xip changes between 2.6.11 and 20050530?

xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

Hello to the list members!

I had XIP in linux-2.6.11 running very well.
Then I patched a mtd snapshot from 20050530 in in order to get code into
there which properly unlock intel K3 flash devices properly (which is
not solved yet).
Now the xipuImage freezes at booting when it comes to

Probing scb9328_flash at physical address 0x10000000 (16-bit buswidth)

Are there any changes between the both versions which I should be aware
of? The xip.h looks still reasonable though, I don't get it.

What I doscovered is the NO_XIP define which was introduced into some
mapping drivers, should I take care of that and put it into mine also?
I did not found out yet, what it does really...

regards, Konsti

Re: xip changes between 2.6.11 and 20050530?

[Todd Poynor]

Konstantin Kletschke wrote:
> Hello to the list members!
> 
> I had XIP in linux-2.6.11 running very well.
> Then I patched a mtd snapshot from 20050530 in in order to get code into
> there which properly unlock intel K3 flash devices properly (which is
> not solved yet).
> Now the xipuImage freezes at booting when it comes to
> 
> Probing scb9328_flash at physical address 0x10000000 (16-bit buswidth)

Do you have a diff of drivers/mtd/chips/cfi_cmdset_0001.c between the 
two versions?  I'll guess a change was made that didn't wrap something 
in the xip disable/check for pending interrupts stuff.

-- 
Todd

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

* Todd Poynor <tpoynor@mvista.com> [Wed, Jul 06, 2005 at 09:44:30AM -0700]:

> Do you have a diff of drivers/mtd/chips/cfi_cmdset_0001.c between the 
> two versions?  I'll guess a change was made that didn't wrap something 
> in the xip disable/check for pending interrupts stuff.

Stupid stupid me! In order to integrate the intel K3 unlocking I
inserted a dozen of printk commands which of couse broke XIP!

*ARGH* How stupid, I simply forgot.

Now I removed them and XIP seems to be fine again.
Now I have to fix sync burst again and then the intel K3 unlocking is
scheduled :)

Regards, Konsti

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

* Todd Poynor <tpoynor@mvista.com> [Wed, Jul 06, 2005 at 09:44:30AM -0700]:

Actually I must be mad. This morning I got it working, now its broken
again and I don't know which file I broke. Damn!

Now the Kernel freezes when I try to write something (touch file), not
before! xip.h still looks reasonable.

Does somebody have tips on how to debug this? Will debug output in sync
over serial line help or sys-req or some sort of that?

> Do you have a diff of drivers/mtd/chips/cfi_cmdset_0001.c between the 
> two versions?  I'll guess a change was made that didn't wrap something 
> in the xip disable/check for pending interrupts stuff.

diff -Nur /usr/src/linux-2.6.11/drivers/mtd/chips/cfi_cmdset_0001.c cfi_cmdset_0001.c
--- /usr/src/linux-2.6.11/drivers/mtd/chips/cfi_cmdset_0001.c	2005-03-02 08:37:52.000000000 +0100
+++ cfi_cmdset_0001.c	2005-07-07 09:48:23.000000000 +0200
@@ -4,7 +4,7 @@
  *
  * (C) 2000 Red Hat. GPL'd
  *
- * $Id: cfi_cmdset_0001.c,v 1.164 2004/11/16 18:29:00 dwmw2 Exp $
+ * $Id: cfi_cmdset_0001.c,v 1.178 2005/05/19 17:05:43 nico Exp $
  *
  * 
  * 10/10/2000	Nicolas Pitre <nico@cam.org>
@@ -29,6 +29,7 @@
 #include <linux/slab.h>
 #include <linux/delay.h>
 #include <linux/interrupt.h>
+#include <linux/reboot.h>
 #include <linux/mtd/xip.h>
 #include <linux/mtd/map.h>
 #include <linux/mtd/mtd.h>
@@ -48,16 +49,26 @@
 #define M50LPW080       0x002F
 
 static int cfi_intelext_read (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
-//static int cfi_intelext_read_user_prot_reg (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
-//static int cfi_intelext_read_fact_prot_reg (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
 static int cfi_intelext_write_words(struct mtd_info *, loff_t, size_t, size_t *, const u_char *);
 static int cfi_intelext_write_buffers(struct mtd_info *, loff_t, size_t, size_t *, const u_char *);
 static int cfi_intelext_erase_varsize(struct mtd_info *, struct erase_info *);
 static void cfi_intelext_sync (struct mtd_info *);
 static int cfi_intelext_lock(struct mtd_info *mtd, loff_t ofs, size_t len);
 static int cfi_intelext_unlock(struct mtd_info *mtd, loff_t ofs, size_t len);
+#ifdef CONFIG_MTD_OTP
+static int cfi_intelext_read_fact_prot_reg (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
+static int cfi_intelext_read_user_prot_reg (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
+static int cfi_intelext_write_user_prot_reg (struct mtd_info *, loff_t, size_t, size_t *, u_char *);
+static int cfi_intelext_lock_user_prot_reg (struct mtd_info *, loff_t, size_t);
+static int cfi_intelext_get_fact_prot_info (struct mtd_info *,
+					    struct otp_info *, size_t);
+static int cfi_intelext_get_user_prot_info (struct mtd_info *,
+					    struct otp_info *, size_t);
+#endif
 static int cfi_intelext_suspend (struct mtd_info *);
 static void cfi_intelext_resume (struct mtd_info *);
+static int cfi_intelext_reboot (struct notifier_block *, unsigned long, void *);
+static void cfi_intelext_instantlock_add(struct mtd_info* mtd);
 
 static void cfi_intelext_destroy(struct mtd_info *);
 
@@ -252,7 +263,8 @@
 		int nb_parts, i;
 
