Closed Bug 10498 Opened 25 years ago Closed 24 years ago

shmat failure on startup in Solaris 7

Categories

(SeaMonkey :: Build Config, defect, P3)

Product:
SeaMonkey
Component:
Build Config
Platform:
Sun
Solaris
Type:
defect

Tracking

(Not tracked)

Status:
VERIFIED FIXED

People

(Reporter: barnes, Assigned: blizzard)

Details

Every time apprunner is started on Solaris 7, this warning is generated:

Gdk-WARNING **: shmat failed!

The warning does not occur in Solaris 2.6.  Last tested with build 1999/07/26.
Full transcript follows:

hydra >./mozilla-apprunner.sh
MOZILLA_FIVE_HOME=/app/build/package
  LD_LIBRARY_PATH=/app/build/package:/opt/SUNWspro/lib
      MOZ_PROGRAM=./apprunner
        moz_debug=0
     moz_debugger=
Registered Ok
*** Registering html library
register filter service
Reading file...
Reading file...Done
Reading file...
Reading file...Done
Reading file...
Reading file...Done
Reading file...
Reading file...Done
File
Reading file...
Reading file...Done
Edit
View
Search
Go
Bookmarks
Tasks
Help
Debug
QA
nsWidget::RealizeSignal(c0a00)

Gdk-WARNING **: shmat failed!
generating focus event
Doing Startup...
etc.
Assignee: don → pavlov
owen?  any ideas?
owen do you have any ideas on this one?
I used truss to find out more about this.  Under Solaris 2.7, shmat is called 8
times; normally the seventh call fails since the default shared memory segment
limit per process is 6.  Under Solaris 2.6, shmat is only called twice which
fits into the default limit, so no error message is produced.  The same
behaviour occurs with gimp, so the "bug" is presumably in GTK+ or GLIB.  You
might want to pass it onto the GTK developers.

It's possible to work around it by increasing the kernel parameter
shmsys:shminfo_shmseg, but system managers aren't going to want to do this on
every machine.
Assignee: pavlov → blizzard
Severity: normal → critical
Target Milestone: M15
blizzard, can you force owen to sit down and figure this one out?
This is probably because gdkrgb creates 6 cached images that use shared memory
segments.  To fix this we need to make that a runtime variable and make it
smaller by default for Solaris since it is lame.
Component: Browser-General → Build Config
QA Contact: leger → granrose
setting component/qa contact.
Status: NEW → RESOLVED
Closed: 25 years ago
Resolution: --- → WONTFIX
This is a feature request for GTK, really.  Doesn't belong here.
Status: RESOLVED → VERIFIED
didn't realize anyone was putting me on QA contact anywhere.

verified wontfix.
Reopening, as this needs to be dealt with in some way before release.  Probably,
this means creating a release note which contains the GNOME bug number so that
all the Solaris users who are annoyed by this can pester the GDK developers to
make the limit configurable (and pester Sun to up the default).

Additionally, this does seem to happen for me in 2.6 as well.
Status: VERIFIED → REOPENED
Clearing WON'T FIX resolution due to reopen.
Resolution: WONTFIX → ---
I've documented this in bug #26315.
Status: REOPENED → RESOLVED
Closed: 25 years ago → 25 years ago
Resolution: --- → FIXED
verified.
Status: RESOLVED → VERIFIED
Please ignore the spam.  Changing address.
Assignee: blizzard → blizzard
Status: VERIFIED → NEW
bustage from my reassign
Status: NEW → RESOLVED
Closed: 25 years ago → 24 years ago
bustage from my reassign
Status: RESOLVED → VERIFIED
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.