Skip to contentSkip to conversation

enable erase-suspend-program for CFI cmdset_0002

PATCH In review Norbert van Bolhuis 3 participants 9 messages Nov 19, 2009 – Nov 24, 2009 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 enable erase-suspend-program for CFI cmdset_0002

erase-suspend for writing is required to avoid blocking applications that wish to write some data (to a NOR block other than the one being erased). Particularly, it solves some huge delays that an application (which writes to a UBIFS) will experience if UBI attaches to empty NOR flash. In this case the UBI background thread will erase a lot of blocks and the application can be blocked for minutes because of the "MTD/CFI chip lock".
This feature has been disabled for years. Maybe this was because the old code turned it on for erase-suspend read-only chips also (cfip->EraseSuspend & 0x1). This is wrong and corrected now.
I tested this patch and it seems to work fine.

---
 drivers/mtd/chips/cfi_cmdset_0002.c |   16 +++-------------
 1 files changed, 3 insertions(+), 13 deletions(-)
drivers/mtd/chips/cfi_cmdset_0002.c +3 -13
index 94bb61e..1d49e18 100644
@@ -490,10 +490,6 @@ static struct mtd_info *cfi_amdstd_setup(struct mtd_info *mtd)
 	}
 #endif
 
-	/* FIXME: erase-suspend-program is broken.  See
-	   http://lists.infradead.org/pipermail/linux-mtd/2003-December/009001.html */
-	printk(KERN_NOTICE "cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness.\n");
-
 	__module_get(THIS_MODULE);
 	return mtd;
 
@@ -589,15 +585,9 @@ static int get_chip(struct map_info *map, struct flchip *chip, unsigned long adr
 		return 0;
 
 	case FL_ERASING:
-		if (mode == FL_WRITING) /* FIXME: Erase-suspend-program appears broken. */
-			goto sleep;
-
-		if (!(   mode == FL_READY
-		      || mode == FL_POINT
-		      || !cfip
-		      || (mode == FL_WRITING && (cfip->EraseSuspend & 0x2))
-		      || (mode == FL_WRITING && (cfip->EraseSuspend & 0x1)
-		    )))
+		if (!cfip || !(cfip->EraseSuspend & (0x1|0x2)) ||
+		    !(mode == FL_READY || mode == FL_POINT ||
+		    (mode == FL_WRITING && (cfip->EraseSuspend & 0x2))))
 			goto sleep;
 
 		/* We could check to see if we're trying to access the sector
Signature · 1 line
1.5.2.2
#1 lore raw
Quoted 12 lines · Norbert van Bolhuis

erase-suspend for writing is required to avoid blocking applications that wish to write some data (to a NOR block other than the one being erased). Particularly, it solves some huge delays that an application (which writes to a UBIFS) will experience if UBI attaches to empty NOR flash. In this case the UBI background thread will erase a lot of blocks and the application can be blocked for minutes because of the "MTD/CFI chip lock".
This feature has been disabled for years. Maybe this was because the old code turned it on for erase-suspend read-only chips also (cfip->EraseSuspend & 0x1). This is wrong and corrected now.
I tested this patch and it seems to work fine.

Quoted 1 line

FYI: I have no experience with non-Intel parts and no good knowledge of the cmdset_0002 code. So I can't review this.

Quoted 41 lines
---
 drivers/mtd/chips/cfi_cmdset_0002.c |   16 +++-------------
 1 files changed, 3 insertions(+), 13 deletions(-)
drivers/mtd/chips/cfi_cmdset_0002.c +3 -13
index 94bb61e..1d49e18 100644
@@ -490,10 +490,6 @@ static struct mtd_info *cfi_amdstd_setup(struct mtd_info *mtd)
 	}
 #endif
 
-	/* FIXME: erase-suspend-program is broken.  See
-	   http://lists.infradead.org/pipermail/linux-mtd/2003-December/009001.html */
-	printk(KERN_NOTICE "cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness.\n");
-
 	__module_get(THIS_MODULE);
 	return mtd;
 
@@ -589,15 +585,9 @@ static int get_chip(struct map_info *map, struct flchip *chip, unsigned long adr
 		return 0;
 
 	case FL_ERASING:
-		if (mode == FL_WRITING) /* FIXME: Erase-suspend-program appears broken. */
-			goto sleep;
-
-		if (!(   mode == FL_READY
-		      || mode == FL_POINT
-		      || !cfip
-		      || (mode == FL_WRITING && (cfip->EraseSuspend & 0x2))
-		      || (mode == FL_WRITING && (cfip->EraseSuspend & 0x1)
-		    )))
+		if (!cfip || !(cfip->EraseSuspend & (0x1|0x2)) ||
+		    !(mode == FL_READY || mode == FL_POINT ||
+		    (mode == FL_WRITING && (cfip->EraseSuspend & 0x2))))
 			goto sleep;
 
