Marvell SATA for libata?
Hello all, Jeff Garzik in particular,
we have interest in having the Marvell 88sx5040 SATA controller work under libata.
Did you happen to start development for this controller, as your latest libata status report seems to tell so?
We have the Marvell datasheet available (but under NDA) as well as the source code from SuperMicro and we would like to convert it to the libata framework.
Regards,
Leon.
Quoted 11 lines · Leon Woestenberg
Hello all, Jeff Garzik in particular,
we have interest in having the Marvell 88sx5040 SATA controller work under libata.
Did you happen to start development for this controller, as your latest libata status report seems to tell so?
We have the Marvell datasheet available (but under NDA) as well as the source code from SuperMicro and we would like to convert it to the libata framework.
In fact, I am working on it right now.
I also have the docs and the driver (albeit officially :)).
A non-queueing driver for Marvell will be fairly straightforward. Adding queueing will be much more of a chore, though.
Jeff
Hello Jeff,
Quoted 9 lines · Jeff Garzik
Quoted 8 lines · Leon Woestenberg
Hello all, Jeff Garzik in particular,
we have interest in having the Marvell 88sx5040 SATA controller work under libata.
Did you happen to start development for this controller, as your latest libata status report seems to tell so?
Quoted 7 lines
Quoted 3 lines
We have the Marvell datasheet available (but under NDA) as well as the source code from SuperMicro and we would like to convert it to the libata framework.
In fact, I am working on it right now.
Excellent. Applying as test-bed candidate. :-)
Quoted 2 lines
I also have the docs and the driver (albeit officially :)).
I happen to find it at this public customer support FTP server: ftp://ftp.supermicro.com/driver/SATA/DAC_SATA-MV8/
which is officially for the MV8 PCI card, which has the forementioned chip on its PCB.
Quoted 2 lines
A non-queueing driver for Marvell will be fairly straightforward. Adding queueing will be much more of a chore, though.
That makes me wonder, what is the typical performance gain when using any of the command queueing methods?
Regards,
Leon.