Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: ld at sdmmc

On Tue, 16 Sep 2008, Rafal Boni wrote:

On Mon, Sep 15, 2008 at 05:31:12AM -0700, Paul Goyette wrote:
Wonderful news.  And I'm sure Jeremy Reed will be happy, too, especially
if it works on his amd64 machine.

I'll compare Kiyohara-san's version against my own and hopefully learn
why my attempt to do this didn't work!

Note that (IIRC) nonaka-san's original patches have some rather poor locking
and can be fairly easily pushed into bad states -- the locks I think were at
too high of a level, forcing recursive lock acquisition and a mess releasing
those locks when failures happen (this is all from memory, it's been a while
since I looked at the SDMMC code).

Yeah, that's what I suspected. My attempts to replace the scsipi layer with ld ran into some "interesting" locking issues.

I haven't paid much attention to it lately, having been occupied with the day job and the dBCool sensor driver. If folks are still interested in getting sdmmc to work, I'll try to get back to it after my vacation.

Maybe this gets better emulating an ld(4) vs. a SCSI device, but I recall
a lot of the problems were in the backend not the disk-emulation layer(s).

I spent a little bit of time trying to untangle this before I realized
that my iPAQ didn't use the PXA built-in SD/MMC controller, but instead
needed a totally new driver for the controller in the HTC ASIC(s); I
punted at that point.


 Time is an illusion; lunchtime, doubly so.     |/\/\|           Rafal Boni
                  -- Ford Prefect               |\/\/|


|   Paul Goyette   | PGP DSS Key fingerprint: |  E-mail addresses:   |
| Customer Service | FA29 0E3B 35AF E8AE 6651 |   |
| Network Engineer | 0786 F758 55DE 53BA 7731 | |

Home | Main Index | Thread Index | Old Index