for CVS sensors modules, which must work with kernels 2.4.9 - 2.4.xx, we use hex ids if they aren't defined in 2.4.9. For a 2.6 kernel patch, the preference is to include a patch to pci_ids.h if necessary and use only names in the driver. mds Gaston, Jason D wrote: > Mark, > > I see that you are using the DID's directly, for ICH5,6,7, rather then > using the pci_ids.h defines. Is this your preference or would you > rather use the pci_ids.h names? > > I am wondering if I should also just use the actual DID's for the 2.6 > patch? > > Thanks, > > Jason Gaston > > > > -----Original Message----- > From: Mark Studebaker [mailto:mds4 at verizon.net] > Sent: Wednesday, December 08, 2004 1:01 PM > To: Gaston, Jason D > Cc: LM Sensors > Subject: Re: RFC: complete rewrite of i2c-i801 for 2.6.x > > since I'm in the middle of 801 driver updates already, I'll go ahead and > add support to > the driver and sensors-detect in CVS. I'll leave the 2.6 patch to you. > let us know when you have test results. > thanks > mds > > Gaston, Jason D wrote: > >>Mark, >> >>I don't think that there are any changes to the ICH7 SMBus Controller, >>other then DID, which is 27DAh. The Datasheet will not be available >>until the product launches. I will however, patch the driver with the >>new DID, test it out and then submit the patch here and to the LKML. > > My > >>plan is to do this within the next six weeks or so. Unless you can or >>want to add the DID now? >> >>Let me know if you would still like me to try something out on ICH5, >>ICH6 or ESB6300. >> >>Thanks, >> >>Jason Gaston >> >> >> >>-----Original Message----- >>From: Mark Studebaker [mailto:mds4 at verizon.net] >>Sent: Wednesday, December 08, 2004 9:37 AM >>To: Gaston, Jason D >>Cc: LM Sensors >>Subject: Re: RFC: complete rewrite of i2c-i801 for 2.6.x >> >>thanks for offer. i2c block read support for ich5 and later is now in >>CVS. >>Best way to test bus drivers is with 'i2cdump'. >>In particular to test i2c block reads, best way is with the "i" mode > > of > >>i2cdump. >>If it works, regular byte ("b") mode and "i" mode should have > > identical > >>results >>when reading from an eeprom. >> >>For the ICH7, first we've heard of it, but not surprised it would be > > the > >>chip >>after ICH6 :) >>I see there's nothing on developer.intel.com about it. Will there be > > any > >>changes in the SMBus controller? What is the device ID? Datasheet >>availability? >>Whether you submit 2.4 or 2.6 patches doesn't really matter. >>But with the device ID, and assurances that the SMBus part of the chip >>didn't change, >>it's real easy. >> >>mds >> >> >>Gaston, Jason D wrote: >> >> >>>Mark, >>> >>>I would be willing to do some testing of the i2c-i801 driver on ICH5, >>>ICH6, ESB6300 and ICH7. I have access to boards/systems, but I am not >>>proficient in i2c testing, other then running "sensors" and seeing the >>>data. If you know of some instructions for the testing you are >> >>looking >> >> >>>for, I am willing to give it a go. >>> >>>In addition, I will soon want to get support for ICH7 added to the >>>i2c-i801 driver. Last time, when I wanted to get support for ICH6 >>>added, I posted a patch, adding the DID's, to the LKML and Greg KH. >>>This made it into the 2.6 kernel. Perhaps I should get support added >>>here first and then the base kernel? >>> >>>Let me know how I can help, >>> >>>Jason Gaston >>>Jason.d.gaston at intel.com >>> >>> >>> >>> >>>-----Original Message----- >>>From: Mark Studebaker [mailto:mds4 at verizon.net] >>>Sent: Tuesday, December 07, 2004 7:02 PM >>>To: Sensors >>>Subject: Re: RFC: complete rewrite of i2c-i801 for 2.6.x >>> >>> >>> >>>Mark M. Hoffman wrote: >>> >>> >>> >>>>Hi Mark: >>>> >>>>(CCs trimmed) >>>> >>>>* Mark D. Studebaker <mds4 at verizon.net> [2004-12-07 11:20:54 -0500]: >>>> >>>> >>>> >>>> >>>>>My reading of the DB datasheet (in 2002 and now) is that it doesn't >>> >>>support a >>> >>> >>> >>>>>standard i2c block read. The driver functionality bitmask reflects >>> >>>that. >>> >>> >>> >>>>>I fixed the error message (which would only print out if a driver >>> >>>ignored >>> >>> >>> >>>>>the functionality mask). >>>> >>>> >>>>Interesting. It looks like that datasheet only shows this: >>>> >>>> S Addr Wr [A] Comm [A] D0 [A] D1 [A] S (read bytes...) >>>> >>>>I agree that's not like any transaction we support. >>> >>> >>>correct. don't know what they were thinking. >>> >>> >>> >>> >>>>OTOH, the ICH5 [1] datasheet shows this: >>>> >>>> S Addr Wr [A] D1 [A] S (read bytes...) >>>> >>>>That's exactly the transfer protocol that eeproms need, right? >>> >>>(That's why I >>> >>> >>> >>>>mentioned to Khali by IRC last weekend that I thought it should be >>> >>>possible.) >>> >>> >>>I think it is what eeproms need (you omitted 'Addr' after the second >>>'S'). >>>It matches what I wrote in doc/smbus-protocol in i2c >>>(that is, the i2c block functions are defined to take a 'command' >> >>byte, >> >> >>>matching >>>what eeproms need for 'random' reads and writes). >>> >>>Does anybody have a ICH5/6? If so I can add support to the driver for >>>testing. >>> >>>mds >>> >>> >>> >>> >>> >>>>It would be really bizarre if these two chipsets behaved differently >>> >>>in this >>> >>> >>> >>>>respect - I wonder if the datasheets are correct? >>>> >>>>[1] Well I just scanned all the other datasheets... only ICH5, ICH6, >>> >>>and >>> >>> >>> >>>>6300ESB appear to support the standard I2C block read. Sorry Khali. >>> >>>:) >>> >>> >>> >>>>Regards, >>>> >>> >>> >>> >> > >
