[MTD] cfi_cmdset_0002.c: Add Macronix CFI V1.0 TopBottom detection

This patch adds TopBottom detection for most Macronix chips with CFI
V1.0.

The main purpose of this patch is to add detection of the MX29LV400C B
used on the LaCie Ethernet Disk mini V2 NAS.

It detects the following parts correctly:-
MX28F640C3B T
MX29LV002C  B
MX29LV002NC B
MX29LV004C  T
MX29LV400C  T/B
MX29LV800C  T/B
MX29LV160C  T/B
MX29SL800C  T/B
MX29SL802C  T/B

It detects the following uniform part as bottom but it should work
correctly:-
MX29LV040C

For T parts it correctly reverses the erase block table.
For other parts it avoids the bogus "Assuming top." message.

It does not detect the following correctly:-
MX28F640C3B B
MX29LV002C  T
MX29LV002NC T
MX29LV004C  B
MX29SL400C  T/B
MX29SL402C  T/B

If desired I could supply a more complicated patch to handle these as
well.

Only the MX29LV400C B has been physically tested; others were checked
against their data sheets.

Signed-off-by: Christopher Moore <moore@free.fr>
---
 drivers/mtd/chips/cfi_cmdset_0002.c |   36
+++++++++++++++++++++++++++++++++++
 1 files changed, 36 insertions(+), 0 deletions(-)

  * This code is GPL
@@ -43,6 +45,7 @@
 
 #define MANUFACTURER_AMD	0x0001
 #define MANUFACTURER_ATMEL	0x001F
+#define MANUFACTURER_MACRONIX	0x00C2
 #define MANUFACTURER_SST	0x00BF
 #define SST49LF004B	        0x0060
 #define SST49LF040B	        0x0050
@@ -144,12 +147,44 @@ static void fixup_amd_bootblock(struct mtd_info
*mtd, void* param)
 
 	if (((major << 8) | minor) < 0x3131) {
 		/* CFI version 1.0 => don't trust bootloc */
+
+		DEBUG(MTD_DEBUG_LEVEL1,
+			"%s: JEDEC Vendor ID is 0x%02X Device ID is 0x%02X\n",
+			map->name, cfi->mfr, cfi->id);
+
+		/* AFAICS all 29LV400 with a bottom boot block have a device ID
+		 * of 0x22BA in 16-bit mode and 0xBA in 8-bit mode.
+		 * These were badly detected as they have the 0x80 bit set
+		 * so treat them as a special case.
+		 */
+		if (((cfi->id == 0xBA) || (cfi->id == 0x22BA)) &&
+
+			/* Macronix added CFI to their 2nd generation
+			 * MX29LV400C B/T but AFAICS no other 29LV400 (AMD,
+			 * Fujitsu, Spansion, EON, ESI and older Macronix)
+			 * has CFI.
+			 *
+			 * Therefore also check the manufacturer.
+			 * This reduces the risk of false detection due to
+			 * the 8-bit device ID.
+			 */
+			(cfi->mfr == MANUFACTURER_MACRONIX)) {
+			DEBUG(MTD_DEBUG_LEVEL1,
+				"%s: Macronix MX29LV400C with bottom boot block"
+				" detected\n", map->name);
+			extp->TopBottom = 2;	/* bottom boot */
+		} else
 		if (cfi->id & 0x80) {
 			printk(KERN_WARNING "%s: JEDEC Device ID is 0x%02X. Assuming broken
CFI table.\n", map->name, cfi->id);
 			extp->TopBottom = 3;	/* top boot */
 		} else {
 			extp->TopBottom = 2;	/* bottom boot */
 		}
+
+		DEBUG(MTD_DEBUG_LEVEL1,
+			"%s: AMD CFI PRI V%c.%c has no boot block field;"
+			" deduced %s from Device ID\n", map->name, major, minor,
+			extp->TopBottom == 2 ? "bottom" : "top");
 	}
 }
 #endif
@@ -243,6 +278,7 @@ static struct cfi_fixup cfi_fixup_table[] = {
 	{ CFI_MFR_ATMEL, CFI_ID_ANY, fixup_convert_atmel_pri, NULL },
 #ifdef AMD_BOOTLOC_BUG
 	{ CFI_MFR_AMD, CFI_ID_ANY, fixup_amd_bootblock, NULL },
+	{ MANUFACTURER_MACRONIX, CFI_ID_ANY, fixup_amd_bootblock, NULL },
 #endif
 	{ CFI_MFR_AMD, 0x0050, fixup_use_secsi, NULL, },
 	{ CFI_MFR_AMD, 0x0053, fixup_use_secsi, NULL, },

On Thu, 2008-10-09 at 23:55 +0200, Christopher Moore wrote:
> This patch adds TopBottom detection for most Macronix chips with CFI
> V1.0.

patching file drivers/mtd/chips/cfi_cmdset_0002.c
patch: **** malformed patch at line 101: CFI V1.0

Did you not test this after you received it back from the list?

David Woodhouse a écrit :
> On Thu, 2008-10-09 at 23:55 +0200, Christopher Moore wrote:
>   
>> This patch adds TopBottom detection for most Macronix chips with CFI
>> V1.0.
>>     
>
> patching file drivers/mtd/chips/cfi_cmdset_0002.c
> patch: **** malformed patch at line 101: CFI V1.0
>
> Did you not test this after you received it back from the list?
>   

(Wiping egg off face) Please accept my apologies; I am afraid that the
line must have been wrapped by the mailer I used  :( 

I ran checkpatch.pl on the patch before sending it.
I even sent the patch to myself first.
The worst is that I did notice that some lines were wrapped to about 72
columns.
However as I had previously used the same mailer to send patches to
other people and to the LAK mailing list without problems, I figured
that the wrapping was happening at the receiving end.
With the intention of improving the presentation for others using the
same mailer, I even modified the commit comment part, which I had
written in 80 columns, down to 72 columns and tried again.

I have now experimented with "git send-email" and the result seems clean
so I shall use this in the future. The only thing that I am now worried
about is getting the headers right so that I do not break the thread.

I shall try sending the patch again later and I apologize in advance in
case it breaks the thread  :( 

Sorry again for wasting your time.

Cheers,
Chris