 		/* Protection Register info */
-		extra_size += (extp->NumProtectionFields - 1) * (4 + 6);
+		extra_size += (extp->NumProtectionFields - 1) *
+			      sizeof(struct cfi_intelext_otpinfo);
 
 		/* Burst Read info */
 		extra_size += 6;
@@ -324,7 +336,9 @@
 	mtd->resume  = cfi_intelext_resume;
 	mtd->flags   = MTD_CAP_NORFLASH;
 	mtd->name    = map->name;
-	
+
+	mtd->reboot_notifier.notifier_call = cfi_intelext_reboot;
+
 	if (cfi->cfi_mode == CFI_MODE_CFI) {
 		/* 
 		 * It's a real CFI chip, not one for which the probe
@@ -377,6 +391,7 @@
 {
 	struct map_info *map = mtd->priv;
 	struct cfi_private *cfi = map->fldrv_priv;
+	struct cfi_pri_intelext *extp = cfi->cmdset_priv;
 	unsigned long offset = 0;
 	int i,j;
 	unsigned long devsize = (1<<cfi->cfiq->DevSize) * cfi->interleave;
@@ -422,17 +437,25 @@
 		       mtd->eraseregions[i].numblocks);
 	}
 
-#if 0
-	mtd->read_user_prot_reg = cfi_intelext_read_user_prot_reg;
+#ifdef CONFIG_MTD_OTP
 	mtd->read_fact_prot_reg = cfi_intelext_read_fact_prot_reg;
+	mtd->read_user_prot_reg = cfi_intelext_read_user_prot_reg;
+	mtd->write_user_prot_reg = cfi_intelext_write_user_prot_reg;
+	mtd->lock_user_prot_reg = cfi_intelext_lock_user_prot_reg;
+	mtd->get_fact_prot_info = cfi_intelext_get_fact_prot_info;
+	mtd->get_user_prot_info = cfi_intelext_get_user_prot_info;
 #endif
 
+	if (extp && (extp->FeatureSupport & (1 << 5)))
+		mtd->notify_add = cfi_intelext_instantlock_add;
+
 	/* This function has the potential to distort the reality
 	   a bit and therefore should be called last. */
 	if (cfi_intelext_partition_fixup(mtd, &cfi) != 0)
 		goto setup_err;
 
 	__module_get(THIS_MODULE);
+	register_reboot_notifier(&mtd->reboot_notifier);
 	return mtd;
 
  setup_err:
@@ -471,7 +494,8 @@
 		int offs, numregions, numparts, partshift, numvirtchips, i, j;
 
 		/* Protection Register info */
-		offs = (extp->NumProtectionFields - 1) * (4 + 6);
+		offs = (extp->NumProtectionFields - 1) *
+		       sizeof(struct cfi_intelext_otpinfo);
 
 		/* Burst Read info */
 		offs += 6;
@@ -563,7 +587,7 @@
  resettime:
 	timeo = jiffies + HZ;
  retry:
-	if (chip->priv && (mode == FL_WRITING || mode == FL_ERASING)) {
+	if (chip->priv && (mode == FL_WRITING || mode == FL_ERASING || mode == FL_OTP_WRITE)) {
 		/*
 		 * OK. We have possibility for contension on the write/erase
 		 * operations which are global to the real chip and not per
@@ -807,10 +831,6 @@
  * assembly to make sure inline functions were actually inlined and that gcc
  * didn't emit calls to its own support functions). Also configuring MTD CFI
  * support to a single buswidth and a single interleave is also recommended.
- * Note that not only IRQs are disabled but the preemption count is also
- * increased to prevent other locking primitives (namely spin_unlock) from
- * decrementing the preempt count to zero and scheduling the CPU away while
- * not in array mode.
  */
 
 static void xip_disable(struct map_info *map, struct flchip *chip,
@@ -818,7 +838,6 @@
 {
 	/* TODO: chips with no XIP use should ignore and return */
 	(void) map_read(map, adr); /* ensure mmu mapping is up to date */
-	preempt_disable();
 	local_irq_disable();
 }
 
@@ -833,7 +852,6 @@
 	(void) map_read(map, adr);
 	asm volatile (".rep 8; nop; .endr"); /* fill instruction prefetch */
 	local_irq_enable();
-	preempt_enable();
 }
 
 /*
@@ -909,7 +927,7 @@
 			(void) map_read(map, adr);
 			asm volatile (".rep 8; nop; .endr");
 			local_irq_enable();
-			preempt_enable();
+			spin_unlock(chip->mutex);
 			asm volatile (".rep 8; nop; .endr");
 			cond_resched();
 
@@ -919,15 +937,15 @@
 			 * a suspended erase state.  If so let's wait
 			 * until it's done.
 			 */
-			preempt_disable();
+			spin_lock(chip->mutex);
 			while (chip->state != newstate) {
 				DECLARE_WAITQUEUE(wait, current);
 				set_current_state(TASK_UNINTERRUPTIBLE);
 				add_wait_queue(&chip->wq, &wait);
-				preempt_enable();
+				spin_unlock(chip->mutex);
 				schedule();
 				remove_wait_queue(&chip->wq, &wait);
-				preempt_disable();
+				spin_lock(chip->mutex);
 			}
 			/* Disallow XIP again */
 			local_irq_disable();
@@ -956,12 +974,14 @@
  * The INVALIDATE_CACHED_RANGE() macro is normally used in parallel while
  * the flash is actively programming or erasing since we have to poll for
  * the operation to complete anyway.  We can't do that in a generic way with
- * a XIP setup so do it before the actual flash operation in this case.
+ * a XIP setup so do it before the actual flash operation in this case
+ * and stub it out from INVALIDATE_CACHE_UDELAY.
  */
-#undef INVALIDATE_CACHED_RANGE
-#define INVALIDATE_CACHED_RANGE(x...)
-#define XIP_INVAL_CACHED_RANGE(map, from, size) \
-	do { if(map->inval_cache) map->inval_cache(map, from, size); } while(0)
+#define XIP_INVAL_CACHED_RANGE(map, from, size)  \
+	INVALIDATE_CACHED_RANGE(map, from, size)
+
+#define INVALIDATE_CACHE_UDELAY(map, chip, adr, len, usec)  \
+	UDELAY(map, chip, adr, usec)
 
