Hello Andreas, On Mon, Nov 28, 2011 at 02:45:26AM +0000, Andreas O wrote: > I am using a real-time 2.6.33.7 version on ARM and I have noticed some > problems with the NAND. After a power cycle I see messages like > "mtd->read( ) returned ECC Error." It is quite strange. > Example: > > mtd->read(0x1fba0 bytes from 0x4e0460) returned ECC error > mtd->read(0x1d520 bytes from 0x4e2ae0) returned ECC error > > Initially, I make write to the NAND the jff2 filesystem and when I > boot for first time > there are no ECC errors at all. After the first power cycler, some ECC > errors will appear. > If I do a second power cycle, more may be would appear. > If I repeat again the same procedure, writing again the rootfile > system to the NAND, after a power cycle the ECC errors will come from > the same region (0x4exxxx). I have used different types of NAND, and > the problem is the same. > > I think it might have to do with a combination of the RT and jffs2, > Jffs2 garbage collector thread. The GC thread might need priority > change as it takes the lock for performing the scan and may sleep for > longer within the RT world. Therefore writes do not happen that often, > therefore those ECC errors from un-flushed wbuf buffers. > > Have you seen similar behaviour before with RT and Jffs2? How do you restart. Just pressing reset or unplugging is expected to yield warnings by jffs2. See http://www.linux-mtd.infradead.org/faq/jffs2.html#L_messages. Maybe rt makes it more likely for these to appear, but the problem is not specific to rt. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | http://www.pengutronix.de/ | -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
