Pinging ... don't take this personal guys, but are we going to keep the original, dangerously broken implementation in 2.6.24, or will my patch go upstream in time? cheers, Peter --------------- Hi all! As the originator of this patch, please let me comment. Jeff Garzik wrote: >Alan Cox wrote: >> Without the valid bits at least one set of TSScorp drives report 0 in >> word 93 for PATA 40 wire, which we (and the specs) say actually means >> SATA. (The SATA version seems to report 80 wire...) >> >> Signed-off-by: Alan Cox <alan@xxxxxxxxxx> >SATA version on what controller? Have you verified where the bridge is, >if it's not reporting word93==0 ? >Is this for 2.6.24-rc, considering that we are late in -rc? The offending drives are native IDE, but their word93 is broken, in that the validation bits are missing. Hence, if connected to a 40-wire cable, they return word93=0x0000 and make themselves look like a SATA drive. This makes the 'relaxed' cable check do the exact wrong thing - allow fast UDMA modes. The patch discussed here corrects the 'relaxed' cable check, so that it actually does what it's supposed to - look at nothing but the cable type bit in word93 and exactly NOT draw any other conclusions from word93. Due the dangerous behaviour of the unpatched code, I strongly vote for inclusion in 2.6.24. regards, Peter - 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