 /*
  * Extra notes:
@@ -984,11 +1004,23 @@
 
 #define xip_disable(map, chip, adr)
 #define xip_enable(map, chip, adr)
-
-#define UDELAY(map, chip, adr, usec)  cfi_udelay(usec)
-
 #define XIP_INVAL_CACHED_RANGE(x...)
 
+#define UDELAY(map, chip, adr, usec)  \
+do {  \
+	spin_unlock(chip->mutex);  \
+	cfi_udelay(usec);  \
+	spin_lock(chip->mutex);  \
+} while (0)
+
+#define INVALIDATE_CACHE_UDELAY(map, chip, adr, len, usec)  \
+do {  \
+	spin_unlock(chip->mutex);  \
+	INVALIDATE_CACHED_RANGE(map, adr, len);  \
+	cfi_udelay(usec);  \
+	spin_lock(chip->mutex);  \
+} while (0)
+
 #endif
 
 static int do_point_onechip (struct map_info *map, struct flchip *chip, loff_t adr, size_t len)
@@ -1176,111 +1208,11 @@
 	return ret;
 }
 
-#if 0
-static int __xipram cfi_intelext_read_prot_reg (struct mtd_info *mtd,
-						loff_t from, size_t len,
-						size_t *retlen,
-						u_char *buf,
-						int base_offst, int reg_sz)
-{
-	struct map_info *map = mtd->priv;
-	struct cfi_private *cfi = map->fldrv_priv;
-	struct cfi_pri_intelext *extp = cfi->cmdset_priv;
-	struct flchip *chip;
-	int ofs_factor = cfi->interleave * cfi->device_type;
-	int count = len;
-	int chip_num, offst;
-	int ret;
-
-	chip_num = ((unsigned int)from/reg_sz);
-	offst = from - (reg_sz*chip_num)+base_offst;
-
-	while (count) {
-	/* Calculate which chip & protection register offset we need */
-
-		if (chip_num >= cfi->numchips)
-			goto out;
-
-		chip = &cfi->chips[chip_num];
-		
-		spin_lock(chip->mutex);
-		ret = get_chip(map, chip, chip->start, FL_JEDEC_QUERY);
-		if (ret) {
-			spin_unlock(chip->mutex);
-			return (len-count)?:ret;
-		}
-
-		xip_disable(map, chip, chip->start);
-
-		if (chip->state != FL_JEDEC_QUERY) {
-			map_write(map, CMD(0x90), chip->start);
-			chip->state = FL_JEDEC_QUERY;
-		}
-
-		while (count && ((offst-base_offst) < reg_sz)) {
-			*buf = map_read8(map,(chip->start+((extp->ProtRegAddr+1)*ofs_factor)+offst));
-			buf++;
-			offst++;
-			count--;
-		}
-
-		xip_enable(map, chip, chip->start);
-		put_chip(map, chip, chip->start);
-		spin_unlock(chip->mutex);
-
-		/* Move on to the next chip */
-		chip_num++;
-		offst = base_offst;
-	}
-	
- out:	
-	return len-count;
-}
-	
-static int cfi_intelext_read_user_prot_reg (struct mtd_info *mtd, loff_t from, size_t len, size_t *retlen, u_char *buf)
-{
-	struct map_info *map = mtd->priv;
-	struct cfi_private *cfi = map->fldrv_priv;
-	struct cfi_pri_intelext *extp=cfi->cmdset_priv;
-	int base_offst,reg_sz;
-	
-	/* Check that we actually have some protection registers */
-	if(!extp || !(extp->FeatureSupport&64)){
-		printk(KERN_WARNING "%s: This flash device has no protection data to read!\n",map->name);
-		return 0;
-	}
-
-	base_offst=(1<<extp->FactProtRegSize);
-	reg_sz=(1<<extp->UserProtRegSize);
-
-	return cfi_intelext_read_prot_reg(mtd, from, len, retlen, buf, base_offst, reg_sz);
-}
-
-static int cfi_intelext_read_fact_prot_reg (struct mtd_info *mtd, loff_t from, size_t len, size_t *retlen, u_char *buf)
-{
-	struct map_info *map = mtd->priv;
-	struct cfi_private *cfi = map->fldrv_priv;
-	struct cfi_pri_intelext *extp=cfi->cmdset_priv;
-	int base_offst,reg_sz;
-	
-	/* Check that we actually have some protection registers */
-	if(!extp || !(extp->FeatureSupport&64)){
-		printk(KERN_WARNING "%s: This flash device has no protection data to read!\n",map->name);
-		return 0;
-	}
-
-	base_offst=0;
-	reg_sz=(1<<extp->FactProtRegSize);
-
-	return cfi_intelext_read_prot_reg(mtd, from, len, retlen, buf, base_offst, reg_sz);
-}
-#endif
-
 static int __xipram do_write_oneword(struct map_info *map, struct flchip *chip,
-				     unsigned long adr, map_word datum)
+				     unsigned long adr, map_word datum, int mode)
 {
 	struct cfi_private *cfi = map->fldrv_priv;
-	map_word status, status_OK;
+	map_word status, status_OK, write_cmd;
 	unsigned long timeo;
 	int z, ret=0;
 
