Cannot access '/dev/pts/292': Value too large for defined data type

hello,

whenever i try open a new pseudo-pty, i get a similar message to the one in the subject, and one like "fstat: Value too large for defined data type" if i open an xterm. the error mentioned in the subject is one i get if i try to re-open a screen session.

i cannot open a screen session in X then, and only on the console (i use screen to be able to move my shell sessions from console to X and back).

this started happening when i opened my current X session.

what can i do to remedy this?

output from scripts/ver_linux, /proc/cpuinfo, and others below. i am not subscribed to the list.

----- scripts/ver_linux (slightly formatted): -----
If some fields are empty or look unusual you may have an old version. Compare to the current minimal requirements in Documentation/Changes.

Linux pylon 2.6.7-1 #1 Sat Jun 19 16:43:25 UTC 2004 ppc GNU/Linux

Gnu C                  3.3.4
Gnu make               3.80
binutils               2.14.90.0.7
util-linux             2.12
mount                  2.12
module-init-tools      3.1-pre2
e2fsprogs              1.35
xfsprogs               2.6.11
quota-tools            3.12.
PPP                    2.4.2
nfs-utils              1.0.6
Linux C Library        2.3.2
Dynamic linker (ldd)   2.3.2
Procps                 3.2.1
Net-tools              1.60
Console-tools          0.2.3
Sh-utils               5.0.91
Modules Loaded         snd_powermac snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd deflate 

des xfrm_user ipcomp esp4 ah4 mach64 sha1 agpgart arc4 nfsd exportfs i2c_keywest cls_fw cls_route cls_u32 cls_rsvp6 cls_rsvp cls_tcindex sch_teql sch_tbf sch_sfq sch_red sch_prio sch_hfsc sch_htb sch_gred sch_ingress sch_csz sch_cbq sch_dsmark ip6table_mangle ip6table_filter ip6t_rt ip6t_ipv6header ip6t_hl ip6t_hbh ip6t_frag ip6t_esp ip6t_dst ip6t_ah ip6t_owner ip6t_mark ip6t_mac ip6t_limit ip6t_length ip6t_multiport ip6t_MARK ip6t_LOG ip6t_eui64 ip6_tables ipt_SAME ipt_NETMAP ipt_iprange ipt_CLASSIFY ip_nat_snmp_basic ip_nat_ftp ip_conntrack_ftp ip_nat_irc ip_conntrack_irc iptable_mangle iptable_filter ipt_ttl ipt_tos ipt_tcpmss ipt_state ipt_pkttype ipt_owner ipt_multiport ipt_mark ipt_mac ipt_recent ipt_limit ipt_length ipt_helper ipt_esp ipt_ecn ipt_dscp ipt_conntrack ipt_ah ipt_ULOG ipt_TOS ipt_TCPMSS ipt_REJECT ipt_REDIRECT ipt_MASQUERADE iptable_nat ip_conntrack ipt_MARK ipt_LOG ipt_ECN ipt_DSCP ip_tables snd_page_alloc sha256 cryptoloop twofish softdog nls_cp437 nls_cp863 tun bsd_comp ppp_async ppp_deflate zlib_deflate pppoe pppox ppp_generic slhc ipddp cdrom isofs zlib_inflate hfs hfsplus nfs lockd sunrpc 3c59x mace
----------

----- /proc/cpuinfo: -----
processor : 0
cpu : 604e
clock : 185MHz
revision : 2.4 (pvr 0009 0204)
bogomips : 356.35
machine : Power Macintosh
motherboard : AAPL,8500 MacRISC
detected as : 16 (PowerMac 8500/8600)
pmac flags : 00000000
L2 cache : 512K unified
memory : 448MB
pmac-generation : OldWorld
----------

----- /proc/ioports: -----
00000000-007fffff : /bandit@F2000000
00000000-000000ff : 0000:00:0f.0
00000400-000004ff : 0000:00:0d.0
00000800-0000087f : 0000:00:0e.0

    00000800-0000087f : 0000:00:0e.0

ff7fe000-ffffdfff : /chaos@F0000000
----------

----- /proc/iomem: -----
80000000-8fffffff : /bandit@F2000000
80800000-8080007f : 0000:00:0e.0
80801000-80801fff : 0000:00:0f.0

