Skip to contentSkip to conversation

Fix recovery after failed write-buffer operation in cfi_cmdset_0002.c

DISCUSSION Applied Harald Nordgard-Hansen 2 participants 4 messages Nov 9, 2012 – Dec 3, 2012 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

Hi.

When working on a problem with some flash chips that lock up during write-buffer operations, I think there may be a bug in the linux handling of chips using cfi_cmdset_0002.c.

The datasheets I have found for a number of these chips all specify that when aborting a write-buffer command, it is not enough to use the standard reset. Rather a "write-to-buffer-reset command" is needed. This command is quite similar for all chips, the main variance seem to be if the final 0xF0 can go to any address or must go to addr_unlock1.

The bug is then in the recovery handling when timing out at the end of do_write_buffer, where using the normal reset command is not sufficient.

Without this change, if the write-buffer command fails then any following operations on the flash also fail.

The small patch here should apply against just about all kernels I've seen over the last 5 years, the code has not changed in this area for a long time...

-Harald Nordgård-Hansen

--------
--- a/drivers/mtd/chips/cfi_cmdset_0002.c 2012-05-21
13:46:28.679794861 +0200
+++ b/drivers/mtd/chips/cfi_cmdset_0002.c 2012-10-30
18:27:49.939109556 +0100
@@ -1536,8 +1536,10 @@

                UDELAY(map, chip, adr, 1);
        }
-       /* reset on all failures. */
-       map_write( map, CMD(0xF0), chip->start );
+       /* write-to-buffer-reset on all failures. */
+       cfi_send_gen_cmd(0xAA, cfi->addr_unlock1, chip->start, map, cfi,

cfi->device_type, NULL);
+ cfi_send_gen_cmd(0x55, cfi->addr_unlock2, chip->start, map, cfi,
cfi->device_type, NULL);
+ cfi_send_gen_cmd(0xF0, cfi->addr_unlock1, chip->start, map, cfi,

cfi->device_type, NULL);
        xip_enable(map, chip, adr);
        /* FIXME - should have reset delay before continuing */

--------

Signature · 1 line
Harald Nordgård-Hansen
#1 lore raw

Hi,

Quoted 13 lines · Harald Nordgard-Hansen

When working on a problem with some flash chips that lock up during write-buffer operations, I think there may be a bug in the linux handling of chips using cfi_cmdset_0002.c.

The datasheets I have found for a number of these chips all specify that when aborting a write-buffer command, it is not enough to use the standard reset. Rather a "write-to-buffer-reset command" is needed. This command is quite similar for all chips, the main variance seem to be if the final 0xF0 can go to any address or must go to addr_unlock1.

The bug is then in the recovery handling when timing out at the end of do_write_buffer, where using the normal reset command is not sufficient.

Quoted 8 lines

Without this change, if the write-buffer command fails then any following operations on the flash also fail.

The small patch here should apply against just about all kernels I've seen over the last 5 years, the code has not changed in this area for a long time...

-Harald Nordgård-Hansen

Hi, your patch is not applicable, it is line-wrapped. Can you please send a patch I can easily apply?

Review line 1540

Quoted 12 lines

--------
--- a/drivers/mtd/chips/cfi_cmdset_0002.c 2012-05-21
13:46:28.679794861 +0200
+++ b/drivers/mtd/chips/cfi_cmdset_0002.c 2012-10-30
18:27:49.939109556 +0100

(quoted diff) +0 -2
@@ -1536,8 +1536,10 @@
                UDELAY(map, chip, adr, 1);
        }
 
-       /* reset on all failures. */
-       map_write( map, CMD(0xF0), chip->start );

Would be nice to put a short comment about what you do and why, may be some reference as well. Just to make sure if someone reads the code, he/she has some clue what are these about.

Thanks!

Quoted 9 lines
(quoted diff) +4 -0
+       /* write-to-buffer-reset on all failures. */
+       cfi_send_gen_cmd(0xAA, cfi->addr_unlock1, chip->start, map, cfi,
 cfi->device_type, NULL);
+       cfi_send_gen_cmd(0x55, cfi->addr_unlock2, chip->start, map, cfi,
 cfi->device_type, NULL);
+       cfi_send_gen_cmd(0xF0, cfi->addr_unlock1, chip->start, map, cfi,
 cfi->device_type, NULL);
        xip_enable(map, chip, adr);
        /* FIXME - should have reset delay before continuing */
Signature · 2 lines
Best Regards,
Artem Bityutskiy
#2 lore raw
Quoted 2 lines · Artem Bityutskiy

Hi, your patch is not applicable, it is line-wrapped. Can you please send a patch I can easily apply?

Sorry about that. It's been way too many years since I regularly sent patches in mail. If this attempt does not work better, I'll just do it all manually instead of trying to use these new-fangled MUAs. :-)

Quoted 3 lines

Would be nice to put a short comment about what you do and why, may be some reference as well. Just to make sure if someone reads the code, he/she has some clue what are these about.

Hopefully in place now.

-Harald

Signature · 1 line
Harald Nordgård-Hansen
drivers/mtd/chips/cfi_cmdset_0002.c +10 -2
diff -ru a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
@@ -1536,8 +1536,16 @@
 		UDELAY(map, chip, adr, 1);
 	}
 
-	/* reset on all failures. */
-	map_write( map, CMD(0xF0), chip->start );
+	/*
+	 * Recovery from write-buffer programming failures requires
+	 * the write-to-buffer-reset sequence.  Since the last part
+	 * of the sequence also works as a normal reset, we can run
+	 * the same commands regardless of why we are here.
+	 * See e.g. http://www.spansion.com/Support/Application%20Notes/MirrorBit_Write_Buffer_Prog_Page_Buffer_Read_AN.pdf
+	 */
+	cfi_send_gen_cmd(0xAA, cfi->addr_unlock1, chip->start, map, cfi, cfi->device_type, NULL);
+	cfi_send_gen_cmd(0x55, cfi->addr_unlock2, chip->start, map, cfi, cfi->device_type, NULL);
+	cfi_send_gen_cmd(0xF0, cfi->addr_unlock1, chip->start, map, cfi, cfi->device_type, NULL);
 	xip_enable(map, chip, adr);
 	/* FIXME - should have reset delay before continuing */
 
#3 lore raw
Quoted 5 lines · Harald Nordgard-Hansen
Quoted 3 lines

Would be nice to put a short comment about what you do and why, may be some reference as well. Just to make sure if someone reads the code, he/she has some clue what are these about.

Hopefully in place now.

Pushed to l2-mtd.git, thanks.

However, I had to do the following:

1. Make a commit message for you by copying parts of your e-mail.
2. Create a Signe-of-by for you.
3. Fix these checkpatch.pl warnings:

WARNING:LONG_LINE: line over 80 characters #48: FILE: drivers/mtd/chips/cfi_cmdset_0002.c:1546:
+ cfi_send_gen_cmd(0xAA, cfi->addr_unlock1, chip->start, map, cfi, cfi->device_type, NULL);

WARNING:LONG_LINE: line over 80 characters #49: FILE: drivers/mtd/chips/cfi_cmdset_0002.c:1547:
+ cfi_send_gen_cmd(0x55, cfi->addr_unlock2, chip->start, map, cfi, cfi->device_type, NULL);

WARNING:LONG_LINE: line over 80 characters #50: FILE: drivers/mtd/chips/cfi_cmdset_0002.c:1548:
+ cfi_send_gen_cmd(0xF0, cfi->addr_unlock1, chip->start, map, cfi, cfi->device_type, NULL);

Signature · 2 lines
Best Regards,
Artem Bityutskiy