JFFS2 Error or Maybe Flash Driver Error

Hi All,

I pulled the JFFS2 code from CVS. And Ported 2.6.12-omap1 to p2-sample. When P2 sample run for a while, Sometimes The following message will show.

But when I boot again, it is ok again.

Does anybody meet the same problem ?

Node CRC 00800080 != calculated CRC 4bf2ff8d for node at 00676518 JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9dfa60. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x0000059c)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9deb3c. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000f24)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9ddcbc. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000e80)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9dce64. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000e58)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9dbf3c. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000f28)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9db5dc. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000960)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9da914. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000cc8)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9d9c94. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000c80)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9d93c8. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x000008cc)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9d883c. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x00000b8c)
JFFS2 notice: (247) read_unknown: node header CRC failed at 0x9d7f74. But it must have been OK earl.Node totlen on flash (0x00800080) != totlen from node ref (0x000008c8)

In our kernel source code, FILE drivers/mtd/chips/cfi_cmdset_0001.c , we change like below to fix a bug.

static void cfi_intelext_resume(struct mtd_info *mtd) {
struct map_info *map = mtd->priv;
struct cfi_private *cfi = map->fldrv_priv; int i;
struct flchip *chip;

   for (i=0; i<cfi->numchips; i++) {

       chip = &cfi->chips[i];

       spin_lock(chip->mutex);

       /* Go to known state. Chip may have been power cycled */
       if (chip->state == FL_PM_SUSPENDED) {
           //map_write(map, CMD(0xFF), cfi->chips[i].start);  //@@@@@@@@@@@@@@@ It will report error!!!
           map_write(map, CMD(0xFF), 0);
           chip->oldstate = chip->state = FL_READY;
           wake_up(&chip->wq);
       }

       spin_unlock(chip->mutex);

}
}

When our p2-sample enter the sleep mode, then resume , the resume code will be called.

So could anyone give me to fix the bug ??

Thanks a lot!

zhonglin

I have seen this problem, yes. Specifically, it has shown up when I try to mount a JFFS2 filesystem as the rootfs. When mounting it as a non-rootfs, I haven't seen the problem. I doubt that really makes a difference though, and it's probably related more to timing than anything.

I haven't figure it out yet. My particular case never has the resume code called, so I'm not sure that's the problem.

josh

Well, if you gather and send the full JFFS2 debugging log I may try to help.

Best Regards,
Artem B. Bityuckiy,
St.-Petersburg, Russia.