    80801000-80801fff : aic7xxx
  80802000-80802fff : 0000:00:0d.0
  81000000-81ffffff : 0000:00:0d.0
    81000000-81ffffff : atyfb
90000000-9fffffff : /chaos@F0000000
  90000000-9000ffff : 0001:01:0b.0
    90001000-90001fff : controlfb regs
  94000000-97ffffff : 0001:01:0b.0
    94000000-97ffffff : controlfb

f1000000-f1ffffff : /chaos@F0000000
f1000000-f10003ff : 0001:01:0d.0
f3000000-f3ffffff : /bandit@F2000000
f3000000-f301ffff : 0000:00:10.0

    f3000000-f301ffff : 0.f3000000:gc
      f3008000-f30080ff : 0.00010000:53c94
        f3008000-f30080ff : mac53c94
      f3008100-f30081ff : 0.00015000:swim3
      f3008200-f30082ff : 0.00011000:mace
        f3008200-f30082ff : mace
      f3008300-f30083ff : 0.00011000:mace
        f3008300-f30083ff : mace
      f3008400-f30084ff : 0.00013020:ch-a
        f3008400-f30084ff : pmac_zilog
      f3008500-f30085ff : 0.00013020:ch-a
        f3008500-f30085ff : pmac_zilog
      f3008600-f30086ff : 0.00013000:ch-b
        f3008600-f30086ff : pmac_zilog
      f3008700-f30087ff : 0.00013000:ch-b
        f3008700-f30087ff : pmac_zilog
      f3008800-f30088ff : 0.00014000:awacs
        f3008800-f30088ff : awacs- Tx DMA
      f3008900-f30089ff : 0.00014000:awacs
        f3008900-f30089ff : awacs- Rx DMA
      f3008a00-f3008aff : 0.00018000:mesh
        f3008a00-f3008aff : mesh
      f3010000-f30100ff : 0.00010000:53c94
        f3010000-f30100ff : mac53c94
      f3011000-f3011fff : 0.00011000:mace
        f3011000-f3011fff : mace
      f3013000-f301301f : 0.00013000:ch-b
        f3013000-f301301f : pmac_zilog
      f3013020-f301303f : 0.00013020:ch-a
        f3013020-f301303f : pmac_zilog
      f3014000-f3014fff : 0.00014000:awacs
        f3014000-f3014fff : awacs
      f3015000-f3015fff : swim3
        f3015000-f3015fff : 0.00015000:swim3
      f3016000-f3017fff : 0.00016000:via-cuda
        f3016000-f3017fff : via-cuda
      f3018000-f30180ff : 0.00018000:mesh
        f3018000-f30180ff : mesh
      f301b000-f301bfff : controlfb cmap
      f301c000-f301c03f : 0.0001c000:sixty6
      f301d000-f301d00f : 0.0001d000:nvram
      f301f000-f301f1ff : 0.0001d000:nvram

f8000000-f80007ff : hammerhead
----------

Puritanism: The haunting fear that someone, somewhere may be happy.

It appears that you're using some variant of the 2.6.7 kernel, yes?

That kernel (and many preceding ones) will create large pty indexes and old (and/or buggy) userspace fails to handle it correctly.

Post-2.6.7, the allocation of pty indexes was switched to first-fit and things should now work OK.

Please test a current kernel and send a report.

2.6.7 plus ftp://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.7-bk9.gz would be suitable.

Ce jour Sat, 26 Jun 2004, Andrew Morton a dit:

yup, just pure mainline, no patches.

ok, i suspected the user space stuff would puke on that, thanks.

oh good news - thank you very much :).

ok. as soon as i can. move it up on my mental TODO list =).

BTW, thanks go to all the kernel hackers and patch submitters (and anyone else that works on or for the kernel :), for doing such a good job, it's appreciated.

Ce jour Sat, 26 Jun 2004, Andrew Morton a dit:

In file included from include/linux/list.h:8,
from include/linux/signal.h:4, from arch/ppc/kernel/asm-offsets.c:12:
include/asm/system.h:85: error: parse error before '{' token include/asm/system.h:85: error: parse error before '<' token make[2]: *** [arch/ppc/kernel/asm-offsets.s] Error 1 make[1]: *** [arch/ppc/kernel/asm-offsets.s] Error 2 make[1]: Leaving directory `/usr/src/linux-2.6'
make: *** [stamp-kernel-configure] Error 2

typos somewhere =)

That was fixed - please try -bk10.

Ce jour Sat, 26 Jun 2004, Andrew Morton a dit:

ok. well, i'm currently running the 2.6.7-bk10, and it runs really nicely for one thing, and the other is that the pty behaviour is exactly like the old one (which i've just been told 'first fit' *is* the old behaviour).

in a way i like the newer behaviour, but hmm, maybe have both behaviours? first fit for some old/buggy apps, and the new one for ones that correctly detect the new capabilities. though IME most of the applications i've used don't recognise the new one, and having varying levels of freakiness - from a warning to outright not working as expected (e.g., my problem with screen, the "Value too large" message).

thanks :)

Software Patents are patently wrong:
http://swpat.ffii.org/papiere/eubsa-swpat0202/ustr0309/index.en.html