 		/* We could check to see if we're trying to access the sector
Signature · 1 line
1.5.2.2
#2 lore raw
Quoted 13 lines · Nicolas Pitre
Quoted 11 lines · Norbert van Bolhuis

erase-suspend for writing is required to avoid blocking applications that wish to write some data (to a NOR block other than the one being erased). Particularly, it solves some huge delays that an application (which writes to a UBIFS) will experience if UBI attaches to empty NOR flash. In this case the UBI background thread will erase a lot of blocks and the application can be blocked for minutes because of the "MTD/CFI chip lock".
This feature has been disabled for years. Maybe this was because the old code turned it on for erase-suspend read-only chips also (cfip->EraseSuspend & 0x1). This is wrong and corrected now.
I tested this patch and it seems to work fine.

Quoted 4 lines
Quoted 1 line

FYI: I have no experience with non-Intel parts and no good knowledge of the cmdset_0002 code. So I can't review this.

OK.

so, who's approving/reviewing patches for cmdset_0002. Nobody ?

#3 lore raw
Quoted 16 lines · Norbert van Bolhuis
Quoted 15 lines · Nicolas Pitre
Quoted 12 lines · Norbert van Bolhuis

erase-suspend for writing is required to avoid blocking applications that wish to write some data (to a NOR block other than the one being erased). Particularly, it solves some huge delays that an application (which writes to a UBIFS) will experience if UBI attaches to empty NOR flash. In this case the UBI background thread will erase a lot of blocks and the application can be blocked for minutes because of the "MTD/CFI chip lock".
This feature has been disabled for years. Maybe this was because the old code turned it on for erase-suspend read-only chips also (cfip->EraseSuspend & 0x1). This is wrong and corrected now.
I tested this patch and it seems to work fine.

Quoted 7 lines
Quoted 3 lines

FYI: I have no experience with non-Intel parts and no good knowledge of the cmdset_0002 code. So I can't review this.

OK.

so, who's approving/reviewing patches for cmdset_0002. Nobody ?

I think we can just take it, since you tested it and confirm it works and solves a real problem. I'll put it to my l2 tree later, do not have time right now.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#4 lore raw
In reply to Norbert van Bolhuis
Quoted 11 lines · Norbert van Bolhuis

erase-suspend for writing is required to avoid blocking applications that wish to write some data (to a NOR block other than the one being erased). Particularly, it solves some huge delays that an application (which writes to a UBIFS) will experience if UBI attaches to empty NOR flash. In this case the UBI background thread will erase a lot of blocks and the application can be blocked for minutes because of the "MTD/CFI chip lock".
This feature has been disabled for years. Maybe this was because the old code turned it on for erase-suspend read-only chips also (cfip->EraseSuspend & 0x1). This is wrong and corrected now.
I tested this patch and it seems to work fine.

Quoted 1 line

I did not follow the patch conversation too closely - but should you be the patch author or Joakim? If you took it as it is - then it is Joakim. If you modified it, then at least Joakim's name should be in the commit log.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#5 lore raw
Quoted 5 lines · Artem Bityutskiy

I did not follow the patch conversation too closely - but should you be the patch author or Joakim? If you took it as it is - then it is Joakim. If you modified it, then at least Joakim's name should be in the commit log.

I also removed the "printk(ERR_NOTICE ...", so I guess I modified it.

you want me to re-submit the patch with CC or Acked-by Joakim ?

#6 lore raw
Quoted 11 lines · Norbert van Bolhuis
Quoted 5 lines · Artem Bityutskiy

I did not follow the patch conversation too closely - but should you be the patch author or Joakim? If you took it as it is - then it is Joakim. If you modified it, then at least Joakim's name should be in the commit log.

I also removed the "printk(ERR_NOTICE ...", so I guess I modified it.

you want me to re-submit the patch with CC or Acked-by Joakim ?

If you just modify slightly, I can just make Joakim the author, which sounds fair. And put a not that you tweaked it - this is the common practice. Is this ok with you?

No need to re-send, I can do these things myself.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#7 lore raw
Quoted 5 lines · Artem Bityutskiy

If you just modify slightly, I can just make Joakim the author, which sounds fair. And put a not that you tweaked it - this is the common practice. Is this ok with you?

yes.

Quoted 2 lines

No need to re-send, I can do these things myself.

ok, good.

#8 lore raw
Quoted 8 lines · Norbert van Bolhuis
Quoted 5 lines · Artem Bityutskiy

If you just modify slightly, I can just make Joakim the author, which sounds fair. And put a not that you tweaked it - this is the common practice. Is this ok with you?

Quoted 6 lines

yes.

Quoted 2 lines

No need to re-send, I can do these things myself.

ok, good.

Pushed the patch to my l2 tree:

http://git.infradead.org/users/dedekind/l2-mtd-2.6.git/commit/6c8793f0c4eae1ce45f654a0fb6cf0b94d7061d6

Thanks!

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)