical Reference Manual. Hewlett Packard Product No. 45918A
+ *
+ *
+ * Driver theory of operation:
+ *
+ * Some access methods and an ISR is defined by the sub-driver
+ * (e.g. hp_sdc_mlc.c). These methods are expected to provide a
+ * few bits of logic in addition to raw access to the HIL MLC,
+ * specifically, the ISR, which is entirely registered by the
+ * sub-driver and invoked directly, must check for record
+ * termination or packet match, at which point a semaphore must
+ * be cleared and then the hil_mlcs_tasklet must be scheduled.
+ *
+ * The hil_mlcs_tasklet processes the state machine for all MLCs
+ * each time it runs, checking each MLC's progress at the current
+ * node in the state machine, and moving the MLC to subsequent nodes
+ * in the state machine when appropriate. It will reschedule
+ * itself if output is pending. (This rescheduling should be replaced
+ * at some point with a sub-driver-specific mechanism.)
+ *
+ * A timer task prods the tasket once per second to prevent
+ * hangups when attached devices do not return expected data
+ * and to initiate probes of the loop for new devices.
+ */
+
+#include <linux/hil_mlc.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/timer.h>
+#include <linux/sched.h>
+#include <linux/list.h>
+
+MODULE_AUTHOR("Brian S. Julin <bri@calyx.com>");
+MODULE_DESCRIPTION("HIL MLC serio");
+MODULE_LICENSE("Dual BSD/GPL");
+
+EXPORT_SYMBOL(hil_mlc_register);
+EXPORT_SYMBOL(hil_mlc_unregister);
+
+#define PREFIX "HIL MLC: "
+
+static LIST_HEAD(hil_mlcs);
+static rwlock_t hil_mlcs_lock = RW_LOCK_UNLOCKED;
+static struct timer_list hil_mlcs_kicker;
+static int hil_mlcs_probe;
+
+static void hil_mlcs_process(unsigned long unused);
+DECLARE_TASKLET_DISABLED(hil_mlcs_tasklet, hil_mlcs_process, 0);
+
+
+/* #define HIL_MLC_DEBUG */
+
+/********************** Device info/instance management **********************/
+
+static void hil_mlc_clear_di_map (hil_mlc *mlc, int val) {
+ int j;
+ for (j = val; j < 7 ; j++) {
+ mlc->di_map[j] = -1;
+ }
+}
+
+static void hil_mlc_clear_di_scratch (hil_mlc *mlc) {
+ memset(&(mlc->di_scratch), 0, sizeof(mlc->di_scratch));
+}
+
+static void hil_mlc_copy_di_scratch (hil_mlc *mlc, int idx) {
+ memcpy(&(mlc->di[idx]), &(mlc->di_scratch), sizeof(mlc->di_scratch));
+}
+
+static int hil_mlc_match_di_scratch (hil_mlc *mlc) {
+ int idx;
+
+ for (idx = 0; idx < HIL_MLC_DEVMEM; idx++) {
+ int j, found;
+
+ /* In-use slots are not eligible. */
+ found = 0;
+ for (j = 0; j < 7 ; j++) {
+ if (mlc->di_map[j] == idx) found++;
+ }
+ if (found) continue;
+ if (!memcmp(mlc->di + idx,
+ &(mlc->di_scratch),
+ sizeof(mlc->di_scratch))) break;
+ }
+ return((idx >= HIL_MLC_DEVMEM) ? -1 : idx);
+}
+
+static int hil_mlc_find_free_di(hil_mlc *mlc) {
+ int idx;
+ /* TODO: Pick all-zero slots first, failing that,
+ * randomize the slot picked among those eligible.