@@ -1288,9 +1220,14 @@
 
 	/* Let's determine this according to the interleave only once */
 	status_OK = CMD(0x80);
+	switch (mode) {
+	case FL_WRITING:   write_cmd = CMD(0x40); break;
+	case FL_OTP_WRITE: write_cmd = CMD(0xc0); break;
+	default: return -EINVAL;
+	}
 
 	spin_lock(chip->mutex);
-	ret = get_chip(map, chip, adr, FL_WRITING);
+	ret = get_chip(map, chip, adr, mode);
 	if (ret) {
 		spin_unlock(chip->mutex);
 		return ret;
@@ -1299,19 +1236,18 @@
 	XIP_INVAL_CACHED_RANGE(map, adr, map_bankwidth(map));
 	ENABLE_VPP(map);
 	xip_disable(map, chip, adr);
-	map_write(map, CMD(0x40), adr);
+	map_write(map, write_cmd, adr);
 	map_write(map, datum, adr);
-	chip->state = FL_WRITING;
+	chip->state = mode;
 
-	spin_unlock(chip->mutex);
-	INVALIDATE_CACHED_RANGE(map, adr, map_bankwidth(map));
-	UDELAY(map, chip, adr, chip->word_write_time);
-	spin_lock(chip->mutex);
+	INVALIDATE_CACHE_UDELAY(map, chip,
+				adr, map_bankwidth(map),
+				chip->word_write_time);
 
 	timeo = jiffies + (HZ/2);
 	z = 0;
 	for (;;) {
-		if (chip->state != FL_WRITING) {
+		if (chip->state != mode) {
 			/* Someone's suspended the write. Sleep */
 			DECLARE_WAITQUEUE(wait, current);
 
@@ -1339,10 +1275,8 @@
 		}
 
 		/* Latency issues. Drop the lock, wait a while and retry */
-		spin_unlock(chip->mutex);
 		z++;
 		UDELAY(map, chip, adr, 1);
-		spin_lock(chip->mutex);
 	}
 	if (!z) {
 		chip->word_write_time--;
@@ -1399,7 +1333,7 @@
 		datum = map_word_load_partial(map, datum, buf, gap, n);
 
 		ret = do_write_oneword(map, &cfi->chips[chipnum],
-					       bus_ofs, datum);
+					       bus_ofs, datum, FL_WRITING);
 		if (ret) 
 			return ret;
 
@@ -1420,7 +1354,7 @@
 		map_word datum = map_word_load(map, buf);
 
 		ret = do_write_oneword(map, &cfi->chips[chipnum],
-				ofs, datum);
+				       ofs, datum, FL_WRITING);
 		if (ret)
 			return ret;
 
@@ -1444,7 +1378,7 @@
 		datum = map_word_load_partial(map, datum, buf, 0, len);
 
 		ret = do_write_oneword(map, &cfi->chips[chipnum],
-					       ofs, datum);
+				       ofs, datum, FL_WRITING);
 		if (ret) 
 			return ret;
 		
@@ -1506,9 +1440,7 @@
 		if (map_word_andequal(map, status, status_OK, status_OK))
 			break;
 
-		spin_unlock(chip->mutex);
 		UDELAY(map, chip, cmd_adr, 1);
-		spin_lock(chip->mutex);
 
 		if (++z > 20) {
 			/* Argh. Not ready for write to buffer */
@@ -1554,10 +1486,9 @@
 	map_write(map, CMD(0xd0), cmd_adr);
 	chip->state = FL_WRITING;
 
-	spin_unlock(chip->mutex);
-	INVALIDATE_CACHED_RANGE(map, adr, len);
-	UDELAY(map, chip, cmd_adr, chip->buffer_write_time);
-	spin_lock(chip->mutex);
+	INVALIDATE_CACHE_UDELAY(map, chip, 
+				cmd_adr, len,
+				chip->buffer_write_time);
 
 	timeo = jiffies + (HZ/2);
 	z = 0;
@@ -1589,10 +1520,8 @@
 		}
 		
 		/* Latency issues. Drop the lock, wait a while and retry */
-		spin_unlock(chip->mutex);
-		UDELAY(map, chip, cmd_adr, 1);
 		z++;
-		spin_lock(chip->mutex);
+		UDELAY(map, chip, cmd_adr, 1);
 	}
 	if (!z) {
 		chip->buffer_write_time--;
@@ -1720,10 +1649,9 @@
 	chip->state = FL_ERASING;
 	chip->erase_suspended = 0;
 
-	spin_unlock(chip->mutex);
-	INVALIDATE_CACHED_RANGE(map, adr, len);
-	UDELAY(map, chip, adr, chip->erase_time*1000/2);
-	spin_lock(chip->mutex);
+	INVALIDATE_CACHE_UDELAY(map, chip,
+				adr, len,
+				chip->erase_time*1000/2);
 
 	/* FIXME. Use a timer to check this, and return immediately. */
 	/* Once the state machine's known to be working I'll do that */
@@ -1768,9 +1696,7 @@
 		}
 		
 		/* Latency issues. Drop the lock, wait a while and retry */
-		spin_unlock(chip->mutex);
 		UDELAY(map, chip, adr, 1000000/HZ);
-		spin_lock(chip->mutex);
 	}
 
 	/* We've broken this before. It doesn't hurt to be safe */
@@ -1780,44 +1706,34 @@
 
 	/* check for lock bit */
 	if (map_word_bitsset(map, status, CMD(0x3a))) {
-		unsigned char chipstatus;
+		unsigned long chipstatus;
 
 		/* Reset the error bits */
 		map_write(map, CMD(0x50), adr);
 		map_write(map, CMD(0x70), adr);
 		xip_enable(map, chip, adr);
 
