Occasional crashes, Bug or dying disk?

[Jonathan T. Looney]

On 2/2/16, 5:53 PM, "owner-freebsd-current at freebsd.org on behalf of
Christian Walther" <owner-freebsd-current at freebsd.org on behalf of
cptsalek at gmail.com> wrote:

>Hello list,
>
>since updating the current installation on my trusty Thinkpad T43 I
>keep getting occasional crashes like the following:
>
>#1  0xc0c23a03 in kern_reboot (howto=260)
>    at /usr/src/sys/kern/kern_shutdown.c:364
>#2  0xc0c23f3d in vpanic (fmt=<value optimized out>, ap=<value optimized
>out>)
>    at /usr/src/sys/kern/kern_shutdown.c:757
>#3  0xc0c23f5b in panic (fmt=0xc1590fad "ffs_blkfree_cg: freeing free
>frag")
>    at /usr/src/sys/kern/kern_shutdown.c:688
>[...]
>I wonder if this might be a bug in FFS or a related subsystem, or my
>hard disk dying. smartctl lists several READ DMA und WRITE DMA related
>errors.

I (and several others) spent time trying to track down "the bug" causing
panics like this on a particular set of hardware. I finally wrote a
program that wrote and read byte patterns from the disk (using O_DIRECT).
I found that what the program read back didn't match what it wrote. In
circumstances like that, you can't expect any filesystem to work reliably.

Given the long pedigree of the UFS/FFS code, I think it is much more
likely to be a hardware problem than a problem with the UFS/FFS code.

Jonathan