Memory leak in 2.3.43pre2 (select.c changes)

[Simon Kirby]

There's a memory leak in 2.3.43pre2 that showed visibly about 10 seconds
after starting X (I could see sawmill draw the window borders and
netscape felt slow), followed shortly by a grinding halt.  I can see the
problem quite easily by typing "free" over and over in an rxvt.  SMP dual
processor box, 128 MB ECC SDRAM.

Reverting the changes to select.c in 2.3.43pre2 seems to have fixed it.

Reviewing the serial console capture log shows this:

VM: killing process panel
VM: killing process rxvt
VM: killing process rxvt
VM: killing process sawmill
VM: killing process init
VM: killing process rpc.mountd
VM: killing process rpc.nfsd
VM: killing process ps
VM: killing process tcsh
VM: killing process irc
VM: killing process rxvt
VM: killing process tcsh
VM: killing process rxvt
VM: killing process tcsh
VM: killing process rxvt
VM: killing process tcsh
VM: killing process rxvt
VM: killing process tcsh
VM: killing process xinit
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process tcsh
VM: killing process apache
VM: killing process tcsh
VM: killing process apache
VM: killing process tcsh
VM: killing process apache
VM: killing process tcsh
VM: killing process tcsh
VM: killing process init
NMI Watchdog detected LOCKUP on CPU0, registers:
CPU:    0
EIP:    0010:[<c0119104>]
EFLAGS: 00000082
eax: c126e098   ebx: 00000002   ecx: c126fe60   edx: 00000021
esi: c4298000   edi: 00000011   ebp: 00000286   esp: c126fe44
ds: 0018   es: 0018   ss: 0018
Process init (pid: 1, stackpage=c126f000)
Stack: c0116aee 00000011 c126fe60 c126e000 c4298000 c027e000 c126e000 00000011
       00000000 00000002 00000235 00000030 00000001 00000010 0000002a c7f6ebfc
       00000001 c7f6ebfc 00000000 00000000 c4298000 c4298000 c027e000 c4298000
Call Trace: [<c0116aee>] [<c010c0d6>] [<c0120018>] [<c012218b>] [<c0122672>] [<c0115912>] [<c0121a6e>]
       [<c010b166>] [<c010c05d>]
Code: 55 89 e5 83 ec 2c 57 56 53 89 45 fc 89 55 f8 85 c0 0f 84 3b
console shuts up ...

...Not a very graceful OOM, but I suppose if it's an in-kernel leak, this
makes sense.

Simon-

[  Stormix Technologies Inc.  ][  NetNation Communcations Inc.  ]
[       [e-mail]       ][       [e-mail]        ]
[ Opinions expressed are not necessarily those of my employers. ]