Skip to contentSkip to conversation

nand sector size is zero in jffs2 cmds

U-Boot Derek Ou 3 participants 8 messages Dec 23, 2008 – Jan 27, 2009

lore
Share & tools
lore

Derek Ou

author
#0 lore raw

Hi, all,

When I tested jffs2 command for NAND, I found that jffs2_1pass.c::
jffs2_1pass_build_lists() fails at

    u32 nr_sectors = part->size/part->sector_size;

because part->sector_size is zero. And cause the "### ERROR ### Please RESET the board ###" error when entering the for loop (i < nr_sectors), I think.

It looks like the sector_size was not updated in mtdparts_init(). According to cmd_jffs2.c, the mtdparts_init in the #ifdef CONFIG_JFFS2_CMDLINE section calls the parse_mtdparts() and eventually reach the following lines in part_validate_nand():

    nand = &nand_info[id->num];
    part->sector_size = nand->erasesize;

However, the other mtdparts_init(), which is not in the #ifdef CONFIG_JFFS2_CMDLINE section, does not call parse_mtdparts() and it may lead to the part->sector_size not initialized.

I am not totally familiar with this jffs2 and nand code. Can someone review this? Thanks.
Derek

author
#1 lore raw

The "### ERROR ### Please RESET the board ###" error and the zero NAND sector size problem can be fixed by the following patch. However, the jffs2 scanning routine does not skip the bad blocks in NAND though.

Derek
************* patch **************
index 7866c80..f6966f3 100644

common/cmd_jffs2.c +2 -0
@@ -1853,6 +1853,8 @@ int mtdparts_init(void)
                if (part->size == SIZE_REMAINING)
                        part->size = id->size - part->offset;
 
+               part->sector_size = nand_info[id->num].erasesize;
+
                DEBUGF("part  : name = %s, size = 0x%08lx, offset = 

0x%08lx\n",

                                part->name, part->size, part->offset);
Quoted 23 lines · Derek Ou

Hi, all,

When I tested jffs2 command for NAND, I found that jffs2_1pass.c::
jffs2_1pass_build_lists() fails at
u32 nr_sectors = part->size/part->sector_size;
because part->sector_size is zero. And cause the "### ERROR ### Please RESET the board ###" error when entering the for loop (i < nr_sectors), I think.

It looks like the sector_size was not updated in mtdparts_init(). According to cmd_jffs2.c, the mtdparts_init in the #ifdef CONFIG_JFFS2_CMDLINE section calls the parse_mtdparts() and eventually reach the following lines in part_validate_nand():

   nand = &nand_info[id->num];
   part->sector_size = nand->erasesize;

However, the other mtdparts_init(), which is not in the #ifdef CONFIG_JFFS2_CMDLINE section, does not call parse_mtdparts() and it may lead to the part->sector_size not initialized.

I am not totally familiar with this jffs2 and nand code. Can someone review this? Thanks.
Derek

#2 lore raw

Dear Derek Ou,

Quoted 10 lines · In message <[email protected]> you wrote:

The "### ERROR ### Please RESET the board ###" error and the zero NAND sector size problem can be fixed by the following patch. However, the jffs2 scanning routine does not skip the bad blocks in NAND though.

Derek
************* patch **************
index 7866c80..f6966f3 100644

common/cmd_jffs2.c int mtdparts_init(void) · line 1860

Quoted 8 lines
common/cmd_jffs2.c +2 -0
@@ -1853,6 +1853,8 @@ int mtdparts_init(void)
                if (part->size == SIZE_REMAINING)
                        part->size = id->size - part->offset;
 
+               part->sector_size = nand_info[id->num].erasesize;
+
                DEBUGF("part  : name = %s, size = 0x%08lx, offset = 
 0x%08lx\n",
                                part->name, part->size, part->offset);

Can you please submit a proper patch (after fixing your mailer configuration which wraps long lines, thus corrupting the patch)?

