Re: Bug#728705: gdb fails on s390x with "Couldn't write registers: Invalid argument"
Control: reassign -1 linux 3.2.51-1
Control: affects -1 gdb
Control: notforwarded -1
So, it turns out that the 0x5001 is PTRACE_POKEUSR_AREA, which is
essentially equivalent to PTRACE_POKEUSER in a loop, except for being
less clear when it fails, because you can't tell which word couldn't be
written.
Accordingly, I've replaced the failing call with a loop.
diff --git a/gdb/s390-linux-nat.c b/gdb/s390-linux-nat.c
index ee9b617..de9d956 100644
--- a/gdb/s390-linux-nat.c
+++ b/gdb/s390-linux-nat.c
@@ -248,6 +248,9 @@ store_regs (const struct regcache *regcache, int tid, int regnum)
{
gregset_t regs;
ptrace_area parea;
+#if 1
+ int i;
+#endif
parea.len = sizeof (regs);
parea.process_addr = (addr_t) ®s;
@@ -257,8 +260,17 @@ store_regs (const struct regcache *regcache, int tid, int regnum)
fill_gregset (regcache, ®s, regnum);
+#if 0
if (ptrace (PTRACE_POKEUSR_AREA, tid, (long) &parea) < 0)
perror_with_name (_("Couldn't write registers"));
+#else
+ for (i = 0; i < parea.len; i += sizeof(void *))
+ {
+ if (ptrace (PTRACE_POKEUSER, tid, parea.kernel_addr + i,
+ *(void**)(parea.process_addr + i)) < 0)
+ perror_with_name (_("Couldn't write registers"));
+ }
+#endif
}
/* Fetch all floating-point registers from process/thread TID and store
Running "strace -e ptrace gdb -batch -ex start -ex bt --args echo foo"
then results in the failing part looking like:
,----
| ptrace(0x5000 /* PTRACE_??? */, 25005, 0x3ffffad7d80, 0x80073fa0) = 0
| ptrace(0x5000 /* PTRACE_??? */, 25005, 0x3ffffad7cf0, 0x80074410) = 0
| ptrace(PTRACE_POKEUSER, 25005, psw_mask, 0x705c00180000000) = -1 EINVAL (Invalid argument)
`----
However, this same value was just read from that register a bit earlier:
,----
| ptrace(PTRACE_PEEKUSER, 25005, psw_mask, [0x705c00180000000]) = 0
`----
So, this looks like a kernel bug from here.
Kernel stuff
============
zelenka is currently running linux-image-3.2.0-4-s390x version 3.2.51-1
,----[ /proc/cpuinfo ]
| vendor_id : IBM/S390
| # processors : 2
| bogomips per cpu: 9398.00
| features : esan3 zarch stfle msa ldisp eimm dfp etf3eh highgprs
| processor 0: version = FF, identification = 06A18A, machine = 2817
| processor 1: version = FF, identification = 06A18A, machine = 2817
`----
While it wasn't exactly easy, I managed to run the kernel bugscript
outside the chroot; the output is attached.
** Version:
Linux version 3.2.0-4-s390x (debian-kernel@lists.debian.org) (gcc version 4.6.3 (Debian 4.6.3-14) ) #1 SMP Debian 3.2.51-1
** Command line:
ro dasd=200-21f vmpoff="LOGOFF" root=/dev/dasdb1 elevator=deadline BOOT_IMAGE=0
** Not tainted
** Kernel log:
[7340394.293000] [30285] 3227 30285 567 1 0 0 0 sh
[7340394.293003] [30286] 3227 30286 2419 6 1 0 0 debuild
[7340394.293006] [30326] 3227 30326 567 1 1 0 0 sh
[7340394.293009] [30327] 3227 30327 2938 8 1 0 0 dpkg-buildpacka
[7340394.293012] [30328] 3227 30328 536 0 0 0 0 tee
[7340394.293015] [30507] 3227 30507 575 0 0 0 0 rules
[7340394.293018] [30517] 3227 30517 3182 0 0 0 0 dh
[7340394.293021] [ 3116] 3227 3116 575 0 0 0 0 rules
[7340394.293024] [ 3130] 3227 3130 567 2 0 0 0 xvfb-run
[7340394.293027] [ 3142] 3227 3142 2690 4 1 0 0 Xvfb
[7340394.293030] [ 3149] 3227 3149 707 0 0 0 0 make
[7340394.293033] [ 5981] 3227 5981 708 1 1 0 0 make
