On 19/12/16 16:25, Andrew Bradford wrote:
I thought I'd poke about with the eMMC in the Galaxy S4 which I have here to see if it supported a sufficient number of AUs to be worth switching to f2fs (I believe you need 7 simultaneous AUs minimum?).
Are you doing these tests on Linux or on Android?
Under Android (within a Debian chroot) - the chroot is on an sdcard, and the phone is booted up using the 'twrp' recovery kernel+userspace combo - hence being able to monkey with the cache partition whilst booted up.
This looks like a Toshiba eMMC. I have a 4 GB Toshiba eMMC which shows similar manfid, name, and oemid values and manfid 0x11 is listed as their Toshiba manfid in the datasheets I have.
OK, thanks I found some references to Toshiba using an 0x11 manfid elsewhere.
# ... not really sure how to interpret that. Am I wrong in attempting # to just operate on the partition, with that offset?
I don't think you're wrong to use that offset, I think the eMMC is hiding its real performance from you quite effectively. I'm not exactly sure how these Toshiba controllers work internally, but they seem pretty effective at evading investigation from flashbench.
[...]
I'd just give f2fs a try, see how the performance is. It'll probably be decent.
OK, thanks. That's the conclusion I'd come to, I had some doubts about the validity of my results, but good to see that a similar eMMC seems to do the same thing in another system (with presumably a different kernel - I've not looked at this Android Samsung-derived kernel's source, but I'm assuming it's not pretty).
Cheers,
Tim.