> -----Original Message----- > From: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> > Sent: Saturday, March 05, 2022 5:43 AM > To: Keller, Jacob E <jacob.e.keller@xxxxxxxxx> > Cc: stable@xxxxxxxxxxxxxxx; Jankowski, Konrad0 > <konrad0.jankowski@xxxxxxxxx>; Nguyen, Anthony L > <anthony.l.nguyen@xxxxxxxxx> > Subject: Re: [PATCH 1/2] ice: Fix race conditions between virtchnl handling and VF > ndo ops > > On Mon, Feb 28, 2022 at 08:57:50PM +0000, Keller, Jacob E wrote: > > > > > > > -----Original Message----- > > > From: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> > > > Sent: Monday, February 28, 2022 3:24 AM > > > To: Keller, Jacob E <jacob.e.keller@xxxxxxxxx> > > > Cc: stable@xxxxxxxxxxxxxxx; Brett Creeley <brett.creeley@xxxxxxxxx>; > Jankowski, > > > Konrad0 <konrad0.jankowski@xxxxxxxxx>; Nguyen, Anthony L > > > <anthony.l.nguyen@xxxxxxxxx> > > > Subject: Re: [PATCH 1/2] ice: Fix race conditions between virtchnl handling > and VF > > > ndo ops > > > > > > On Mon, Feb 28, 2022 at 12:10:20PM +0100, Greg KH wrote: > > > > On Fri, Feb 25, 2022 at 12:21:00PM -0800, Jacob Keller wrote: > > > > > From: Brett Creeley <brett.creeley@xxxxxxxxx> > > > > > > > > > > commit e6ba5273d4ede03d075d7a116b8edad1f6115f4d upstream. > > > > > > > > > > [I had to fix the cherry-pick manually as the patch added a line around > > > > > some context that was missing.] > > > > > > > > What stable tree(s) is this for?$ > > > > > > Looks like it applied only to 5.15.y. Can you also provide backports > > > for the other older kernels that these are needed for? > > > > > > > Hi Greg! > > > > I sent a series that should apply from v5.8 through 5.12 (excepting one patch > that was already on 5.10 but not the other trees, for some reason). I also sent a > series for 5.13 and one for 5.14. > > > > The code prior to 5.8 still has this bug, going back to the first release with ice, > (4.20), but it is different enough that I had trouble determining what the correct > patch is. Especially for the first patch which is necessary for the later fixes, since it > introduces the lock we need. I'm going to spend a bit more time later this evening > to see if I can sort it out. If not, we may have to live without the fixes on those > trees. > > I've applied the 2 missing changes to the 5.10.y tree. The other ones > are all end-of-life, sorry. > > thanks, > > greg k-h Thanks Greg! -Jake
