Closed Bug 61247 Opened 24 years ago Closed 24 years ago

Crash with SYSTEM-F-ACCVIO

Categories

(Core Graveyard :: Security: UI, defect, P3)

1.0 Branch
DEC
OpenVMS
defect

Tracking

(Not tracked)

VERIFIED WORKSFORME

People

(Reporter: gfruth, Assigned: colin)

References

()

Details

From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0) BuildID: 2000100907 Here is the output: Starting mozilla-bin... Setting content window *** Pulling out the charset Loading page specified via openDialog in SetSecurityButton Document http://www.mozilla.org/ loaded successfully we don't handle eBorderStyle_close yet... please fix me %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=000000000000 001C, PC=00000000013D10E0, PS=0000001B %TRACE-F-TRACEBACK, symbolic stack dump follows image module routine line rel PC abs PC LIBGDK GDKWINDOW gdk_window_ref 31869 0000000000001440 00000000013D10E0 LIBGDK GDKEVENTS gdk_event_copy 20001 0000000000000AB4 00000000013BA0A4 LIBGDK GDKEVENTS gdk_event_put 19949 0000000000000984 00000000013B9F74 LIBWIDGET_GTK NSGTKEVENTHANDLER handle_key_release_event 72270 00000000000012A4 0000000001202804 LIBWIDGET_GTK NSGTKEVENTHANDLER dispatch_superwin_event 72517 0000000000001950 0000000001202EB0 LIBWIDGET_GTK NSGTKEVENTHANDLER handle_gdk_event 72367 0000000000001530 0000000001202A90 LIBGDK GDKEVENTS gdk_event_dispatch 31216 0000000000002690 00000000013BBC80 LIBGLIB GMAIN g_main_dispatch 19203 0000000000000AD0 0000000001311E20 LIBGLIB GMAIN g_main_iterate 19424 000000000000128C 00000000013125DC LIBGLIB GMAIN g_main_run 19482 0000000000001470 00000000013127C0 LIBGTK GTKMAIN gtk_main 20665 0000000000000A58 0000000001562508 LIBWIDGET_GTK NSAPPSHELL Run 72218 00000000000014A4 00000000011F3544 MOZILLA-BIN NSAPPRUNNER main1 58655 0000000000007270 0000000000037270 MOZILLA-BIN NSAPPRUNNER main 58836 0000000000007838 0000000000037838 MOZILLA-BIN NSAPPRUNNER __MAIN 0 0000000000000070 0000000000030070 MOZILLA-BIN 0 0000000000035938 0000000000045938 PTHREAD$RTL 0 000000000003115C 000000007BBC115C PTHREAD$RTL 0 0000000000012B48 000000007BBA2B48 0 FFFFFFFF8FDED3F4 FFFFFFFF8FDED3F4 $ Reproducible: Always Steps to Reproduce: 1. Visit https://abpa-tpa.com/cgi-bin/webm002aa.exe
is there a way you could build a newer version of the mozilla tree? your build is very outdated (m18)?
I'm using the binary distribution from Compaq. I did not build this version.
Three comments: 1. I couldn't reproduce this, even with an OpenVMS build. 2. The PSM code hasn't been ported yet, so this may be possibly a 'crashes on all https: pages' problem. Have you tried other secure sites? 3. The Mozilla 0.6 port should be available now.
Garry Fruth, can you try this with something more recent. Thanks.
I installed M0.6; the result is very similar. However, I found that if I used the pull-down in the address box to visit the URL, I get the message "The connection was refused when attempting to contact abpa-tpa.com". Immediatly after, if I typed in the URL into the address box, I get the following dump: Starting mozilla-bin... Setting content window *** Pulling out the charset Loading page specified via openDialog in SetSecurityButton Document http://www.mozilla.org/releases/mozilla0.6/ loaded successfully we don't handle eBorderStyle_close yet... please fix me Error loading URL https://abpa-tpa.com/cgi-bin/webm002aa: 804b000d we don't handle eBorderStyle_close yet... please fix me ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Got docshell from appCOre we don't handle eBorderStyle_close yet... please fix me %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=000000000000 001C, PC=0000000000C170E0, PS=0000001B %TRACE-F-TRACEBACK, symbolic stack dump follows image module routine line rel PC abs PC LIBGDK GDKWINDOW gdk_window_ref 31869 0000000000001440 0000000000C170E0 LIBGDK GDKEVENTS gdk_event_copy 20001 0000000000000AB4 0000000000C000A4 LIBGDK GDKEVENTS gdk_event_put 19949 0000000000000984 0000000000BFFF74 LIBWIDGET_GTK NSGTKEVENTHANDLER handle_key_release_event 72280 00000000000012A4 0000000000A48804 LIBGTK GTKMARSHAL gtk_marshal_BOOL__POINTER 7901 00000000000000E0 0000000000DAA390 LIBGTK GTKSIGNAL gtk_handlers_run 13692 0000000000003A04 0000000000DE1B34 LIBGTK GTKSIGNAL gtk_signal_real_emit 13252 0000000000002AEC 0000000000DE0C1C LIBGTK GTKSIGNAL gtk_signal_emit 12327 0000000000000B20 0000000000DDEC50 LIBGTK GTKWIDGET gtk_widget_event 26128 0000000000006EF4 0000000000E20214 LIBGTK GTKMAIN gtk_propagate_event 21487 00000000000026B4 0000000000DAA164 LIBGTK GTKMAIN gtk_main_do_event 20959 0000000000001440 0000000000DA8EF0 LIBWIDGET_GTK NSGTKEVENTHANDLER handle_gdk_event 72497 0000000000001860 0000000000A48DC0 LIBGDK GDKEVENTS gdk_event_dispatch 31216 0000000000002690 0000000000C01C80 LIBGLIB GMAIN g_main_dispatch 19203 0000000000000AD0 0000000000B57E20 LIBGLIB GMAIN g_main_iterate 19424 000000000000128C 0000000000B585DC LIBGLIB GMAIN g_main_run 19482 0000000000001470 0000000000B587C0 LIBGTK GTKMAIN gtk_main 20665 0000000000000A58 0000000000DA8508 LIBWIDGET_GTK NSAPPSHELL Run 72239 00000000000014A4 0000000000A39544 MOZILLA-BIN NSAPPRUNNER main1 58700 0000000000006F70 0000000000036F70 MOZILLA-BIN NSAPPRUNNER main 58881 0000000000007538 0000000000037538 MOZILLA-BIN NSAPPRUNNER __MAIN 0 0000000000000070 0000000000030070 MOZILLA-BIN 0 0000000000035638 0000000000045638 PTHREAD$RTL 0 000000000003115C 000000007BBC115C PTHREAD$RTL 0 0000000000012B48 000000007BBA2B48 0 FFFFFFFF8D0A33F4 FFFFFFFF8D0A33F4 The problem does seem to be related to HTTPS. The URL http://abpa- tpa.com/apache/ works with no problems. However https://abpa-tpa.com/apache does not. I was not able to turn off TLS, as the 'Privacy and Security' menu item did not respond. I think that M18 and M0.6 are the most recent binaries for OpenVMS.
we shouldn't crash on https: even if PSM is not installed. Over to crypto for a look.
Assignee: asa → ddrinan
Status: UNCONFIRMED → NEW
Component: Browser-General → Security: Crypto
Ever confirmed: true
QA Contact: doronr → junruh
When specifying an https URL on the startup line... In M18 I don't get a crash, I just get an empty window where the page didn't load. In M0.6 I get a "The connection was refused" popup before the empty window. And as previously mentioned in this report, security is not implemented yet for OpenVMS, so its not meant to work! Of course, it should not crash. If you wish to persue this further can you post the output for a SHOW PROCESS/QUOTA command. Thanks.
Assignee: ddrinan → colin
Accepting bug.
Status: NEW → ASSIGNED
Am I to understand that the OpenVMS build does not support https? The comment "The PSM code hasn't been ported yet" did not mean anything to this user. The release notes at http://www.openvms.compaq.com/openvms/products/ips/mozilla_relnotes.html have a comment that says: "Netscape 6 contains the following additional components: Security component". Was this supposed to inform me that https does not work? If HTTPS is not supposed to work for this release of Mozilla for OpenVMS, then I will leave it up to you to decide whether or not this should be persued. As you requested, here is the output of SHOW PROCESS/QUOTA/ACCOUNTING before and after running Mozilla. I noticed that peak virtual size is nearly 300000; the release notes estimate that a PGFLQUOTA of 150,000 is sufficient - I question that estimate and would suggest 300,000. $ ty before.txt 19-DEC-2000 10:06:28.58 User: SYSTEM Process ID: 0000036B Node: MARS Process name: "_FTA4:" Process Quotas: Account name: SYSTEM CPU limit: Infinite Direct I/O limit: 150 Buffered I/O byte count quota: 99616 Buffered I/O limit: 150 Timer queue entry quota: 20 Open file quota: 399 Paging file quota: 495936 Subprocess quota: 10 Default page fault cluster: 64 AST quota: 248 Enqueue quota: 2000 Shared file limit: 0 Max detached processes: 0 Max active jobs: 0 Accounting information: Buffered I/O count: 162 Peak working set size: 4352 Direct I/O count: 56 Peak virtual size: 169280 Page faults: 553 Mounted volumes: 0 Images activated: 6 Elapsed CPU time: 0 00:00:00.07 Connect time: 0 00:00:18.71 -------------------- $ ty after.txt 19-DEC-2000 10:07:19.67 User: SYSTEM Process ID: 0000036B Node: MARS Process name: "_FTA4:" Process Quotas: Account name: SYSTEM CPU limit: Infinite Direct I/O limit: 150 Buffered I/O byte count quota: 99616 Buffered I/O limit: 150 Timer queue entry quota: 20 Open file quota: 399 Paging file quota: 495808 Subprocess quota: 10 Default page fault cluster: 64 AST quota: 248 Enqueue quota: 2000 Shared file limit: 0 Max detached processes: 0 Max active jobs: 0 Accounting information: Buffered I/O count: 3180 Peak working set size: 77840 Direct I/O count: 6152 Peak virtual size: 284768 Page faults: 4614 Mounted volumes: 0 Images activated: 9 Elapsed CPU time: 0 00:00:05.54 Connect time: 0 00:01:09.80
Correct. The is no HTTPS support in the OpenVMS version at the moment. Sorry the release note was cryptic; I'll get that changed. The think the page file quota estimate of 150,000 is reasonable since a process does not need a page file quota equal to its virtual page count (only writeable pages need space in the page file space and much of Mozilla's bulk is read-only code). I still have no idea why trying to access secure pages causes an accvio for you but not for me (and others). Clearly it must be something specific to your environment. Can you think of anything that might be relevant? Have you editted any Mozilla configuration files? Was M18 the first version of Mozilla you tried?
Follow up to my previous reply. I think we'll be leaving the release notes as they are. The note on Security (HTTPS) seems reasonably clear. http://www.openvms.compaq.com/openvms/products/ips/mozilla_relnotes.html#notes Security for Mozilla on OpenVMS Q: When will security be available for Mozilla on OpenVMS? A: Netscape, through iPlanet, has made the Personal Security Manager for Mozilla and NSS open source cryptographic code publicly available. This code is currently in the final stages of integration into the Mozilla build. We expect that when integration is complete, security can be made available in the OpenVMS version of Mozilla.
I don't consider the environment that I am running Mozilla in particularly unique. We aquired the box in October with VMS 7.2-1 pre-installed. I have installed a few other items on it. For the record, here is the output of PRODUCT SHOW HISTORY $ PRODUCT SHOW HIST ----------------------------------- ----------- ----------- -------------------- PRODUCT KIT TYPE OPERATION DATE AND TIME ----------------------------------- ----------- ----------- -------------------- CPQ AXPVMS MOZILLA M0.6 Full LP Install 12-DEC-2000 12:58:28 CPQ AXPVMS MOZILLA M0.6 Full LP Remove 12-DEC-2000 12:54:52 CPQ AXPVMS MOZILLA M0.6 Full LP Install 12-DEC-2000 12:34:12 CPQ AXPVMS MOZILLA I5.0 Full LP Remove 12-DEC-2000 12:34:12 DEC AXPVMS VMS721_BACKUP V1.0 Patch Install 16-NOV-2000 11:23:46 DEC AXPVMS DFU V2.7 Full LP Install 25-OCT-2000 13:33:42 DEC AXPVMS VMS721_FIBRECHAN V3.0 Patch Install 18-OCT-2000 15:45:47 DEC AXPVMS VMS721_F11X V2.0 Patch Install 18-OCT-2000 15:36:30 CPQ AXPVMS MOZILLA I5.0 Full LP Install 18-OCT-2000 11:48:16 CPQ AXPVMS CSWS V1.0 Full LP Install 05-OCT-2000 10:10:32 CPQ AXPVMS CSWS T1.0 Full LP Remove 05-OCT-2000 10:10:32 CPQ AXPVMS CSWS T1.0 Full LP Install 29-SEP-2000 18:40:06 DEC AXPVMS DECNET_PHASE_IV V7.2-1 Full LP Install 28-SEP-2000 12:13:36 DEC AXPVMS DECNET_OSI V7.2-1 Full LP Remove 28-SEP-2000 12:13:04 DEC AXPVMS DNVOSIECO01 V7.2 Patch Remove 28-SEP-2000 12:13:04 DEC AXPVMS VMS721_SYS V5.0 Patch Install 11-APR-2000 09:46:24 DEC AXPVMS VMS721_PCSI V1.0 Patch Install 11-APR-2000 09:41:38 DEC AXPVMS NS_NAV_EXPORT V3.3 Full LP Install 30-MAR-2000 09:47:29 DEC AXPVMS DNVOSIECO01 V7.2 Patch Install 30-MAR-2000 09:40:34 DEC AXPVMS VMS721_LAN V1.0 Patch Install 30-MAR-2000 08:45:11 DEC AXPVMS DWMOTIF125_UPD V1.0 Patch Install 29-MAR-2000 15:49:13 DEC AXPVMS DECNET_OSI V7.2-1 Full LP Install 29-MAR-2000 15:04:17 DEC AXPVMS VMS721_DQCONFIG V2.0 Patch Install 29-MAR-2000 14:54:50 DEC AXPVMS VMS721_UPDATE V1.0 Patch Install 29-MAR-2000 14:50:16 DEC AXPVMS DWMOTIF V1.2-5 Full LP Install 29-MAR-2000 08:57:50 DEC AXPVMS OPENVMS V7.2-1 Platform Install 29-MAR-2000 08:57:50 DEC AXPVMS TCPIP V5.0-10 Full LP Install 29-MAR-2000 08:57:50 DEC AXPVMS VMS V7.2-1 Oper System Install 29-MAR-2000 08:57:50 ----------------------------------- ----------- ----------- -------------------- The box is a 2 CPU DS20E with 1 Gig of member. Other than myself, there are no users. It is running the Apache web server, using DEC-TCPIP 5.0A, and Decnet phase IV. I'm pretty good using the VMS debugger. I don't have access to a C compiler here. If you are willing to send a version with all the modules compiled with debug, along with the source from which they were compiled, I'll be happy to step through it and see if I can find the bug. As far as the release notes, we could differ on what is clear. But this is unimportant since you will have SSL support soon. We are planning on using this box as our primary Internet server (HTTPS, FTP...). It is embarrassing to have a www server that does not have a decent browser. Any hint when I can expect to see a version of Mozilla that supports SSL?
There's certainly nothing unusual looking about your system. In fact, I have the same system myself. The HTTPS code (PSM and NSS) was meant to be checked in to the Mozilla tree today. But when I checked a few minutes ago it wasn't present. Once the code is checked in, the porting to OpenVMS phase can start. Hopefully the porting won't take too long, but with it being the holidays its not going to happen overnight. If all goes well I hope to have it running by mid January, which means that it should appear in the M0.8 released (estimated early Feb). Given that there is going to be much change in the code for support of HTTPS, I really don't think its worth my spending a lot of time trying to fix this problem. With your permission, I'd like to close the report and move on. Please note that current estimates from Mozilla put the official release of Mozilla 1.0 in April at best, and possibly slipping to May or later. Until then, all releases should be considered beta quality and I strongly suggest that you DO NOT run them in a production environment.
You have my permission to close this and move on. Thanks for asking.
As discussed, closing this bug report.
Status: ASSIGNED → RESOLVED
Closed: 24 years ago
Resolution: --- → WORKSFORME
Verified per above comments.
Status: RESOLVED → VERIFIED
Mass changing Security:Crypto to PSM
Component: Security: Crypto → Client Library
Product: Browser → PSM
Version: other → 2.1
Mass changing Security:Crypto to PSM
Product: PSM → Core
Version: psm2.1 → 1.0 Branch
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.