[7340394.293036] [ 5984] 3227 5984 1165290 241254 1 0 0 lt-octave-cli
[7340394.293039] [ 7843] 0 7843 1174 65 0 0 0 cron
[7340394.293042] [ 7844] 0 7844 1174 65 0 0 0 cron
[7340394.293045] [ 7845] 0 7845 1173 109 1 0 0 cron
[7340394.293048] [ 7846] 0 7846 1173 114 1 0 0 cron
[7340394.293050] Out of memory: Kill process 5984 (lt-octave-cli) score 961 or sacrifice child
[7340394.293053] Killed process 5984 (lt-octave-cli) total-vm:4661160kB, anon-rss:965016kB, file-rss:0kB
[7344121.447118] lt-octave-cli invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
[7344121.447124] lt-octave-cli cpuset=/ mems_allowed=0
[7344121.447128] CPU: 0 Not tainted 3.2.0-4-s390x #1 Debian 3.2.51-1
[7344121.447130] Process lt-octave-cli (pid: 21694, task: 000000003dd7e738, ksp: 0000000020b1b8d8)
[7344121.447133] 0000000000000000 0000000020b1b880 0000000000000002 0000000000000000
[7344121.447135] 0000000020b1b920 0000000020b1b898 0000000020b1b898 00000000003b9c0e
[7344121.447139] 00000000003c77c8 0000000000000000 0000000000000000 000000000000000a
[7344121.447142] 0000000000000000 000000000000000c 0000000020b1b880 0000000020b1b8f8
[7344121.447145] 00000000003d2260 0000000000011a5e 0000000020b1b880 0000000020b1b8d0
[7344121.447148] Call Trace:
[7344121.447150] ([<00000000000201da>] sys_dump_fcp_lun_store+0x32/0x60)
[7344121.447157] Mem-Info:
[7344121.447158] DMA per-cpu:
[7344121.447160] CPU 0: hi: 186, btch: 31 usd: 9
[7344121.447162] CPU 1: hi: 186, btch: 31 usd: 177
[7344121.447166] active_anon:121262 inactive_anon:121314 isolated_anon:0
[7344121.447167] active_file:434 inactive_file:480 isolated_file:0
[7344121.447169] unevictable:26 dirty:0 writeback:27 unstable:0
[7344121.447170] free:1012 slab_reclaimable:1825 slab_unreclaimable:2749
[7344121.447171] mapped:579 shmem:26 pagetables:2070 bounce:0
[7344121.447177] DMA free:4048kB min:4064kB low:5080kB high:6096kB active_anon:485048kB inactive_anon:485256kB active_file:1736kB inactive_file:1920kB unevictable:104kB isolated(anon):0kB isolated(file):0kB present:1034240kB mlocked:104kB dirty:0kB writeback:108kB mapped:2316kB shmem:104kB slab_reclaimable:7300kB slab_unreclaimable:10996kB kernel_stack:1360kB pagetables:8280kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:5875 all_unreclaimable? yes
[7344121.447183] lowmem_reserve[]: 0 0 0
[7344121.447186] DMA: 500*4kB 0*8kB 0*16kB 0*32kB 0*64kB 0*128kB 0*256kB 0*512kB 2*1024kB = 4048kB
[7344121.447192] 2466 total pagecache pages
[7344121.447194] 1535 pages in swap cache
[7344121.447196] Swap cache stats: add 9230940, delete 9229405, find 105881464/106245976
[7344121.447198] Free swap = 0kB
[7344121.447199] Total swap = 2620852kB
[7344121.453288] 262144 pages RAM
[7344121.453290] 6585 pages reserved
[7344121.453292] 671 pages shared
[7344121.453293] 253044 pages non-shared
[7344121.453295] [ pid ] uid tgid total_vm rss cpu oom_adj oom_score_adj name
[7344121.453302] [ 306] 0 306 1134 2 1 -17 -1000 udevd
[7344121.453306] [ 2016] 1 2016 481 1 1 0 0 uptimed
[7344121.453310] [ 2160] 0 2160 972 0 1 0 0 atd
[7344121.453313] [ 2206] 110 2206 1630 1 1 0 0 stunnel4
[7344121.453316] [ 2207] 110 2207 1630 1 1 0 0 stunnel4
[7344121.453320] [ 2208] 110 2208 1630 1 1 0 0 stunnel4
[7344121.453323] [ 2209] 110 2209 1630 1 1 0 0 stunnel4
[7344121.453326] [ 2210] 110 2210 1630 1 1 0 0 stunnel4
