于 2013/11/18 18:00, Fan Du 写道: > > > On 2013年11月18日 15:00, Li Zefan wrote: >> On 2013/11/18 14:48, Fan Du wrote: >>> >>> >>> On 2013年11月18日 14:42, Zhao Hongjiang wrote: >>>> On 2013/11/18 14:36, Fan Du wrote: >>>>> Hi, HongJiang >>>>> >>>>> On 2013年11月18日 14:25, Zhao Hongjiang wrote: >>>>>> Hi, Fan or Andrew, >>>>>> >>>>>> 74e3d1e17b2e11d175970b85acd44f5927000ba2 >>>>>> include/linux/fs.h: disable preempt when acquire i_size_seqcount write lock >>>>>> >>>>>> >>>>>> This looks applicable to stable-3.4, that fix a lockup while two rt tasks >>>>>> bind to one CPU core. It was built successful for me. What do you think? >>>>> >>>>> I don't see any stable tree tagged as "stable-3.4", do you mean the "longterm 3.4.69"? >>>> >>>> Yes! I mean the longterm 3.4.69. >>> >>> It should be integrated into longterm 3.4.69, but usually the author of this fix is expected >>> to request for stable tree only, I'm not familiar with the longterm tree maintenance procedure. >>> >>> I cannot help you with this, maybe Andrew will clarify your request. >>> >> >> Yes you can. >> >> A longterm stable tree is a stable tree. As you're the author of this >> patch, if you agree it fits for 3.4, Greg will do the remaining work. >> > > I'm happy to do it, however I don't know in the first place how I should write > the request, and to whom I should send this. Is there any template I can refer to? Yes. Documentation/stable_kernel_rules.txt > Informal/improper such request seems always upset maintainers。。。 > > IMO, the subsystem maintainer of this patch will make his/her judgment whether > this fix should be queued into longterm tree according to the severity of this fix. > Apparently it's not the call of the patch author. > The patch author can add a stable tag in the changelog if he feels the patch should be backported to stable kernels. If he didn't do that, the maintainer may tag it for stable. After the patch hits mainline, it will be queued into stable queue automatically. If neither of them did that, and then someone else feels the fix is a stable thing, this person can ask Greg to include it for stable, as long as no one objects to it. What you need to do here, is ack or nack the stable request from Hongjiang. -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html