+ */
+ for (idx = 0; idx < HIL_MLC_DEVMEM; idx++) {
+ int j, found;
+ found = 0;
+ for (j = 0; j < 7 ; j++) {
+ if (mlc->di_map[j] == idx) found++;
+ }
+ if (!found) break;
+ }
+ return(idx); /* Note: It is guaranteed at least one above will match */
+}
+
+static inline void hil_mlc_clean_serio_map(hil_mlc *mlc) {
+ int idx;
+ for (idx = 0; idx < HIL_MLC_DEVMEM; idx++) {
+ int j, found;
+ found = 0;
+ for (j = 0; j < 7 ; j++) {
+ if (mlc->di_map[j] == idx) found++;
+ }
+ if (!found) mlc->serio_map[idx].di_revmap = -1;
+ }
+}
+
+static void hil_mlc_send_polls(hil_mlc *mlc) {
+ int did, i, cnt;
+ struct serio *serio;
+ struct serio_dev *dev;
+
+ i = cnt = 0;
+ did = (mlc->ipacket[0] & HIL_PKT_ADDR_MASK) >> 8;
+ serio = did ? &(mlc->serio[mlc->di_map[did - 1]]) : NULL;
+ dev = (serio != NULL) ? serio->dev : NULL;
+
+ while (mlc->icount < 15 - i) {
+ hil_packet p;
+ p = mlc->ipacket[i];
+ if (did != (p & HIL_PKT_ADDR_MASK) >> 8) {
+ if (dev == NULL || dev->interrupt == NULL) goto skip;
+
+ dev->interrupt(serio, 0, 0);
+ dev->interrupt(serio, HIL_ERR_INT >> 16, 0);
+ dev->interrupt(serio, HIL_PKT_CMD >> 8, 0);
+ dev->interrupt(serio, HIL_CMD_POL + cnt, 0);
+ skip:
+ did = (p & HIL_PKT_ADDR_MASK) >> 8;
+ serio = did ? &(mlc->serio[mlc->di_map[did-1]]) : NULL;
+ dev = (serio != NULL) ? serio->dev : NULL;
+ cnt = 0;
+ }
+ cnt++; i++;
+ if (dev == NULL || dev->interrupt == NULL) continue;
+ dev->interrupt(serio, (p >> 24), 0);
+ dev->interrupt(serio, (p >> 16) & 0xff, 0);
+ dev->interrupt(serio, (p >> 8) & ~HIL_PKT_ADDR_MASK, 0);
+ dev->interrupt(serio, p & 0xff, 0);
+ }
+}
+
+/*************************** State engine *********************************/
+
+#define HILSEN_SCHED 0x000100 /* Schedule the tasklet */
+#define HILSEN_BREAK 0x000200 /* Wait until next pass */
+#define HILSEN_UP 0x000400 /* relative node#, decrement */
+#define HILSEN_DOWN 0x000800 /* relative node#, increment */
+#define HILSEN_FOLLOW 0x001000 /* use retval as next node# */
+
+#define HILSEN_MASK 0x0000ff
+#define HILSEN_START 0
+#define HILSEN_RESTART 1
+#define HILSEN_DHR 9
+#define HILSEN_DHR2 10
+#define HILSEN_IFC 14
+#define HILSEN_HEAL0 16
+#define HILSEN_HEAL 18
+#define HILSEN_ACF 21
+#define HILSEN_ACF2 22
+#define HILSEN_DISC0 25
+#define HILSEN_DISC 27
+#define HILSEN_MATCH 40
+#define HILSEN_OPERATE 41
+#define HILSEN_PROBE 44
+#define HILSEN_DSR 52
+#define HILSEN_REPOLL 55
+#define HILSEN_IFCACF 58
+#define HILSEN_END 60
+
+#define HILSEN_NEXT (HILSEN_DOWN | 1)
+#define HILSEN_SAME (HILSEN_DOWN | 0)
+#define HILSEN_LAST (HILSEN_UP | 1)
+
+#define HILSEN_DOZE (HILSEN_SAME | HILSEN_SCHED | HILSEN_BREAK)
+#define HILSEN_SLEEP (HILSEN_SAME | HILSEN_BREAK)
+
+static int hilse_match(hil_mlc *mlc, int unused) {
+ int rc;
+ rc = hil_mlc_match_di_scratch(mlc);
+ if (rc == -1) {
+ rc = hil_mlc_find_free_di(mlc);
+ if (rc == -1) goto err;
+#ifdef HIL_MLC_DEBUG
+ printk(KERN_DEBUG PREFIX "new in slot %i\n", rc);
+#endif
+ hil_mlc_copy_di_scratch(mlc, rc);
+ mlc->di_map[mlc->ddi] = rc;
+ mlc->serio_map[rc].di_revmap = mlc->ddi;
+ hil_mlc_clean_serio_map(mlc);
+ serio_rescan(mlc->serio + rc);
+ return -1;
+ }
+ mlc->di_map[mlc->ddi] = rc;
+#ifdef HIL_MLC_DEBUG
+ printk(KERN_DEBUG PREFIX "same in slot %i\n", rc);
+#endif
+ mlc->serio_map[rc].di_revmap = mlc->ddi;
+ hil_mlc_clean_serio_map(mlc);
+ return 0;
+ err:
+ printk(KERN_ERR PREFIX "Residual device slots exhausted, close some serios!\n");
+ return 1;
+}
+
+/* An LCV used to prevent runaway loops, forces 5 second sleep when reset. */
+static int hilse_init_lcv(hil_mlc *mlc, int unused) {
+ struct timeval tv;
+
+ do_gettimeofday(&tv);
+
+ if(mlc->lcv == 0) goto restart; /* First init, no need to dally */
+ if(tv.tv_sec - mlc->lcv_tv.tv_sec < 5) return -1;
+ restart:
+ mlc->lcv_tv = tv;
+ mlc->lcv = 0;
+ return 0;
+}
+
+static int hilse_inc_lcv(hil_mlc *mlc, int lim) {
+ if (mlc->lcv++ >= lim) return -1;
+ return 0;
+}
+
+#if 0
+static int hilse_set_lcv(hil_mlc *mlc, int val) {
+ mlc->lcv = val;
+ return 0;
+}
+#endif
+
+/* Management of the discovered device index (zero based, -1 means no devs) */
+static int hilse_set_ddi(hil_mlc *mlc, int val) {
+ mlc->ddi = val;
+ hil_mlc_clear_di_map(mlc, val + 1);
+ return 0;
+}
+
+static int hilse_dec_ddi(hil_mlc *mlc, int unused) {
+ mlc->ddi--;
+ if (mlc->ddi <= -1) {
+ mlc->ddi = -1;
+ hil_mlc_clear_di_map(mlc, 0);
+ return -1;
+ }
+ hil_mlc_clear_di_map(mlc, mlc->ddi + 1);
+ return 0;
+}
+
+static int hilse_inc_ddi(hil_mlc *mlc, int unused) {
+ if (mlc->ddi >= 6) {
+ BUG();
+ return -1;
+ }
+ mlc->ddi++;
+ return 0;
+}
+
+static int hilse_take_idd(hil_mlc *mlc, int unused) {
+ int i;
+
+ /* Help the state engine:
+ * Is this a real IDD response or just an echo?
