2.6.x SMP UltraSPARC instability
PATCH 1/1 2.6.x SMP UltraSPARC instability
Can folks seeing problems on SMP sparc64 boxes with 2.6.x give this patch a try?
arch/sparc64/kernel/irq.c renamed from arch/sparc64/kernel/irq.c.~1~ +3 -3
@@ -756,7 +756,7 @@ clear_softint(clr_mask); }#else -int should_forward =1 ; +int should_forward =0 ;clear_softint (1 << irq);#endif
@@ -1007,10 +1007,10 @@ }upa_writel (tid |IMAP_VALID , imap); -while (!cpu_online (goal_cpu)) { +do {if (++goal_cpu >=NR_CPUS ) goal_cpu =0 ; - } + }while (!cpu_online (goal_cpu));return goal_cpu; }
Quoted 2 lines · David S. Miller
Can folks seeing problems on SMP sparc64 boxes with 2.6.x give this patch a try?
Both cases compiled with gcc-3_3-branch dated 20040623 (post 3.3.4)
2.6.11 without patch:
crashes after about 2 minutes of uptime
2.6.12-rc3 with patch:
so far it's been up for about 10 minutes with no problems so far... I'm going to leave it doing some high load tasks while I head out to go see a movie... hopefully it'll still be up when I get back.
--Jeremy
system is an ultra2, and the following info is as reported under 2.4.29smp
/proc/cpuinfo
cpu : TI UltraSparc II (BlackBird)
fpu : UltraSparc II integrated FPU
promlib : Version 3 Revision 25
prom : 3.25.0
type : sun4u
ncpus probed : 2
ncpus active : 2
Cpu0Bogo : 591.46
Cpu0ClkTck : 0000000011a4959b
Cpu1Bogo : 591.46
Cpu1ClkTck : 0000000011a4959b
MMU Type : Spitfire
State:
CPU0: online CPU1: online
I've heard scsi is sometimes a factor with some of the 2.6 issues, so including this for good measure:
/proc/scsi/esp/0
Sparc ESP Host Adapter:
PROM node f006635c
PROM name SUNW,fas
ESP Model Happy Meal FAS
DMA Revision Rev HME/FAS
Live Targets [ 0 1 6 ]Target # config3 Sync Capabilities Disconnect
Wide
0 000000c3 [0f,04] yes
yes
1 000000c3 [0f,04] yes
yes
6 00000081 [0f,04] yes
no
/proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: FUJITSU Model: MAB3045S SUN4.2G Rev: 1705
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: FUJITSU Model: MAB3045S SUN4.2G Rev: 1705
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 06 Lun: 00
Vendor: TOSHIBA Model: XM5701TASUN12XCD Rev: 0997
Type: CD-ROM ANSI SCSI revision: 02
Quoted 9 lines · Jeremy Huddleston
Quoted 2 lines · David S. Miller
Can folks seeing problems on SMP sparc64 boxes with 2.6.x give this patch a try?
Both cases compiled with gcc-3_3-branch dated 20040623 (post 3.3.4)
2.6.11 without patch:
crashes after about 2 minutes of uptime
Quoted 8 lines
2.6.12-rc3 with patch:
so far it's been up for about 10 minutes with no problems so far... I'm going to leave it doing some high load tasks while I head out to go see a movie... hopefully it'll still be up when I get back.--Jeremy
system is an ultra2
Looking good still after almost 7 hours... also looks like I'm not affected by the SCSI bugs others have hit as I've been doing a ton of disk i/o with no problems... =)
$ uname -a Linux yuffie 2.6.12-rc3-gentoo #1 SMP Tue May 3 20:21:05 PDT 2005 sparc64 sun4u TI UltraSparc II (BlackBird) GNU/Linux
$ uptime 03:51:23 up 6:43, 2 users, load average: 3.00, 3.00, 2.75