On Mon, Sep 19, 2011 at 7:35 PM, thibaut bethune <thibaut.bethune@xxxxxxxxx> wrote: > Hello again > > I've formatted my SSD using ext4 and enabled TRIM using the "discard" option. > Then i've run that test > https://sites.google.com/site/lightrush/random-1/checkiftrimonext4isenabledandworking > and it failed : no TRIM with that configuration. > > Since it's theorically possible to get TRIM without AHCI, only using > 48-bit LBA addressing, do you know if a can find a driver that would > enable TRIM ? > > thanks in advance > > Thibaut > France > > 2011/9/14 thibaut bethune <thibaut.bethune@xxxxxxxxx>: >> Hello, >> >> I'd like to know if my system is ATA TRIM compatible : >> >> 1°) I've just bought a SATA SSD (Crucial M4) which handles TRIM >> 2°) I'm running Linux 2.6.33 whch also handles TRIM (ext4 + "discard" option) >> 3°) My motherboard (ASRock Conroe 1333 DVI/H R2.02, featuring the ICH7 >> southbridge) doesn't handle AHCI. Yet 48-bit LBA addressing is >> supported. >> >> I did some research and I've read that 48-bit LBA addressing should be >> enough to get TRIM IF the driver is compatible. >> >> Is it the case ? >> >> Thanks in advance >> >> Thibaut >> 35' >> France I don't have any answers for you, but I do want to argue with the apparent assumption that mount -t ext4 -o discard is a good thing. To date I have not read of any devices that truly benefit from that. On the other hand, recent versions of e2fsprogs have a userspace tool, fstrim. It works by invoking a kernel ioctl that has been around for about a year. Maybe more. You can setup cron to invoke fstrim daily (or whatever). It has the advantage of keeping your ssd fairly well "trimmed-up" but not the negative aspect of interspersing trim commands in the middle of your normal data packets. (Typically this seems to cause a cache flush on the ssd, so it really drops performance.) FYI: I wrote most of: http://en.opensuse.org/SDB:SSD_discard_%28trim%29_support If you find anything relevant out related to drivers that do or don't support trim, it be great if you would update the wiki article. Or post it here and I will try to do so. Greg -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