+ *
+ * Real IDD response does not start with a command.
+ */
+ if (mlc->ipacket[0] & HIL_PKT_CMD) goto bail;
+ /* Should have the command echoed further down. */
+ for (i = 1; i < 16; i++) {
+ if (((mlc->ipacket[i] & HIL_PKT_ADDR_MASK) ==
+ (mlc->ipacket[0] & HIL_PKT_ADDR_MASK)) &&
+ (mlc->ipacket[i] & HIL_PKT_CMD) &&
+ ((mlc->ipacket[i] & HIL_PKT_DATA_MASK) == HIL_CMD_IDD))
+ break;
+ }
+ if (i > 15) goto bail;
+ /* And the rest of the packets should still be clear. */
+ while (++i < 16) {
+ if (mlc->ipacket[i]) break;
+ }
+ if (i < 16) goto bail;
+ for (i = 0; i < 16; i++) {
+ mlc->di_scratch.idd[i] =
+ mlc->ipacket[i] & HIL_PKT_DATA_MASK;
+ }
+ /* Next step is to see if RSC supported */
+ if (mlc->di_scratch.idd[1] & HIL_IDD_HEADER_RSC)
+ return HILSEN_NEXT;
+ if (mlc->di_scratch.idd[1] & HIL_IDD_HEADER_EXD)
+ return HILSEN_DOWN | 4;
+ return 0;
+ bail:
+ mlc->ddi--;
+ return -1; /* This should send us off to ACF */
+}
+
+static int hilse_take_rsc(hil_mlc *mlc, int unused) {
+ int i;
+
+ for (i = 0; i < 16; i++) {
+ mlc->di_scratch.rsc[i] =
+ mlc->ipacket[i] & HIL_PKT_DATA_MASK;
+ }
+ /* Next step is to see if EXD supported (IDD has already been read) */
+ if (mlc->di_scratch.idd[1] & HIL_IDD_HEADER_EXD)
+ return HILSEN_NEXT;
+ return 0;
+}
+
+static int hilse_take_exd(hil_mlc *mlc, int unused) {
+ int i;
+
+ for (i = 0; i < 16; i++) {
+ mlc->di_scratch.exd[i] =
+ mlc->ipacket[i] & HIL_PKT_DATA_MASK;
+ }
+ /* Next step is to see if RNM supported. */
+ if (mlc->di_scratch.exd[0] & HIL_EXD_HEADER_RNM)
+ return HILSEN_NEXT;
+ return 0;
+}
+
+static int hilse_take_rnm(hil_mlc *mlc, int unused) {
+ int i;
+
+ for (i = 0; i < 16; i++) {
+ mlc->di_scratch.rnm[i] =
+ mlc->ipacket[i] & HIL_PKT_DATA_MASK;
+ }
+ do {
+ char nam[17];
+ snprintf(nam, 16, "%s", mlc->di_scratch.rnm);
+ nam[16] = '\0';
+ printk(KERN_INFO PREFIX "Device name gotten: %s\n", nam);
+ } while (0);
+ return 0;
+}
+
+static int hilse_operate(hil_mlc *mlc, int repoll) {
+
+ if (mlc->opercnt == 0) hil_mlcs_probe = 0;
+ mlc->opercnt = 1;
+
+ hil_mlc_send_polls(mlc);
+
+ if (!hil_mlcs_probe) return 0;
+ hil_mlcs_probe = 0;
+ mlc->opercnt = 0;
+ return 1;
+}
+
+#define FUNC(funct, funct_arg, zero_rc, neg_rc, pos_rc) \
+{ HILSE_FUNC, { func: &funct }, funct_arg, zero_rc, neg_rc, pos_rc },
+#define OUT(pack) \
+{ HILSE_OUT, { packet: pack }, 0, HILSEN_NEXT, HILSEN_DOZE, 0 },
+#define CTS \
+{ HILSE_CTS, { packet: 0 }, 0, HILSEN_NEXT | HILSEN_SCHED | HILSEN_BREAK, HILSEN_DOZE, 0 },
+#define EXPECT(comp, to, got, got_wrong, timed_out) \
+{ HILSE_EXPECT, { packet: comp }, to, got, got_wrong, timed_out },
+#define EXPECT_LAST(comp, to, got, got_wrong, timed_out) \
+{ HILSE_EXPECT_LAST, { packet: comp }, to, got, got_wrong, timed_out },
+#define EXPECT_DISC(comp, to, got, got_wrong, timed_out) \
+{ HILSE_EXPECT_DISC, { packet: comp }, to, got, got_wrong, timed_out },
+#define IN(to, got, got_error, timed_out) \
+{ HILSE_IN, { packet: 0 }, to, got, got_error, timed_out },
+#define OUT_DISC(pack) \
+{ HILSE_OUT_DISC, { packet: pack }, 0, 0, 0, 0 },
+#define OUT_LAST(pack) \
+{ HILSE_OUT_LAST, { packet: pack }, 0, 0, 0, 0 },
+
+struct hilse_node hil_mlc_se[HILSEN_END] = {
+
+ /* 0 HILSEN_START */
+ FUNC(hilse_init_lcv, 0, HILSEN_NEXT, HILSEN_SLEEP, 0)
+
+ /* 1 HILSEN_RESTART */
+ FUNC(hilse_inc_lcv, 10, HILSEN_NEXT, HILSEN_START, 0)
+ OUT(HIL_CTRL_ONLY) /* Disable APE */
+ CTS
+
+#define TEST_PACKET(x) \
+(HIL_PKT_CMD | (x << HIL_PKT_ADDR_SHIFT) | x << 4 | x)
+
+ OUT(HIL_DO_ALTER_CTRL | HIL_CTRL_TEST | TEST_PACKET(0x5))
+ EXPECT(HIL_ERR_INT | TEST_PACKET(0x5),
+ 2000, HILSEN_NEXT, HILSEN_RESTART, HILSEN_RESTART)
+ OUT(HIL_DO_ALTER_CTRL | HIL_CTRL_TEST | TEST_PACKET(0xa))
+ EXPECT(HIL_ERR_INT | TEST_PACKET(0xa),
+ 2000, HILSEN_NEXT, HILSEN_RESTART, HILSEN_RESTART)
+ OUT(HIL_CTRL_ONLY | 0) /* Disable test mode */
+
+ /* 9 HILSEN_DHR */
+ FUNC(hilse_init_lcv, 0, HILSEN_NEXT, HILSEN_SLEEP, 0)
+
+ /* 10 HILSEN_DHR2 */
+ FUNC(hilse_inc_lcv, 10, HILSEN_NEXT, HILSEN_START, 0)
+ FUNC(hilse_set_ddi, -1, HILSEN_NEXT, 0, 0)
+ OUT(HIL_PKT_CMD | HIL_CMD_DHR)
+ IN(300000, HILSEN_DHR2, HILSEN_DHR2, HILSEN_NEXT)
+
+ /* 14 HILSEN_IFC */
+ OUT(HIL_PKT_CMD | HIL_CMD_IFC)
+ EXPECT(HIL_PKT_CMD | HIL_CMD_IFC | HIL_ERR_INT,
+ 20000, HILSEN_DISC, HILSEN_DHR2, HILSEN_NEXT )
+
+ /* If devices are there, they weren't in PUP or other loopback mode.
+ * We're more concerned at this point with restoring operation
+ * to devices than discovering new ones, so we try to salvage
+ * the loop configuration by closing off the loop.
