Hi Jeff, On 08/24/2015 12:41 PM, Jeff Layton wrote: > v2: > - fix the fault injection code to handle the prototype change to > unhash_lock_stateid > > Slightly updated set, to fix a problem that Wu found when enabling the > nfsd fault injection code. Original cover-letter follows: > > -------------------------[snip]------------------- > > I hadn't heard from Andrew in a while so I'm not 100% clear on whether > his patches had fixed the race that he had spotted in the NFSv4 > unhashing code. > > Still, once he had done the analysis, the problem was clear. I think > this patchset should fix the problem in a relatively clean way. I've > run some tests against this and it seems to not have broken anything, > but I haven't been able to reproduce the actual race so I can't > verify that this fixes it. > > Andrew, Anna, if you are able to do so could you test these patches > and let us know whether this fixes those races? I just did a quick test, and it looks like my problem has been fixed. I'll do more in-depth testing tomorrow, just to make sure. Thanks for working on this! Anna > > Jeff Layton (2): > nfsd: ensure that the ol stateid hash reference is only put once > nfsd: ensure that delegation stateid hash references are only put once > > fs/nfsd/nfs4state.c | 84 +++++++++++++++++++++++++++++++++-------------------- > 1 file changed, 53 insertions(+), 31 deletions(-) > -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