-		chipstatus = status.x[0];
-		if (!map_word_equal(map, status, CMD(chipstatus))) {
-			int i, w;
-			for (w=0; w<map_words(map); w++) {
-				for (i = 0; i<cfi_interleave(cfi); i++) {
-					chipstatus |= status.x[w] >> (cfi->device_type * 8);
-				}
-			}
-			printk(KERN_WARNING "Status is not identical for all chips: 0x%lx. Merging to give 0x%02x\n",
-			       status.x[0], chipstatus);
-		}
+		chipstatus = MERGESTATUS(status);
 
 		if ((chipstatus & 0x30) == 0x30) {
-			printk(KERN_NOTICE "Chip reports improper command sequence: status 0x%x\n", chipstatus);
+			printk(KERN_NOTICE "Chip reports improper command sequence: status 0x%lx\n", chipstatus);
 			ret = -EIO;
 		} else if (chipstatus & 0x02) {
 			/* Protection bit set */
 			ret = -EROFS;
 		} else if (chipstatus & 0x8) {
 			/* Voltage */
-			printk(KERN_WARNING "Chip reports voltage low on erase: status 0x%x\n", chipstatus);
+			printk(KERN_WARNING "Chip reports voltage low on erase: status 0x%lx\n", chipstatus);
 			ret = -EIO;
 		} else if (chipstatus & 0x20) {
 			if (retries--) {
-				printk(KERN_DEBUG "Chip erase failed at 0x%08lx: status 0x%x. Retrying...\n", adr, chipstatus);
+				printk(KERN_DEBUG "Chip erase failed at 0x%08lx: status 0x%lx. Retrying...\n", adr, chipstatus);
 				timeo = jiffies + HZ;
 				put_chip(map, chip, adr);
 				spin_unlock(chip->mutex);
 				goto retry;
 			}
-			printk(KERN_DEBUG "Chip erase failed at 0x%08lx: status 0x%x\n", adr, chipstatus);
+			printk(KERN_DEBUG "Chip erase failed at 0x%08lx: status 0x%lx\n", adr, chipstatus);
 			ret = -EIO;
 		}
 	} else {
@@ -1882,6 +1798,7 @@
 		
 		if (chip->state == FL_SYNCING) {
 			chip->state = chip->oldstate;
+			chip->oldstate = FL_READY;
 			wake_up(&chip->wq);
 		}
 		spin_unlock(chip->mutex);
@@ -1897,8 +1814,9 @@
 	struct cfi_private *cfi = map->fldrv_priv;
 	int status, ofs_factor = cfi->interleave * cfi->device_type;
 
+	adr += chip->start;
 	xip_disable(map, chip, adr+(2*ofs_factor));
-	cfi_send_gen_cmd(0x90, 0x55, 0, map, cfi, cfi->device_type, NULL);
+	map_write(map, CMD(0x90), adr+(2*ofs_factor));
 	chip->state = FL_JEDEC_QUERY;
 	status = cfi_read_query(map, adr+(2*ofs_factor));
 	xip_enable(map, chip, 0);
@@ -1915,6 +1833,7 @@
 				       unsigned long adr, int len, void *thunk)
 {
 	struct cfi_private *cfi = map->fldrv_priv;
+	struct cfi_pri_intelext *extp = cfi->cmdset_priv;
 	map_word status, status_OK;
 	unsigned long timeo = jiffies + HZ;
 	int ret;
@@ -1944,9 +1863,13 @@
 	} else
 		BUG();
 
-	spin_unlock(chip->mutex);
-	UDELAY(map, chip, adr, 1000000/HZ);
-	spin_lock(chip->mutex);
+	/*
+	 * If Instant Individual Block Locking supported then no need
+	 * to delay.
+	 */
+
+	if (!extp || !(extp->FeatureSupport & (1 << 5)))
+		UDELAY(map, chip, adr, 1000000/HZ);
 
 	/* FIXME. Use a timer to check this, and return immediately. */
 	/* Once the state machine's known to be working I'll do that */
@@ -1973,9 +1896,7 @@
 		}
 		
 		/* Latency issues. Drop the lock, wait a while and retry */
-		spin_unlock(chip->mutex);
 		UDELAY(map, chip, adr, 1);
-		spin_lock(chip->mutex);
 	}
 	
