Patchwork mtd cs553x_nand: Initialise ecc.strength before nand_scan()

login
register
mail settings
Submitter Nathan Williams
Date Nov. 21, 2012, 11:42 p.m.
Message ID <1353541372.12585.14.camel@laptop>
Download mbox | patch
Permalink /patch/200885/
State Accepted
Commit d1f3b65d2d6fdb4bf0edd4b67e86e191af48daee
Headers show

Comments

Nathan Williams - Nov. 21, 2012, 11:42 p.m.
Loading cs553x_nand with Hynix H27U1G8F2BTR NAND flash causes this bug:

kernel BUG at drivers/mtd/nand/nand_base.c:3345!
invalid opcode: 0000 [#1] 
Modules linked in: cs553x_nand(+) vfat fat usb_storage ehci_hcd usbcore usb_comr
Pid: 436, comm: modprobe Not tainted 3.6.7 #1  
EIP: 0060:[<c118d205>] EFLAGS: 00010296 CPU: 0
EIP is at nand_scan_tail+0x64c/0x69c
EAX: 00000034 EBX: cea6ed98 ECX: 00000000 EDX: 00000000
ESI: cea6ec00 EDI: cea6ec00 EBP: 20000000 ESP: cdd17e48                         
 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068                                   
CR0: 8005003b CR2: 0804e119 CR3: 0d850000 CR4: 00000090                         
DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000                         
DR6: ffff0ff0 DR7: 00000400                                                     
Process modprobe (pid: 436, ti=cdd16000 task=cdd1c320 task.ti=cdd16000)         
Stack:                                                                          
 c12e962c c118f7ef 00000003 cea6ed98 d014b25c 20000000 fffff007 00000001        
 00000000 cdd53b00 d014b000 c1001021 cdd53b00 d01493c0 cdd53b00 cdd53b00        
 d01493c0 c1047f83 d014b4a0 00000000 cdd17f9c ce4be454 cdd17f48 cdd1c320        
Call Trace:                                                                     
 [<c118f7ef>] ? nand_scan+0x1b/0x4d                                             
 [<d014b25c>] ? init_module+0x25c/0x2de [cs553x_nand]                           
 [<d014b000>] ? 0xd014afff                                                      
 [<c1001021>] ? do_one_initcall+0x21/0x111                                      
 [<c1047f83>] ? sys_init_module+0xe4/0x1261                                     
 [<c1031207>] ? task_work_run+0x36/0x43                                         
 [<c1265ced>] ? syscall_call+0x7/0xb                                            
Code: fa ff ff c7 86 d8 00 00 00 01 00 00 00 e9 5f fc ff ff 68 f8 26 2e c1 e8 a7
EIP: [<c118d205>] nand_scan_tail+0x64c/0x69c SS:ESP 0068:cdd17e48

Initialising ecc.strength before the call to nand_scan() fixes this.

Signed-off-by: Nathan Williams <nathan@traverse.com.au>
---
 drivers/mtd/nand/cs553x_nand.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)
Brian Norris - Nov. 22, 2012, 4:08 a.m.
On Wed, Nov 21, 2012 at 3:42 PM, Nathan Williams <nathan@traverse.com.au> wrote:
> Loading cs553x_nand with Hynix H27U1G8F2BTR NAND flash causes this bug:
>
> kernel BUG at drivers/mtd/nand/nand_base.c:3345!
...
> Initialising ecc.strength before the call to nand_scan() fixes this.

Acked-by: Brian Norris <computersforpeace@gmail.com>

BTW, I made a quick read through of the other NAND drivers to look for
similar errors. I think this is the only one where the only
ecc.strenght initialization occurred after the nand_scan() or
nand_scan_tail() call.

Thanks,
Brian
Artem Bityutskiy - Nov. 22, 2012, 11:09 a.m.
On Thu, 2012-11-22 at 10:42 +1100, Nathan Williams wrote:
> Loading cs553x_nand with Hynix H27U1G8F2BTR NAND flash causes this bug:
> 
> kernel BUG at drivers/mtd/nand/nand_base.c:3345!

Pushed to l2-mtd.git, added

Cc: stable@vgen.kernel.org [3.4+]

Thanks!
Mike Dunn - Nov. 27, 2012, 6:11 p.m.
On 11/21/2012 08:08 PM, Brian Norris wrote:
> On Wed, Nov 21, 2012 at 3:42 PM, Nathan Williams <nathan@traverse.com.au> wrote:
>> Loading cs553x_nand with Hynix H27U1G8F2BTR NAND flash causes this bug:
>>
>> kernel BUG at drivers/mtd/nand/nand_base.c:3345!
> ...
>> Initialising ecc.strength before the call to nand_scan() fixes this.
> 
> Acked-by: Brian Norris <computersforpeace@gmail.com>


Acked-by: Mike Dunn <mikedunn@newsguy.com>


> 
> BTW, I made a quick read through of the other NAND drivers to look for
> similar errors. I think this is the only one where the only
> ecc.strenght initialization occurred after the nand_scan() or
> nand_scan_tail() call.


Thanks for that Brian.  I thought I did this when another driver had a similiar
problem a while back.  Maybe I just intended to :(

Mike

Patch

diff --git a/drivers/mtd/nand/cs553x_nand.c b/drivers/mtd/nand/cs553x_nand.c
index adb6c3e..2cdeab8 100644
--- a/drivers/mtd/nand/cs553x_nand.c
+++ b/drivers/mtd/nand/cs553x_nand.c
@@ -237,6 +237,7 @@  static int __init cs553x_init_one(int cs, int mmio, unsigned long adr)
 	this->ecc.hwctl  = cs_enable_hwecc;
 	this->ecc.calculate = cs_calculate_ecc;
 	this->ecc.correct  = nand_correct_data;
+	this->ecc.strength = 1;
 
 	/* Enable the following for a flash based bad block table */
 	this->bbt_options = NAND_BBT_USE_FLASH;
@@ -247,8 +248,6 @@  static int __init cs553x_init_one(int cs, int mmio, unsigned long adr)
 		goto out_ior;
 	}
 
-	this->ecc.strength = 1;
-
 	new_mtd->name = kasprintf(GFP_KERNEL, "cs553x_nand_cs%d", cs);
 
 	cs553x_mtd[cs] = new_mtd;