[PC-BSD Testing] Removable media

Rod Clark sc at sdf.lonestar.org
Tue Apr 21 22:55:23 PDT 2009


On Wed, 1 Oct 2008, Arthur Koziol wrote:
> I've been sort of loosely watching this thread. Is there any easy
> test this to see if this is isolated perhaps to certain chipsets on
> certain mothers boards (AMD chipset v. Intel v. NForce v. VIA, etc.)
> that the rest of us can test and confirm? Typically, USB flash drives
> UNDER 2GB will come out of the factory formatted to 2GB. Anything
> OVER 2GB has ot be FAT32 or some other size due to FATs 2GB ceiling.
> I've tried my 4GB and 8GB Sandisk drives with no loss or data error
> but they are formatted FAT32. I have plenty of time to run some tests
> with USB drivers ranging from 128 FAT or FAT32 to 8GB FAT32 to play with.

Arthur,

One thing that has been causing some of the crashes on my
6-STABLE machine is, strangely enough, the SD card write-protect
switch. This wasn't apparent until the other day because I had
never flipped the position of the write-protect switch until
upgrading the camera software recently.

The CHDK software uses the SD card's write protect switch to
flag whether to start itself up or (with the write-protect
switch off) to use the camera's own standard firmware. (CHDK
writes photos to the card even though the switch is on.)

But the write protect switch being on makes a fatal difference.
The SD card and reader operate normally with the switch off, but
setting the switch on makes FreeBSD 6 unstable on insertion and
generally crashes it completely during use or on removal. I
haven't tried it yet on FreeBSD 7, but will do that tonight.

Could you or others confirm this, to see whether this is
isolated or widespread? The card is a 2GB Sandisk Ultra II SD. I
can try it on a couple of other old 16MB and 32MB cards, but I
don't have an SDHC (4GB or larger) card here to test. Crashes
have happened with other card readers too, but I have only one
of them here now.

The other USB thumb drive instability problems that I had
earlier were fixed in recent versions of HAL, so the big
remaining problem here was with the SD card readers.

-rsc


More information about the Testing mailing list