[7344121.453329] [ 2211] 110 2211 2069 0 0 0 0 stunnel4
[7344121.453332] [ 2214] 105 2214 2341 53 1 0 0 ntpd
[7344121.453336] [ 2219] 0 2219 527 0 0 0 0 ulogd
[7344121.453339] [ 2301] 111 2301 5000 259 1 0 0 unbound
[7344121.453342] [ 2448] 106 2448 4955 3 1 0 0 exim4
[7344121.453345] [ 2999] 0 2999 836 0 1 0 0 bash
[7344121.453348] [ 3000] 0 3000 1417 22 0 0 0 monit
[7344121.453352] [21883] 0 21883 2864 0 0 0 0 sshd
[7344121.453355] [21889] 113 21889 2968 0 1 0 0 sshd
[7344121.453358] [23316] 112 23316 5087 0 0 0 0 bacula-fd
[7344121.453361] [12696] 0 12696 475 0 1 0 0 ekeyd-egd-linux
[7344121.453364] [20307] 2830 20307 1676 0 1 0 0 screen
[7344121.453367] [20308] 2830 20308 2005 1 1 0 0 bash
[7344121.453371] [20471] 2830 20471 2011 3 1 0 0 bash
[7344121.453373] [ 9746] 0 9746 2932 106 0 0 0 munin-node
[7344121.453377] [10002] 113 10002 2831 0 1 0 0 /usr/share/muni
[7344121.453380] [28336] 0 28336 4286 57 1 0 0 samhain
[7344121.453383] [29409] 0 29409 1057 1 0 0 0 syslog-ng
[7344121.453386] [29410] 0 29410 4307 211 0 0 0 syslog-ng
[7344121.453389] [ 4726] 0 4726 1133 0 0 -17 -1000 udevd
[7344121.453393] [ 4727] 0 4727 1133 0 1 -17 -1000 udevd
[7344121.453396] [ 7734] 0 7734 1712 24 1 -17 -1000 sshd
[7344121.453399] [25295] 0 25295 1069 5 0 0 0 cron
[7344121.453402] [ 1600] 0 1600 2864 2 0 0 0 sshd
[7344121.453405] [21657] 0 21657 4286 0 1 0 0 samhain
[7344121.453408] [21671] 102 21671 2279 2 1 0 0 zabbix_agentd
[7344121.453411] [21672] 102 21672 2279 37 1 0 0 zabbix_agentd
[7344121.453415] [21673] 102 21673 2279 3 1 0 0 zabbix_agentd
[7344121.453418] [21674] 102 21674 2279 3 0 0 0 zabbix_agentd
[7344121.453421] [21676] 102 21676 2279 3 0 0 0 zabbix_agentd
[7344121.453424] [ 9686] 0 9686 2864 1 0 0 0 sshd
[7344121.453427] [17558] 108 17558 1638 5 0 0 0 nrpe
[7344121.453430] [19683] 0 19683 2911 0 0 0 0 sshd
[7344121.453433] [19689] 3227 19689 2911 27 0 0 0 sshd
[7344121.453436] [19690] 3227 19690 1976 3 1 0 0 bash
[7344121.453440] [19729] 3227 19729 8151 1 1 0 0 schroot
[7344121.453443] [19730] 3227 19730 993 3 1 0 0 bash
[7344121.453446] [21694] 3227 21694 903176 240715 0 0 0 lt-octave-cli
[7344121.453448] Out of memory: Kill process 21694 (lt-octave-cli) score 970 or sacrifice child
[7344121.453451] Killed process 21694 (lt-octave-cli) total-vm:3612704kB, anon-rss:960744kB, file-rss:2116kB
** Model information
processor 0: version = FF, identification = 06A18A, machine = 2817
processor 1: version = FF, identification = 06A18A, machine = 2817
** Loaded modules:
nf_conntrack_ipv6
nf_defrag_ipv6
ip6t_REJECT
ip6t_LOG
ip6table_filter
ip6_tables
nf_conntrack_ipv4
nf_defrag_ipv4
ipt_REJECT
ipt_ULOG
xt_tcpudp
xt_state
nf_conntrack
xt_hashlimit
xt_multiport
iptable_filter
ip_tables
x_tables
qeth_l2
dm_snapshot
appldata_net_sum
appldata_mem
appldata_os
qeth
ccwgroup
vmur
ext3
mbcache
jbd
dm_mod
dasd_fba_mod
dasd_eckd_mod
dasd_mod
** PCI devices:
not available
** USB devices:
not available
The error return is presumably from this code:
<http://sources.debian.net/src/linux/3.2.51-1/arch/s390/kernel/ptrace.c#L294>
I'd be happy to get you any further information, though I assume you
would find it easier to simply "ssh zelenka.debian.org" to answer most
of the questions you might typically ask.
Of course, to try another kernel, you'd need help from DSA.
Reply to: