On Wed, Oct 23, 2013 at 08:45 +0100, Stuart Henderson wrote: > The panic seems odd (0x594f69b5 != > 0x594f69b5 - these are the same value) , maybe someone else will be > able to comment on that. > > It would be useful to know what isakmpd is doing at the time e.g. run it in > the foreground with fairly high debug settings (maybe "isakmpd -K -d -DA=99" > and "ipsecctl -f /etc/ipsrc.conf" - there will be a lot of output so run from > ssh not serial console). > > You may be able to workaround by using 'boot -c' at the bootloader prompt, > then 'disable glxsb' and 'quit', or the kernel on-disk could be modified > using 'config -ef /bsd'. (glxsb is for hardware AES acceleration for Geode LX > systems). > > iamatt <iam...@gmail.com> wrote: > ># Data modified on freelist: word 153288032 of object 0xd17218e0 size > >0x18 > >previ > >ous type ??? (invalid addr > >0x3557935e) > >panic: Data modified on freelist: word 3 of object 0xd17218e0 size 0x18 > >previous > > type ??? (0x594f69b5 != > >0x594f69b5) > > > > > >Stopped at Debugger+0x4: popl > >%ebp > >RUN AT LEAST 'trace' AND 'ps' AND INCLUDE OUTPUT WHEN REPORTING THIS > >PANIC! > >DO NOT EVEN BOTHER REPORTING THIS WITHOUT INCLUDING THAT > >INFORMATION! > >ddb> > >trace > >Debugger(d0976848,f53be668,d0952384,f53be668,d0a95014) at > >Debugger+0x4 > >panic(d0952384,d095232c,3,d17218e0,18) at > >panic+0x67 > >malloc(18,6c,a,f53be6fc,d1613000) at > >malloc+0x5e4 > >glxsb_crypto_newsession(f53be73c,f53be794,0,10,40) at > >glxsb_crypto_newsession+0 > >x133 > > > >crypto_newsession(d17264f0,f53be794,0,f53be7b8,d0b05340) at > >crypto_newsession+0 > >x135 > > > >esp_init(d1726400,d0a96cd8,f53be900,d1724f50,d1724f68) at > >esp_init+0x245 > >pfkeyv2_send(d5d1f008,d1724e00,198,d1724f98,d0b05340) at > >pfkeyv2_send+0x1a6e > >pfkey_register(d5d8cc00,d5d1f008,f53bedbc,d03c0a4a,88ab98a0) at > >pfkey_register+ > >0x278 > > > >raw_usrreq(d5d1f008,9,d5d8cc00,0,0) at > >raw_usrreq+0x221 > >sosend(d5d1f008,0,f53beeb0,d5d8cc00,0) at > >sosend+0x48d > >soo_write(d5d230ac,d5d230c8,f53beeb0,d5e202d0,55187908) at > >soo_write+0x3b > >dofilewritev(d5d3c618,8,d5d230ac,84fa7500,11) at > >dofilewritev+0x133 > >sys_writev(d5d3c618,f53bef64,f53bef84,f53befa8,d5e11600) at > >sys_writev+0x7c > >syscall() at > >syscall+0x1f9 > >--- syscall (number 5) > >--- > >0x2: > > > >ddb> > >ps > > PID PPID PGRP UID S FLAGS WAIT > >COMMAND > > 31132 1 31132 0 3 0x83 ttyin > >ksh > > 6230 1 6230 0 3 0x80 select > >cron > > 1331 1 1331 601 3 0x90 kqread > >unbound > > 30501 1 30501 535 3 0x90 nanosleep > >symon > > 16688 1 16688 99 3 0x90 poll > >sndiod > > 27774 30105 30105 0 3 0xb0 netcon2 > >sendmail > > 30105 1 30105 0 3 0xb0 select > >sendmail > > 11400 1 11400 77 3 0x90 poll > >dhcpd > > 1054 1 1054 0 3 0x80 kqread > >ifstated > > 7646 1 7646 0 3 0x80 select > >sshd > > 18014 0 0 0 3 0x100280 nfsidl > >nfsio > > 20915 0 0 0 3 0x100280 nfsidl > >nfsio > > 8480 0 0 0 3 0x100280 nfsidl > >nfsio > > 13835 0 0 0 3 0x100280 nfsidl > >nfsio > >*25527 31834 31834 68 7 0x10 > >isakmpd > > 31834 1 31834 0 3 0x80 netio > >isakmpd > > 19715 24693 17897 83 3 0x90 poll > >ntpd > > 24693 17897 17897 83 3 0x90 poll > >ntpd > > 17897 1 17897 0 3 0x80 poll > >ntpd > > 630 4742 4742 74 3 0x90 bpf > >pflogd > > 4742 1 4742 0 3 0x80 netio > >pflogd > > 9803 9138 9138 73 2 0x90 > >syslogd > > 9138 1 9138 0 3 0x80 netio > >syslogd > > 18744 1 18744 77 3 0x90 poll > >dhclient > > 22501 1 22501 0 3 0x80 poll > >dhclient > > 13 0 0 0 3 0x100200 aiodoned > >aiodoned > > 12 0 0 0 3 0x100200 syncer > >update > > 11 0 0 0 3 0x100200 cleaner > >cleaner > > 10 0 0 0 3 0x100200 reaper > >reaper > > 9 0 0 0 3 0x100200 pgdaemon > >pagedaemon > > 8 0 0 0 3 0x100200 bored > >crypto > > 7 0 0 0 3 0x100200 pftm > >pfpurge > > 6 0 0 0 3 0x100200 usbtsk > >usbtask > > 5 0 0 0 3 0x100200 usbatsk > >usbatsk > > 4 0 0 0 3 0x100200 bored > >syswq > > 3 0 0 0 3 0x40100200 > >idle0 > > 2 0 0 0 3 0x100200 kmalloc > >kmthread > > 1 0 1 0 3 0x82 wait > >init > > 0 -1 0 0 3 0x200 scheduler > >swapper > >ddb> > > > > > > > >On Mon, Oct 21, 2013 at 5:20 PM, Stuart Henderson <st...@openbsd.org> > >wrote: > > > >> On 2013/10/21 10:29, iamatt wrote: > >> > I do not have the console debug screen but I do have the files from > >> > /var/crash/ Is there some commands I can run on them using gdb > >that can > >> > be of use? > >> > >> If you have a crashdump, possibly, see "man crash". > >> > >> As the panic message goes, > >> > >> RUN AT LEAST 'trace' AND 'ps' AND INCLUDE OUTPUT WHEN REPORTING THIS > >PANIC! > >> DO NOT EVEN BOTHER REPORTING THIS WITHOUT INCLUDING THAT INFORMATION! > >> > >> This is much easier than for somebody else to install Linux > >> and the shrewsoft client before they can even start looking at the > >bug > >> (and even then, they might not be able to replicate it). >
hi, this is partially my screwup. please try the diff below. we forget to actually allocate the memory for our key schedule but free it later on in the glxsb_crypto_freesession. that's why you get memory corruption all the way. it has another seemingly unrelated chunk (pctr.h -> cpufunc.h change) that i would like to commit as well (perhaps separately) to limit pctr.h usage in kernel. so it would be nice to test the whole thing together. i also reformat one malloc block to shorten it (: ok's are welcome (assuming this will pass the test). diff --git sys/arch/i386/pci/glxsb.c sys/arch/i386/pci/glxsb.c index dacb500..59a7e06 100644 --- sys/arch/i386/pci/glxsb.c +++ sys/arch/i386/pci/glxsb.c @@ -32,7 +32,7 @@ #include <sys/timeout.h> #include <machine/bus.h> -#include <machine/pctr.h> +#include <machine/cpufunc.h> #include <dev/rndvar.h> #include <dev/pci/pcivar.h> @@ -398,14 +398,24 @@ glxsb_crypto_newsession(uint32_t *sidp, struct cryptoini *cri) case CRYPTO_AES_CBC: if (c->cri_klen != 128) { - swd = malloc(sizeof(struct swcr_data), M_CRYPTO_DATA, - M_NOWAIT|M_ZERO); + swd = malloc(sizeof(struct swcr_data), + M_CRYPTO_DATA, M_NOWAIT|M_ZERO); if (swd == NULL) { glxsb_crypto_freesession(sesn); return (ENOMEM); } ses->ses_swd_enc = swd; txf = &enc_xform_rijndael128; + if (txf->ctxsize > 0) { + swd->sw_kschedule = + malloc(txf->ctxsize, + M_CRYPTO_DATA, + M_NOWAIT|M_ZERO); + if (swd->sw_kschedule == NULL) { + glxsb_crypto_freesession(sesn); + return (EINVAL); + } + } if (txf->setkey(&(swd->sw_kschedule), c->cri_key, c->cri_klen / 8) < 0) { glxsb_crypto_freesession(sesn);