> None of the other setups produce the "..could net get a free page.."
> message. This looks like a subtle race condition with this particular scsi
> driver. I will try some other scenarios.
That may be part of it, but it certainly isn't all of it.
Since 2.0.35 at least, my little 8Mb-RAM 486 firewall, esperi, has been
bombing out every ten days or so with this message. It seems slightly
less frequent in 2.0.35 than 2.0.36, so until a day ago (when it went to
2.0.37) it's been on 2.0.35.
It normally happens at times of load, when it's masquerading, doing some
name service, SLIP transfer, mail handling and NFS service at
once. (Yes, it can do all this without swapping to death. Linux is
great.)
Once this has happened, all forks fail forevermore. Existing processes
continue running, very slowly, but as I can't run /bin/su or
/sbin/shutdown this is not very useful :( also, something is sitting in
the kernel (in get_free_page(), maybe?) eating up all the time it can
get, too.
Note that existing processes do *not* keel over with malloc errors, and
swap pressure is not too high (well, there's 40Mb free swap, 20Mb used,
so it's only swapped out twice its total physical memory capacity...)
Unfortunately it is headless so the shift-scroll-lock key can't be used
to get extra info :( but if anyone has a clue how I can try to get extra
info anyway, please say.
There are no modules on this system, and .config looks like
CONFIG_EXPERIMENTAL=y
CONFIG_MODULES=y
CONFIG_KERNELD=y
CONFIG_NET=y
CONFIG_SYSVIPC=y
CONFIG_BINFMT_ELF=y
CONFIG_BINFMT_JAVA=m
CONFIG_KERNEL_ELF=y
CONFIG_M486=y
CONFIG_BLK_DEV_FD=m
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_LOOP=m
CONFIG_FIREWALL=y
CONFIG_INET=y
CONFIG_IP_FORWARD=y
CONFIG_SYN_COOKIES=y
CONFIG_IP_FIREWALL=y
CONFIG_IP_FIREWALL_VERBOSE=y
CONFIG_IP_MASQUERADE=y
CONFIG_IP_MASQUERADE_IPAUTOFW=y
CONFIG_IP_MASQUERADE_ICMP=y
CONFIG_IP_ALWAYS_DEFRAG=y
CONFIG_IP_ACCT=y
CONFIG_IP_ROUTER=y
CONFIG_IP_NOSR=y
CONFIG_NETDEVICES=y
CONFIG_DUMMY=m
CONFIG_SLIP=y
CONFIG_SLIP_COMPRESSED=y
CONFIG_SLIP_SMART=y
CONFIG_NET_ETHERNET=y
CONFIG_NET_VENDOR_3COM=y
CONFIG_EL3=y
CONFIG_CD_NO_IDESCSI=y
CONFIG_CDU31A=m
CONFIG_QUOTA=y
CONFIG_MINIX_FS=m
CONFIG_EXT2_FS=y
CONFIG_NLS=m
CONFIG_ISO9660_FS=m
CONFIG_FAT_FS=m
CONFIG_MSDOS_FS=m
CONFIG_UMSDOS_FS=m
CONFIG_VFAT_FS=m
CONFIG_NLS_CODEPAGE_437=m
CONFIG_NLS_CODEPAGE_850=m
CONFIG_NLS_ISO8859_1=m
CONFIG_PROC_FS=y
CONFIG_NFS_FS=y
CONFIG_HPFS_FS=m
CONFIG_SERIAL=y
CONFIG_PRINTER=y
CONFIG_RTC=y
CONFIG_SOUND=m
CONFIG_SB=y
CONFIG_ADLIB=y
CONFIG_AUDIO=y
CONFIG_MIDI=y
CONFIG_YM3812=y
SBC_BASE=220
SBC_IRQ=5
SBC_DMA=1
SB_DMA2=5
SB_MPU_BASE=0
SB_MPU_IRQ=-1
DSP_BUFFSIZE=65536
-- `dont forget that Linux became only possible because 20 years of OS research was carefully studied, analyzed, discussed and thrown away.' -- mingo on linux-kernel- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/