+ */
+
+ /* 16 HILSEN_HEAL0 */
+ FUNC(hilse_dec_ddi, 0, HILSEN_NEXT, HILSEN_ACF, 0)
+ FUNC(hilse_inc_ddi, 0, HILSEN_NEXT, 0, 0)
+
+ /* 18 HILSEN_HEAL */
+ OUT_LAST(HIL_CMD_ELB)
+ EXPECT_LAST(HIL_CMD_ELB | HIL_ERR_INT,
+ 20000, HILSEN_REPOLL, HILSEN_DSR, HILSEN_NEXT)
+ FUNC(hilse_dec_ddi, 0, HILSEN_HEAL, HILSEN_NEXT, 0)
+
+ /* 21 HILSEN_ACF */
+ FUNC(hilse_init_lcv, 0, HILSEN_NEXT, HILSEN_DOZE, 0)
+
+ /* 22 HILSEN_ACF2 */
+ FUNC(hilse_inc_lcv, 10, HILSEN_NEXT, HILSEN_START, 0)
+ OUT(HIL_PKT_CMD | HIL_CMD_ACF | 1)
+ IN(20000, HILSEN_NEXT, HILSEN_DSR, HILSEN_NEXT)
+
+ /* 25 HILSEN_DISC0 */
+ OUT_DISC(HIL_PKT_CMD | HIL_CMD_ELB)
+ EXPECT_DISC(HIL_PKT_CMD | HIL_CMD_ELB | HIL_ERR_INT,
+ 20000, HILSEN_NEXT, HILSEN_DSR, HILSEN_DSR)
+
+ /* Only enter here if response just received */
+ /* 27 HILSEN_DISC */
+ OUT_DISC(HIL_PKT_CMD | HIL_CMD_IDD)
+ EXPECT_DISC(HIL_PKT_CMD | HIL_CMD_IDD | HIL_ERR_INT,
+ 20000, HILSEN_NEXT, HILSEN_DSR, HILSEN_START)
+ FUNC(hilse_inc_ddi, 0, HILSEN_NEXT, HILSEN_START, 0)
+ FUNC(hilse_take_idd, 0, HILSEN_MATCH, HILSEN_IFCACF, HILSEN_FOLLOW)
+ OUT_LAST(HIL_PKT_CMD | HIL_CMD_RSC)
+ EXPECT_LAST(HIL_PKT_CMD | HIL_CMD_RSC | HIL_ERR_INT,
+ 30000, HILSEN_NEXT, HILSEN_DSR, HILSEN_DSR)
+ FUNC(hilse_take_rsc, 0, HILSEN_MATCH, 0, HILSEN_FOLLOW)
+ OUT_LAST(HIL_PKT_CMD | HIL_CMD_EXD)
+ EXPECT_LAST(HIL_PKT_CMD | HIL_CMD_EXD | HIL_ERR_INT,
+ 30000, HILSEN_NEXT, HILSEN_DSR, HILSEN_DSR)
+ FUNC(hilse_take_exd, 0, HILSEN_MATCH, 0, HILSEN_FOLLOW)
+ OUT_LAST(HIL_PKT_CMD | HIL_CMD_RNM)
+ EXPECT_LAST(HIL_PKT_CMD | HIL_CMD_RNM | HIL_ERR_INT,
+ 30000, HILSEN_NEXT, HILSEN_DSR, HILSEN_DSR)
+ FUNC(hilse_take_rnm, 0, HILSEN_MATCH, 0, 0)
+
+ /* 40 HILSEN_MATCH */
+ FUNC(hilse_match, 0, HILSEN_NEXT, HILSEN_NEXT, /* TODO */ 0)
+
+ /* 41 HILSEN_OPERATE */
+ OUT(HIL_PKT_CMD | HIL_CMD_POL)
+ EXPECT(HIL_PKT_CMD | HIL_CMD_POL | HIL_ERR_INT,
+ 20000, HILSEN_NEXT, HILSEN_DSR, HILSEN_NEXT)
+ FUNC(hilse_operate, 0, HILSEN_OPERATE, HILSEN_IFC, HILSEN_NEXT)
+
+ /* 44 HILSEN_PROBE */
+ OUT_LAST(HIL_PKT_CMD | HIL_CMD_EPT)
+ IN(10000, HILSEN_DISC, HILSEN_DSR, HILSEN_NEXT)
+ OUT_DISC(HIL_PKT_CMD | HIL_CMD_ELB)
+ IN(10000, HILSEN_DISC, HILSEN_DSR, HILSEN_NEXT)
+ OUT(HIL_PKT_CMD | HIL_CMD_ACF | 1)
+ IN(10000, HILSEN_DISC0, HILSEN_DSR, HILSEN_NEXT)
+ OUT_LAST(HIL_PKT_CMD | HIL_CMD_ELB)
+ IN(10000, HILSEN_OPERATE, HILSEN_DSR, HILSEN_DSR)
+
+ /* 52 HILSEN_DSR */
+ FUNC(hilse_set_ddi, -1, HILSEN_NEXT, 0, 0)
+ OUT(HIL_PKT_CMD | HIL_CMD_DSR)
+ IN(20000, HILSEN_DHR, HILSEN_DHR, HILSEN_IFC)
+
+ /* 55 HILSEN_REPOLL */
+ OUT(HIL_PKT_CMD | HIL_CMD_RPL)
+ EXPECT(HIL_PKT_CMD | HIL_CMD_RPL | HIL_ERR_INT,
+ 20000, HILSEN_NEXT, HILSEN_DSR, HILSEN_NEXT)
+ FUNC(hilse_operate, 1, HILSEN_OPERATE, HILSEN_IFC, HILSEN_PROBE)
+
+ /* 58 HILSEN_IFCACF */
+ OUT(HIL_PKT_CMD | HIL_CMD_IFC)
+ EXPECT(HIL_PKT_CMD | HIL_CMD_IFC | HIL_ERR_INT,
+ 20000, HILSEN_ACF2, HILSEN_DHR2, HILSEN_HEAL)
+
+ /* 60 HILSEN_END */
+};
+
+static inline void hilse_setup_input(hil_mlc *mlc, struct hilse_node *node) {
+
+ switch (node->act) {
+ case HILSE_EXPECT_DISC:
+ mlc->imatch = node->object.packet;
+ mlc->imatch |= ((mlc->ddi + 2) << HIL_PKT_ADDR_SHIFT);
+ break;
+ case HILSE_EXPECT_LAST:
+ mlc->imatch = node->object.packet;
+ mlc->imatch |= ((mlc->ddi + 1) << HIL_PKT_ADDR_SHIFT);
+ break;
+ case HILSE_EXPECT:
+ mlc->imatch = node->object.packet;
+ break;
+ case HILSE_IN:
+ mlc->imatch = 0;
+ break;
+ default:
+ BUG();
+ }
+ mlc->istarted = 1;
+ mlc->intimeout = node->arg;
+ do_gettimeofday(&(mlc->instart));
+ mlc->icount = 15;
+ memset(mlc->ipacket, 0, 16 * sizeof(hil_packet));
+ if (down_trylock(&(mlc->isem))) BUG();
+
+ return;
+}
+
+#ifdef HIL_MLC_DEBUG
+static int doze = 0;
+static int seidx; /* For debug */
+static int kick = 1;
+#endif
+
+static int hilse_donode (hil_mlc *mlc) {
+ struct hilse_node *node;
+ int nextidx = 0;
+ int sched_long = 0;
+ unsigned long flags;
+
+#ifdef HIL_MLC_DEBUG
+ if (mlc->seidx && (mlc->seidx != seidx) && mlc->seidx != 41 && mlc->seidx != 42 && mlc->seidx != 43) {
+ printk(KERN_DEBUG PREFIX "z%i \n%s {%i}", doze, kick ? "K" : "", mlc->seidx);
+ doze = 0;
+ }
+ kick = 0;
+
+ seidx = mlc->seidx;
+#endif
+ node = hil_mlc_se + mlc->seidx;
+
+ switch (node->act) {
+ int rc;
+ hil_packet pack;
+
+ case HILSE_FUNC:
+ if (node->object.func == NULL) break;
+ rc = node->object.func(mlc, node->arg);
+ nextidx = (rc > 0) ? node->ugly :
+ ((rc < 0) ? node->bad : node->good);
+ if (nextidx == HILSEN_FOLLOW) nextidx = rc;
+ break;
+ case HILSE_EXPECT_LAST:
+ case HILSE_EXPECT_DISC:
+ case HILSE_EXPECT:
+ case HILSE_IN:
+ /* Already set up from previous HILSE_OUT_* */
+ write_lock_irqsave(&(mlc->lock), flags);
+ rc = mlc->in(mlc, node->arg);
+ if (rc == 2) {
+ nextidx = HILSEN_DOZE;
+ sched_long = 1;
+ write_unlock_irqrestore(&(mlc->lock), flags);
+ break;
+ }
+ if (rc == 1) nextidx = node->ugly;
+ else if (rc == 0) nextidx = node->good;
+ else nextidx = node->bad;
+ mlc->istarted = 0;
+ write_unlock_irqrestore(&(mlc->lock), flags);
+ break;
+ case HILSE_OUT_LAST:
+ write_lock_irqsave(&(mlc->lock), flags);
+ pack = node->object.packet;
+ pack |= ((mlc->ddi + 1) << HIL_PKT_ADDR_SHIFT);
+ goto out;
+ case HILSE_OUT_DISC:
+ write_lock_irqsave(&(mlc->lock), flags);
+ pack = node->object.packet;
+ pack |= ((mlc->ddi + 2) << HIL_PKT_ADDR_SHIFT);
+ goto out;
+ case HILSE_OUT:
+ write_lock_irqsave(&(mlc->lock), flags);
+ pack = node->object.packet;
+ out:
+ if (mlc->istarted) goto out2;
+ /* Prepare to receive input */
+ if ((node + 1)->act & HILSE_IN)
+ hilse_setup_input(mlc, node + 1);
+
+ out2:
+ write_unlock_irqrestore(&(mlc->lock), flags);
+
+ if (down_trylock(&mlc->osem)) {
+ nextidx = HILSEN_DOZE;
+ break;
+ }
+ up(&mlc->osem);
+
+ write_lock_irqsave(&(mlc->lock), flags);
+ if (!(mlc->ostarted)) {
+ mlc->ostarted = 1;
+ mlc->opacket = pack;
+ mlc->out(mlc);
+ nextidx = HILSEN_DOZE;
+ write_unlock_irqrestore(&(mlc->lock), flags);
+ break;
+ }
+ mlc->ostarted = 0;
+ do_gettimeofday(&(mlc->instart));
+ write_unlock_irqrestore(&(mlc->lock), flags);
+ nextidx = HILSEN_NEXT;
+ break;
+ case HILSE_CTS:
+ nextidx = mlc->cts(mlc) ? node->bad : node->good;
+ break;
+ default:
+ BUG();
+ nextidx = 0;
+ break;
+ }
+
+#ifdef HIL_MLC_DEBUG
+ if (nextidx == HILSEN_DOZE) doze++;
+#endif
+
+ while (nextidx & HILSEN_SCHED) {
+ struct timeval tv;
+
+ if (!sched_long) goto sched;
+
+ do_gettimeofday(&tv);
+ tv.tv_usec += 1000000 * (tv.tv_sec - mlc->instart.tv_sec);
+ tv.tv_usec -= mlc->instart.tv_usec;
+ if (tv.tv_usec >= mlc->intimeout) goto sched;
+ tv.tv_usec = (mlc->intimeout - tv.tv_usec) * HZ / 1000000;
+ if (!tv.tv_usec) goto sched;
+ mod_timer(&hil_mlcs_kicker, jiffies + tv.tv_usec);
+ break;
+ sched:
+ tasklet_schedule(&hil_mlcs_tasklet);
+ break;
+ }
+ if (nextidx & HILSEN_DOWN) mlc->seidx += nextidx & HILSEN_MASK;
+ else if (nextidx & HILSEN_UP) mlc->seidx -= nextidx & HILSEN_MASK;
+ else mlc->seidx = nextidx & HILSEN_MASK;
+
+ if (nextidx & HILSEN_BREAK) return 1;
+ return 0;
+}
+
+/******************** tasklet context functions **************************/
+static void hil_mlcs_process(unsigned long unused) {
+ struct list_head *tmp;
+
+ read_lock(&hil_mlcs_lock);
+ list_for_each(tmp, &hil_mlcs) {
+ struct hil_mlc *mlc = list_entry(tmp, hil_mlc, list);
+ while (hilse_donode(mlc) == 0) {
+#ifdef HIL_MLC_DEBUG
+ if (mlc->seidx != 41 &&
+ mlc->seidx != 42 &&
+ mlc->seidx != 43)
+ printk(KERN_DEBUG PREFIX " + ");
+#endif
+ };
+ }
+ read_unlock(&hil_mlcs_lock);
+}
+
+/************************* Keepalive timer task *********************/
+
+void hil_mlcs_timer (unsigned long data) {
+ hil_mlcs_probe = 1;
+ tasklet_schedule(&hil_mlcs_tasklet);
+ /* Re-insert the periodic task. */
+ if (!timer_pending(&hil_mlcs_kicker))
+ mod_timer(&hil_mlcs_kicker, jiffies + HZ);
+}
+
+/******************** user/kernel context functions **********************/
+
+static int hil_mlc_serio_write(struct serio *serio, unsigned char c) {
+ struct hil_mlc_serio_map *map;
+ struct hil_mlc *mlc;
+ struct serio_dev *dev;
+ uint8_t *idx, *last;
+
+ map = serio->driver;
+ if (map == NULL) {
+ BUG();
+ return -EIO;
+ }
+ mlc = map->mlc;
+ if (mlc == NULL) {
+ BUG();
+ return -EIO;
+ }
+ mlc->serio_opacket[map->didx] |=
+ ((hil_packet)c) << (8 * (3 - mlc->serio_oidx[map->didx]));
+
+ if (mlc->serio_oidx[map->didx] >= 3) {
+ /* for now only commands */
+ if (!(mlc->serio_opacket[map->didx] & HIL_PKT_CMD))
+ return -EIO;
+ switch (mlc->serio_opacket[map->didx] & HIL_PKT_DATA_MASK) {
+ case HIL_CMD_IDD:
+ idx = mlc->di[map->didx].idd;
+ goto emu;
+ case HIL_CMD_RSC:
+ idx = mlc->di[map->didx].rsc;
+ goto emu;
+ case HIL_CMD_EXD:
+ idx = mlc->di[map->didx].exd;
+ goto emu;
+ case HIL_CMD_RNM:
+ idx = mlc->di[map->didx].rnm;
+ goto emu;
+ default:
+ break;
+ }
+ mlc->serio_oidx[map->didx] = 0;
+ mlc->serio_opacket[map->didx] = 0;
+ }
+
+ mlc->serio_oidx[map->didx]++;
+ return -EIO;
+ emu:
+ dev = serio->dev;
+ if (dev == NULL) {
+ BUG();
+ return -EIO;
+ }
+ last = idx + 15;
+ while ((last != idx) && (*last == 0)) last--;
+
+ while (idx != last) {
+ dev->interrupt(serio, 0, 0);
+ dev->interrupt(serio, HIL_ERR_INT >> 16, 0);
+ dev->interrupt(serio, 0, 0);
+ dev->interrupt(serio, *idx, 0);
+ idx++;
+ }
+ dev->interrupt(serio, 0, 0);
+ dev->interrupt(serio, HIL_ERR_INT >> 16, 0);
+ dev->interrupt(serio, HIL_PKT_CMD >> 8, 0);
+ dev->interrupt(serio, *idx, 0);
+
+ mlc->serio_oidx[map->didx] = 0;
+ mlc->serio_opacket[map->didx] = 0;
+
+ return 0;
+}
+
+static int hil_mlc_serio_open(struct serio *serio) {
+ struct hil_mlc_serio_map *map;