See http://www.denx.de/wiki/U-Boot/Patches for instructions.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Perl itself is  usually  pretty  good  about  telling  you  what  you
shouldn't do. :-)     - Larry Wall in <[email protected]>
author
#3 lore raw

Hi, Wolfgang,

The following patch was tested some time ago. I am in the middle of another task and can not
verify this patch with the latest code. Should I just reformat my patch and submit it or should I
wait till I can test it with the latest code?

Also, I found that jffs2 scanning routine does not skip the bad blocks in NAND in my
previous test. Does anybody else see this behavior?

BTW, what is the best way to submit patches? I tried git-format-patch approach described in
http://www.denx.de/wiki/U-Boot/GnatsEmail
But git-send-email is not easy to use. I tried just copy and paste but thunderbird convert all the
tabs into spaces. I ended up using copy and paste and webmail interface to send out another
patch. But I will love to hear a better way of doing it.

Thanks,
Derek

Quoted 30 lines · Wolfgang Denk

Dear Derek Ou,

Quoted 11 lines · In message <[email protected]> you wrote:

The "### ERROR ### Please RESET the board ###" error and the zero NAND sector size problem can be fixed by the following patch. However, the jffs2 scanning routine does not skip the bad blocks in NAND though.

Derek
************* patch **************
index 7866c80..f6966f3 100644

common/cmd_jffs2.c int mtdparts_init(void) · line 1860

Quoted 8 lines
common/cmd_jffs2.c +2 -0
@@ -1853,6 +1853,8 @@ int mtdparts_init(void)
                if (part->size == SIZE_REMAINING)
                        part->size = id->size - part->offset;
 
+               part->sector_size = nand_info[id->num].erasesize;
+
                DEBUGF("part  : name = %s, size = 0x%08lx, offset = 
 0x%08lx\n",
                                part->name, part->size, part->offset);

Can you please submit a proper patch (after fixing your mailer configuration which wraps long lines, thus corrupting the patch)?

See http://www.denx.de/wiki/U-Boot/Patches for instructions.

Best regards,
Wolfgang Den

#4 lore raw

Dear Derek Ou,

Quoted 6 lines · In message <[email protected]> you wrote:

The following patch was tested some time ago. I am in the middle of another task and can not
verify this patch with the latest code. Should I just reformat my patch and submit it or should I
wait till I can test it with the latest code?

I think the fix is obvious, so feel free to repost after adding your signed-off-by line and fixing yourmailer not to cwrap long lines. I recommend you use git-send-email to submit the patch.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Business is like a wheelbarrow. Nothing ever happens until you  start
pushing.
author
#5 lore raw

[U-Boot] env_nand malloc not released

In reply to Derek Ou

Hi, all,

In the common/env_nand.c, function env_relocate_spec (for CONFIG_ENV_OFFSET_REDUND) malloc two memory areas in line 296 and 297. If it hits line 308 when both environment areas not found, it will return without freeing the two memory areas. Is this a bug or I am missing something?

Thanks,
Derek

#6 lore raw

[U-Boot] env_nand malloc not released

Quoted 7 lines · Derek Ou

Hi, all,

In the common/env_nand.c, function env_relocate_spec (for CONFIG_ENV_OFFSET_REDUND) malloc two memory areas in line 296 and 297. If it hits line 308 when both environment areas not found, it will return without freeing the two memory areas. Is this a bug or I am missing something?

Looks like a bug.

-Scott

#7 lore raw

[U-Boot] env_nand malloc not released

In reply to Derek Ou

Dear Derek Ou,

Quoted 6 lines · In message <[email protected]> you wrote:

In the common/env_nand.c, function env_relocate_spec (for CONFIG_ENV_OFFSET_REDUND) malloc two memory areas in line 296 and 297. If it hits line 308 when both environment areas not found, it will return without freeing the two memory areas. Is this a bug or I am missing something?

Can you please submit a patch?

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Little known fact about Middle Earth:   The Hobbits had a very sophi-
sticated computer network!   It was a Tolkien Ring...