![](https://seccdn.libravatar.org/avatar/3035b38ff33cf86f480bb169b8500b80.jpg?s=120&d=mm&r=g)
https://bugzilla.novell.com/show_bug.cgi?id=881727
https://bugzilla.novell.com/show_bug.cgi?id=881727#c7
--- Comment #7 from Vlastimil Babka
(In reply to comment #4)
Could you please attach output of dmesg from the system?
The system got naturally rebooted in the meantime but here it is in multiple parts due to limitations in bugzilla.
Rebooting is OK, I was interested in the memory initialization messages. And seems they indeed explain a lot...
[ 0.000000] SRAT: Node 0 PXM 0 [mem 0x00000000-0x0009ffff] [ 0.000000] SRAT: Node 0 PXM 0 [mem 0x000e8000-0xdfffffff] [ 0.000000] SRAT: Node 0 PXM 0 [mem 0x100000000-0x19fffffff] [ 0.000000] SRAT: Node 1 PXM 1 [mem 0x1a0000000-0x31fffffff] [ 0.000000] NUMA: Node 0 [mem 0x00000000-0x0009ffff] + [mem 0x000e8000-0xdfffffff] -> [mem 0x00000000-0xdfffffff] [ 0.000000] NUMA: Node 0 [mem 0x00000000-0xdfffffff] + [mem 0x100000000-0x19fffffff] -> [mem 0x00000000-0x19fffffff] [ 0.000000] Initmem setup node 0 [mem 0x00000000-0x19fffffff] [ 0.000000] NODE_DATA [mem 0x3637b000-0x3637cfff] [ 0.000000] Initmem setup node 1 [mem 0x1a0000000-0x31fffffff] [ 0.000000] Cannot find 8192 bytes in node 1
So, in fact this is a NUMA machine with two machines. But for some reason, node 1 could not get its NODE_DATA allocated so it acts as a single node machine, as your zoneinfo has shown.
[ 0.000000] 11918MB HIGHMEM available. [ 0.000000] 881MB LOWMEM available. [ 0.000000] max_low_pfn = 371fe, highstart_pfn = 371fe [ 0.000000] Low memory ends at vaddr f71fe000 [ 0.000000] High memory starts at vaddr f71fe000 [ 0.000000] mapped low ram: 0 - 371fe000 [ 0.000000] low ram: 0 - 371fe000 [ 0.000000] BRK [0x00cd2000, 0x00cd2fff] PGTABLE [ 0.000000] Node: 0, start_pfn: 1, end_pfn: 9c [ 0.000000] Setting physnode_map array to node 0 for pfns: [ 0.000000] 1 [ 0.000000] Node: 0, start_pfn: 100, end_pfn: defa5 [ 0.000000] Setting physnode_map array to node 0 for pfns: [ 0.000000] 100 4100 8100 c100 10100 14100 18100 1c100 20100 24100 28100 2c100 30100 34100 38100 3c100 40100 44100 48100 4c100 50100 54100 58100 5c100 60100 64100 68100 6c100 70100 74100 78100 7c100 80100 84100 88100 8c100 90100 94100 98100 9c100 a0100 a4100 a8100 ac100 b0100 b4100 b8100 bc100 c0100 c4100 c8100 cc100 d0100 d4100 d8100 dc100 [ 0.000000] Node: 0, start_pfn: 100000, end_pfn: 1a0000 [ 0.000000] Setting physnode_map array to node 0 for pfns: [ 0.000000] 100000 104000 108000 10c000 110000 114000 118000 11c000 120000 124000 128000 12c000 130000 134000 138000 13c000 140000 144000 148000 14c000 150000 154000 158000 15c000 160000 164000 168000 16c000 170000 174000 178000 17c000 180000 184000 188000 18c000 190000 194000 198000 19c000 [ 0.000000] Node: 1, start_pfn: 1a0000, end_pfn: 320000 [ 0.000000] Setting physnode_map array to node 1 for pfns: [ 0.000000] 1a0000 1a4000 1a8000 1ac000 1b0000 1b4000 1b8000 1bc000 1c0000 1c4000 1c8000 1cc000 1d0000 1d4000 1d8000 1dc000 1e0000 1e4000 1e8000 1ec000 1f0000 1f4000 1f8000 1fc000 200000 204000 208000 20c000 210000 214000 218000 21c000 220000 224000 228000 22c000 230000 234000 238000 23c000 240000 244000 248000 24c000 250000 254000 258000 25c000 260000 264000 268000 26c000 270000 274000 278000 27c000 280000 284000 288000 28c000 290000 294000 298000 29c000 2a0000 2a4000 2a8000 2ac000 2b0000 2b4000 2b8000 2bc000 2c0000 2c4000 2c8000 2cc000 2d0000 2d4000 2d8000 2dc000 2e0000 2e4000 2e8000 2ec000 2f0000 2f4000 2f8000 2fc000 300000 304000 308000 30c000 310000 314000 318000 31c000
Despite NODE_DATA not being allocated for node 1, physnode_map was filled for zone 1. As a side note, it is questionable why anyone would want to run a 32bit kernel on a NUMA machine with 12GB RAM. 64bit kernel would make much more sense. I'm guessing not a lot of people run 32bit on systems like this, which is why this hasn't been reported sooner. Besides that, there are clearly two problems here: 1) physnode_map should not have been filled for node 1 - it's the direct cause of this bug. 2) why wasn't NODE_DATA for node 1 allocated? Losing access to 6GB of memory because 8192 bytes could not be allocated is not exactly the best tradeoff I can imagine. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.