+ struct hil_mlc *mlc;
+
+ if (serio->private != NULL) return -EBUSY;
+
+ map = serio->driver;
+ if (map == NULL) {
+ BUG();
+ return -ENODEV;
+ }
+ mlc = map->mlc;
+ if (mlc == NULL) {
+ BUG();
+ return -ENODEV;
+ }
+
+ mlc->inc_use_count();
+
+ return 0;
+}
+
+static void hil_mlc_serio_close(struct serio *serio) {
+ struct hil_mlc_serio_map *map;
+ struct hil_mlc *mlc;
+
+ map = serio->driver;
+ if (map == NULL) {
+ BUG();
+ return;
+ }
+ mlc = map->mlc;
+ if (mlc == NULL) {
+ BUG();
+ return;
+ }
+
+ mlc->dec_use_count();
+
+ serio->private = NULL;
+ serio->dev = NULL;
+ /* TODO wake up interruptable */
+}
+
+int hil_mlc_register(hil_mlc *mlc) {
+ int i;
+ unsigned long flags;
+
+ MOD_INC_USE_COUNT;
+ if (mlc == NULL) {
+ MOD_DEC_USE_COUNT;
+ return -EINVAL;
+ }
+
+ mlc->istarted = 0;
+ mlc->ostarted = 0;
+
+ mlc->lock = RW_LOCK_UNLOCKED;
+ init_MUTEX(&(mlc->osem));
+
+ init_MUTEX(&(mlc->isem));
+ mlc->icount = -1;
+ mlc->imatch = 0;
+
+ mlc->opercnt = 0;
+
+ init_MUTEX_LOCKED(&(mlc->csem));
+
+ hil_mlc_clear_di_scratch(mlc);
+ hil_mlc_clear_di_map(mlc, 0);
+ for (i = 0; i < HIL_MLC_DEVMEM; i++) {
+ hil_mlc_copy_di_scratch(mlc, i);
+ memset(&(mlc->serio[i]), 0, sizeof(mlc->serio[0]));
+ mlc->serio[i].type = SERIO_HIL | SERIO_HIL_MLC;
+ mlc->serio[i].write = hil_mlc_serio_write;
+ mlc->serio[i].open = hil_mlc_serio_open;
+ mlc->serio[i].close = hil_mlc_serio_close;
+ mlc->serio[i].driver = &(mlc->serio_map[i]);
+ mlc->serio_map[i].mlc = mlc;
+ mlc->serio_map[i].didx = i;
+ mlc->serio_map[i].di_revmap = -1;
+ mlc->serio_opacket[i] = 0;
+ mlc->serio_oidx[i] = 0;
+ serio_register_port(&(mlc->serio[i]));
+ }
+
+ mlc->tasklet = &hil_mlcs_tasklet;
+
+ write_lock_irqsave(&hil_mlcs_lock, flags);
+ list_add_tail(&mlc->list, &hil_mlcs);
+ mlc->seidx = HILSEN_START;
+ write_unlock_irqrestore(&hil_mlcs_lock, flags);
+
+ tasklet_schedule(&hil_mlcs_tasklet);
+ return 0;
+}
+
+int hil_mlc_unregister(hil_mlc *mlc) {
+ struct list_head *tmp;
+ unsigned long flags;
+ int i;
+
+ if (mlc == NULL)
+ return -EINVAL;
+
+ write_lock_irqsave(&hil_mlcs_lock, flags);
+ list_for_each(tmp, &hil_mlcs) {
+ if (list_entry(tmp, hil_mlc, list) == mlc)
+ goto found;
+ }
+
+ /* not found in list */
+ write_unlock_irqrestore(&hil_mlcs_lock, flags);
+ tasklet_schedule(&hil_mlcs_tasklet);
+ return -ENODEV;
+
+ found:
+ list_del(tmp);
+ write_unlock_irqrestore(&hil_mlcs_lock, flags);
+ MOD_DEC_USE_COUNT;
+
+ for (i = 0; i < HIL_MLC_DEVMEM; i++)
+ serio_unregister_port(&(mlc->serio[i]));
+
+ tasklet_schedule(&hil_mlcs_tasklet);
+ return 0;
+}
+
+/**************************** Module interface *************************/
+
+static int __init hil_mlc_init(void)
+{
+ init_timer(&hil_mlcs_kicker);
+ hil_mlcs_kicker.expires = jiffies + HZ;
+ hil_mlcs_kicker.function = &hil_mlcs_timer;
+ add_timer(&hil_mlcs_kicker);
+
+ tasklet_enable(&hil_mlcs_tasklet);
+
+ return 0;
+}
+
+static void __exit hil_mlc_exit(void)
+{
+ del_timer(&hil_mlcs_kicker);
+
+ tasklet_disable(&hil_mlcs_tasklet);
+ tasklet_kill(&hil_mlcs_tasklet);
+}
+
+module_init(hil_mlc_init);
+module_exit(hil_mlc_exit);
FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen (who was at: slshen@lbl.gov)