 	/* Done and happy. */
@@ -2034,6 +1955,311 @@
 	return ret;
 }
 
+#ifdef CONFIG_MTD_OTP
+
+typedef int (*otp_op_t)(struct map_info *map, struct flchip *chip, 
+			u_long data_offset, u_char *buf, u_int size,
+			u_long prot_offset, u_int groupno, u_int groupsize);
+
+static int __xipram
+do_otp_read(struct map_info *map, struct flchip *chip, u_long offset,
+	    u_char *buf, u_int size, u_long prot, u_int grpno, u_int grpsz)
+{
+	struct cfi_private *cfi = map->fldrv_priv;
+	int ret;
+
+	spin_lock(chip->mutex);
+	ret = get_chip(map, chip, chip->start, FL_JEDEC_QUERY);
+	if (ret) {
+		spin_unlock(chip->mutex);
+		return ret;
+	}
+
+	/* let's ensure we're not reading back cached data from array mode */
+	INVALIDATE_CACHED_RANGE(map, chip->start + offset, size);
+
+	xip_disable(map, chip, chip->start);
+	if (chip->state != FL_JEDEC_QUERY) {
+		map_write(map, CMD(0x90), chip->start);
+		chip->state = FL_JEDEC_QUERY;
+	}
+	map_copy_from(map, buf, chip->start + offset, size);
+	xip_enable(map, chip, chip->start);
+
+	/* then ensure we don't keep OTP data in the cache */
+	INVALIDATE_CACHED_RANGE(map, chip->start + offset, size);
+
+	put_chip(map, chip, chip->start);
+	spin_unlock(chip->mutex);
+	return 0;
+}
+
+static int
+do_otp_write(struct map_info *map, struct flchip *chip, u_long offset,
+	     u_char *buf, u_int size, u_long prot, u_int grpno, u_int grpsz)
+{
+	int ret;
+
+	while (size) {
+		unsigned long bus_ofs = offset & ~(map_bankwidth(map)-1);
+		int gap = offset - bus_ofs;
+		int n = min_t(int, size, map_bankwidth(map)-gap);
+		map_word datum = map_word_ff(map);
+
+		datum = map_word_load_partial(map, datum, buf, gap, n);
+		ret = do_write_oneword(map, chip, bus_ofs, datum, FL_OTP_WRITE);
+		if (ret) 
+			return ret;
+
+		offset += n;
+		buf += n;
+		size -= n;
+	}
+
+	return 0;
+}
+
+static int
+do_otp_lock(struct map_info *map, struct flchip *chip, u_long offset,
+	    u_char *buf, u_int size, u_long prot, u_int grpno, u_int grpsz)
+{
+	struct cfi_private *cfi = map->fldrv_priv;
+	map_word datum;
+
+	/* make sure area matches group boundaries */
+	if (size != grpsz)
+		return -EXDEV;
+
+	datum = map_word_ff(map);
+	datum = map_word_clr(map, datum, CMD(1 << grpno));
+	return do_write_oneword(map, chip, prot, datum, FL_OTP_WRITE);
+}
+
+static int cfi_intelext_otp_walk(struct mtd_info *mtd, loff_t from, size_t len,
+				 size_t *retlen, u_char *buf,
+				 otp_op_t action, int user_regs)
+{
+	struct map_info *map = mtd->priv;
+	struct cfi_private *cfi = map->fldrv_priv;
+	struct cfi_pri_intelext *extp = cfi->cmdset_priv;
+	struct flchip *chip;
+	struct cfi_intelext_otpinfo *otp;
+	u_long devsize, reg_prot_offset, data_offset;
+	u_int chip_num, chip_step, field, reg_fact_size, reg_user_size;
+	u_int groups, groupno, groupsize, reg_fact_groups, reg_user_groups;
+	int ret;
+
+	*retlen = 0;
+
+	/* Check that we actually have some OTP registers */
+	if (!extp || !(extp->FeatureSupport & 64) || !extp->NumProtectionFields)
+		return -ENODATA;
+
+	/* we need real chips here not virtual ones */
+	devsize = (1 << cfi->cfiq->DevSize) * cfi->interleave;
+	chip_step = devsize >> cfi->chipshift;
+	chip_num = 0;
+
+	/* Some chips have OTP located in the _top_ partition only.
+	   For example: Intel 28F256L18T (T means top-parameter device) */
+	if (cfi->mfr == MANUFACTURER_INTEL) {
+		switch (cfi->id) {
+		case 0x880b:
+		case 0x880c:
+		case 0x880d:
+			chip_num = chip_step - 1;
+		}
+	}
+
+	for ( ; chip_num < cfi->numchips; chip_num += chip_step) {
+		chip = &cfi->chips[chip_num];
+		otp = (struct cfi_intelext_otpinfo *)&extp->extra[0];
+
+		/* first OTP region */
+		field = 0;
+		reg_prot_offset = extp->ProtRegAddr;
+		reg_fact_groups = 1;
+		reg_fact_size = 1 << extp->FactProtRegSize;
+		reg_user_groups = 1;
+		reg_user_size = 1 << extp->UserProtRegSize;
+
+		while (len > 0) {
+			/* flash geometry fixup */
+			data_offset = reg_prot_offset + 1;
+			data_offset *= cfi->interleave * cfi->device_type;
+			reg_prot_offset *= cfi->interleave * cfi->device_type;
+			reg_fact_size *= cfi->interleave;
+			reg_user_size *= cfi->interleave;
+
+			if (user_regs) {
+				groups = reg_user_groups;
+				groupsize = reg_user_size;
+				/* skip over factory reg area */
+				groupno = reg_fact_groups;
+				data_offset += reg_fact_groups * reg_fact_size;
+			} else {
+				groups = reg_fact_groups;
+				groupsize = reg_fact_size;
+				groupno = 0;
+			}
+
+			while (len > 0 && groups > 0) {
+				if (!action) {
+					/*
+					 * Special case: if action is NULL
+					 * we fill buf with otp_info records.
+					 */
+					struct otp_info *otpinfo;
+					map_word lockword;
+					len -= sizeof(struct otp_info);
+					if (len <= 0)
+						return -ENOSPC;
+					ret = do_otp_read(map, chip,
+							  reg_prot_offset,
+							  (u_char *)&lockword,
+							  map_bankwidth(map),
+							  0, 0,  0);
+					if (ret)
+						return ret;
+					otpinfo = (struct otp_info *)buf;
+					otpinfo->start = from;
+					otpinfo->length = groupsize;
+					otpinfo->locked =
+					   !map_word_bitsset(map, lockword,
+							     CMD(1 << groupno));
+					from += groupsize;
+					buf += sizeof(*otpinfo);
+					*retlen += sizeof(*otpinfo);
+				} else if (from >= groupsize) {
+					from -= groupsize;
+					data_offset += groupsize;
+				} else {
+					int size = groupsize;
+					data_offset += from;
+					size -= from;
+					from = 0;
+					if (size > len)
+						size = len;
+					ret = action(map, chip, data_offset,
+						     buf, size, reg_prot_offset,
+						     groupno, groupsize);
+					if (ret < 0)
+						return ret;
+					buf += size;
+					len -= size;
+					*retlen += size;
+					data_offset += size;
+				}
+				groupno++;
+				groups--;
+			}
+
+			/* next OTP region */
+			if (++field == extp->NumProtectionFields)
+				break;
+			reg_prot_offset = otp->ProtRegAddr;
+			reg_fact_groups = otp->FactGroups;
+			reg_fact_size = 1 << otp->FactProtRegSize;
+			reg_user_groups = otp->UserGroups;
+			reg_user_size = 1 << otp->UserProtRegSize;
+			otp++;
+		}
+	}
+
+	return 0;
+}
+
+static int cfi_intelext_read_fact_prot_reg(struct mtd_info *mtd, loff_t from,
+					   size_t len, size_t *retlen,
+					    u_char *buf)
+{
+	return cfi_intelext_otp_walk(mtd, from, len, retlen,
+				     buf, do_otp_read, 0);
+}
+
+static int cfi_intelext_read_user_prot_reg(struct mtd_info *mtd, loff_t from,
+					   size_t len, size_t *retlen,
+					    u_char *buf)
+{
+	return cfi_intelext_otp_walk(mtd, from, len, retlen,
+				     buf, do_otp_read, 1);
+}
+
+static int cfi_intelext_write_user_prot_reg(struct mtd_info *mtd, loff_t from,
+					    size_t len, size_t *retlen,
+					     u_char *buf)
+{
+	return cfi_intelext_otp_walk(mtd, from, len, retlen,
+				     buf, do_otp_write, 1);
+}
+
+static int cfi_intelext_lock_user_prot_reg(struct mtd_info *mtd,
+					   loff_t from, size_t len)
+{
+	size_t retlen;
+	return cfi_intelext_otp_walk(mtd, from, len, &retlen,
+				     NULL, do_otp_lock, 1);
+}
+
+static int cfi_intelext_get_fact_prot_info(struct mtd_info *mtd, 
+					   struct otp_info *buf, size_t len)
+{
+	size_t retlen;
+	int ret;
+
+	ret = cfi_intelext_otp_walk(mtd, 0, len, &retlen, (u_char *)buf, NULL, 0);
+	return ret ? : retlen;
+}
+
+static int cfi_intelext_get_user_prot_info(struct mtd_info *mtd,
+					   struct otp_info *buf, size_t len)
+{
+	size_t retlen;
+	int ret;
+
+	ret = cfi_intelext_otp_walk(mtd, 0, len, &retlen, (u_char *)buf, NULL, 1);
+	return ret ? : retlen;
+}
+
+#endif
+
+static void cfi_intelext_unlockall(struct mtd_info *mtd)
+{
+#if 0
+	int i;
+	for (i = 0; i <= mtd->numeraseregions; i++) {
+		printk("mtd->numeraseregions %x\n", mtd->numeraseregions);
+		int j;
+		printk("rumms\n");
+		printk("mtd->eraseregions[0].numblocks %x\n", 
+				mtd->eraseregions[i].numblocks);
+		printk("bumms\n");
+		for (j = 0; j < mtd->eraseregions[i].numblocks; j++){
+			printk("rumms\n");
+			printk("mtd->eraseregions[i].numblocks %x\n", 
+					mtd->eraseregions[i].numblocks);
+			printk("bumms\n");
+			printk("mtd->eraseregions[i].offset %x\n", 
+					mtd->eraseregions[i].offset);
+			printk("mtd->eraseregions[i].erasesize%x\n", 
+					mtd->eraseregions[i].erasesize);
+			mtd->unlock(mtd, mtd->eraseregions[i].offset +
+				    j * mtd->eraseregions[i].erasesize,
+				    mtd->eraseregions[i].erasesize);
+		}
+	} 
+#endif
+#if 1
+	cfi_intelext_unlock(mtd, 0, mtd->size);
+#endif
+}
+
+static void cfi_intelext_instantlock_add(struct mtd_info *mtd)
+{
+	if (mtd->flags & MTD_WRITEABLE)
+		cfi_intelext_unlockall(mtd);
+}
+
 static int cfi_intelext_suspend(struct mtd_info *mtd)
 {
 	struct map_info *map = mtd->priv;
@@ -2107,6 +2333,7 @@
 	struct cfi_private *cfi = map->fldrv_priv;
 	int i;
 	struct flchip *chip;
+	struct cfi_pri_intelext *extp = cfi->cmdset_priv;
 
