Re: linux-next regression: SNP Guest boot hangs with certain cpu/mem config combination

From: Kirill A. Shutemov
Date: Tue Mar 25 2025 - 09:47:06 EST


On Tue, Mar 25, 2025 at 02:40:00PM +0530, Aithal, Srikanth wrote:
> Hello,
>
>
> Starting linux-next build next-20250312, including recent build 20250324, we
> are seeing an issue where the SNP guest boot hangs at the "boot smp config"
> step:
>
>
>  [ 2.294722] smp: Bringing up secondary CPUs ...
> [    2.295211] smpboot: Parallel CPU startup disabled by the platform
> [    2.309687] smpboot: x86: Booting SMP configuration:
> [    2.310214] .... node  #0, CPUs:          #1   #2   #3   #4 #5   #6  
> #7   #8   #9  #10  #11  #12  #13  #14  #15  #16  #17 #18  #19  #20  #21 
> #22  #23  #24  #25  #26  #27  #28  #29  #30 #31  #32  #33  #34  #35  #36 
> #37  #38  #39  #40  #41  #42  #43 #44  #45  #46  #47  #48  #49  #50  #51 
> #52  #53  #54  #55  #56 #57  #58  #59  #60  #61  #62  #63  #64  #65  #66 
> #67  #68  #69 #70  #71  #72  #73  #74  #75  #76  #77  #78  #79  #80  #81 
> #82 #83  #84  #85  #86  #87  #88  #89  #90  #91  #92  #93  #94  #95 #96 
> #97  #98  #99 #100 #101 #102 #103 #104 #105 #106 #107 #108 #109 #110 #111
> #112 #113 #114 #115 #116 #117 #118 #119 #120 #121 #122 #123 #124 #125 #126
> #127 #128 #129 #130 #131 #132 #133 #134 #135 #136 #137 #138 #139 #140 #141
> #142 #143 #144 #145 #146 #147 #148 #149 #150 #151 #152 #153 #154 #155 #156
> #157 #158 #159 #160 #161 #162 #163 #164 #165 #166 #167 #168 #169 #170 #171
> #172 #173 #174 #175 #176 #177 #178 #179 #180 #181 #182 #183 #184 #185 #186
> #187 #188 #189 #190 #191 #192 #193 #194 #195 #196 #197 #198
> --> The guest hangs forever at this point.
>
>
> I have observed that certain vCPU and memory combinations work, while others
> do not. The VM configuration I am using does not have any NUMA nodes.
>
> vcpus             Mem        SNP guest boot
> <=240            19456M    Boots fine
> >=241,<255   19456M    Hangs
> 1-255              2048M    Boots fine
> 1-255              4096M    Boots fine
> >71                 8192M    Hangs
> >41                 6144M    Hangs
>
> When I bisected this issue, it pointed to the following commit :
>
>
> *commit 800f1059c99e2b39899bdc67a7593a7bea6375d8*
> Author: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
> Date:   Mon Mar 10 10:28:55 2025 +0200
>
>     mm/page_alloc: fix memory accept before watermarks gets initialized

Hm. It is puzzling for me. I don't see how this commit can cause the hang.

Could you track down where hang happens?

--
Kiryl Shutsemau / Kirill A. Shutemov