On 2017/5/8 下午12:48, tang.junhui@xxxxxxxxxx wrote: > Hello Coly, > > > ＞ Hi Junhui, > > ＞ > > ＞ I think this is a design on purpose. Current multiple partitions bcache > > ＞ design is per-device, not per major number, and each bcache devices as > > ＞ 16 partitions at max. > > ＞ > > ＞ This is why when you see a bcache device is initialized, the first minor > > ＞ of the device will *BCACHE_MINORS. > > ＞ > > ＞ Your patch won't make all minor numbers of the partitions of a bcache > > ＞ device being continuous, IMHO this is not correct. > > ＞ > > ＞ Thanks. > > ＞ > > ＞ Coly > > > We get minor number by ida_simple_get(), but the minor what we are going to > > relase is minor*BCACHE_MINORS, so we release a wrong minor > minor*BCACHE_MINORS, > > and the right minor number leakage occures. I mean if there are partitions on bcache devices, like /dev/bcache0p1, /dev/bcache0p2 ... /dev/bcache0p16 /dev/bcache1p1, /dev/bcache1p2 ... /dev/bcache0p16 bcache0 is not the last level to allocate minor number, bcache0p1 is. What we should fix might be: /dev/bcache0 has minor 0 /dev/bcache1 has minor 16 /dev/bcache2 has minor 32 Maybe I didn't understand you correctly. Can your patch make sure all partitions of a bcache device have adjacent minor numbers ? Correct me if I am wrong. -- Coly Li -- To unsubscribe from this list: send the line "unsubscribe linux-bcache" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