 	for (i=0; i<cfi->numchips; i++) {
 	
@@ -2123,12 +2350,53 @@
 
 		spin_unlock(chip->mutex);
 	}
+
+
+	if (extp && (extp->FeatureSupport & (1 << 5)) &&
+	    (mtd->flags & MTD_WRITEABLE))
+		cfi_intelext_unlockall(mtd);	
+}
+
+static int cfi_intelext_reset(struct mtd_info *mtd)
+{
+	struct map_info *map = mtd->priv;
+	struct cfi_private *cfi = map->fldrv_priv;
+	int i, ret;
+
+	for (i=0; i < cfi->numchips; i++) {
+		struct flchip *chip = &cfi->chips[i];
+
+		/* force the completion of any ongoing operation
+		   and switch to array mode so any bootloader in 
+		   flash is accessible for soft reboot. */
+		spin_lock(chip->mutex);
+		ret = get_chip(map, chip, chip->start, FL_SYNCING);
+		if (!ret) {
+			map_write(map, CMD(0xff), chip->start);
+			chip->state = FL_READY;
+		}
+		spin_unlock(chip->mutex);
+	}
+
+	return 0;
+}
+
+static int cfi_intelext_reboot(struct notifier_block *nb, unsigned long val,
+			       void *v)
+{
+	struct mtd_info *mtd;
+
+	mtd = container_of(nb, struct mtd_info, reboot_notifier);
+	cfi_intelext_reset(mtd);
+	return NOTIFY_DONE;
 }
 
 static void cfi_intelext_destroy(struct mtd_info *mtd)
 {
 	struct map_info *map = mtd->priv;
 	struct cfi_private *cfi = map->fldrv_priv;
+	cfi_intelext_reset(mtd);
+	unregister_reboot_notifier(&mtd->reboot_notifier);
 	kfree(cfi->cmdset_priv);
 	kfree(cfi->cfiq);
 	kfree(cfi->chips[0].priv);





o_O regards, Konsti

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

* Konstantin Kletschke <lists@ku-gbr.de> [Wed, Jul 06, 2005 at 04:45:45PM +0200]:

> I had XIP in linux-2.6.11 running very well.

Meanwhile I found out a solution.
I switched off CONFIG_FPE_NWFPE=y which the XIP System seems to depend
on!

K

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

Am 2005-07-28 14:17 +0200 schrieb Konstantin Kletschke:

> I switched off CONFIG_FPE_NWFPE=y which the XIP System seems to depend
> on!

Also I realize, that the only toolchain creating a functional xipuImage
is gcc-3.3.3 with binutils-2.14.90.
Any newer components are creating a freezing xipuImage.

-- 
GPG KeyID EF62FCEF
Fingerprint: 13C9 B16B 9844 EC15 CC2E  A080 1E69 3FDA EF62 FCEF

Re: xip changes between 2.6.11 and 20050530?

[Nicolas Pitre]

On Thu, 28 Jul 2005, Konstantin Kletschke wrote:

> * Konstantin Kletschke <lists@ku-gbr.de> [Wed, Jul 06, 2005 at 04:45:45PM +0200]:
> 
> > I had XIP in linux-2.6.11 running very well.
> 
> Meanwhile I found out a solution.
> I switched off CONFIG_FPE_NWFPE=y which the XIP System seems to depend
> on!

It sure doesn't.

Something else must be going on.


Nicolas

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

* Nicolas Pitre <nico@cam.org> [Thu, Jul 28, 2005 at 11:01:24AM -0400]:

> > I switched off CONFIG_FPE_NWFPE=y which the XIP System seems to depend
> > on!
> 
> It sure doesn't.
> 
> Something else must be going on.


Hm.. funny. Beside the trouble with my toolchains the working one
creates definately only a bootable xipuImage if CONFIG_FPE_NWFPE is
enabled.  I wonder myself, why...

I rechecked again now.

Konsti

Re: xip changes between 2.6.11 and 20050530?

[Todd Poynor]

Konstantin Kletschke wrote:
> * Nicolas Pitre <nico@cam.org> [Thu, Jul 28, 2005 at 11:01:24AM -0400]:
> 
> 
>>>I switched off CONFIG_FPE_NWFPE=y which the XIP System seems to depend
>>>on!
>>
>>It sure doesn't.
>>
>>Something else must be going on.
> 
> 
> 
> Hm.. funny. Beside the trouble with my toolchains the working one
> creates definately only a bootable xipuImage if CONFIG_FPE_NWFPE is
> enabled.  I wonder myself, why...

Dunno about that option, but check the assembly code generated for the 
xip_* routines to see if any functions are called that are not marked 
__xipram?  If so, linux may be attempting to execute from flash memory 
while the flash is not in the proper (read array or whatever) state.

-- 
Todd

Re: xip changes between 2.6.11 and 20050530?

[Konstantin Kletschke]

Am 2005-07-28 11:15 -0700 schrieb Todd Poynor:

> >Hm.. funny. Beside the trouble with my toolchains the working one
> >creates definately only a bootable xipuImage if CONFIG_FPE_NWFPE is
> >enabled.  I wonder myself, why...
> 
> Dunno about that option, but check the assembly code generated for the 
> xip_* routines to see if any functions are called that are not marked 
> __xipram?  If so, linux may be attempting to execute from flash memory 

Ok, I can check that. 
To be sure:
Any xip_* function only calls functions marked __xipram (which are
located in RAM then hopefully also) and only __xipram marked functions
call functions located still in flash?

> while the flash is not in the proper (read array or whatever) state.

Something like that seems to habben, indeed...

Konsti

-- 
GPG KeyID EF62FCEF
Fingerprint: 13C9 B16B 9844 EC15 CC2E  A080 1E69 3FDA EF62 FCEF

Re: xip changes between 2.6.11 and 20050530?

[Todd Poynor]

Konstantin Kletschke wrote:

> To be sure:
> Any xip_* function only calls functions marked __xipram (which are
> located in RAM then hopefully also) and only __xipram marked functions
> call functions located still in flash?

Nico's comments in cfi_cmdset_0001.c describe it better than I did:

* No interrupt what so ever can be serviced while the flash isn't in array
  * mode.  This is ensured by the xip_disable() and xip_enable() functions
  * enclosing any code path where the flash is known not to be in array 
mode.
  * And within a XIP disabled code path, only functions marked with __xipram
  * may be called and nothing else (it's a good thing to inspect generated
  * assembly to make sure inline functions were actually inlined and 
that gcc
  * didn't emit calls to its own support functions). Also configuring 
MTD CFI
  * support to a single buswidth and a single interleave is also 
recommended.

The reason for "single buswidth and a single interleave" is to avoid 
division operations to compute various things driven by those two 
parameters, which will probably call a library function placed by 
default in flash to do the divide; if a single buswidth and interleave 
is configured then the derived values are compiled as constants and need 
not be computed at runtime.

Yes, it's all a little delicate. ;)

-- 
Todd