leaking pixmaps

Bug #177680 reported by David
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
firefox (Ubuntu)
Invalid
Undecided
Unassigned
flashplugin-nonfree (Ubuntu)
Expired
Undecided
Unassigned

Bug Description

Binary package hint: xorg

I first noticed this in Firefox. I thought it was related to https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/137764 but it appears to be more than that.
When I open a page containing large numbers of images, memory does balloon, as described. However, when I close that tab, not all memory is returned: about 1-2 megs remains. Opening, browsing, and closing dozens of tabs can quickly lose hundreds of megs. Even if I return to a single empty tab, this apparently "leaked" memory is never returned.
This does not just happen in Firefox, but in fact in anything that caches pixmap data apparently. I also tried in web browsers Epiphany and Opera, but noticed the same behaviour in Nautilus when looking at directories with a large number of images or videos that get thumbnails.
I have an ATI Radeon 9000 card, and a colleague mentioned he'd had memory leak problems with the ATI driver when he had a Radeon 9800. This may be related.
Hopefully I've provided enough information

Tags: packaging
Revision history for this message
David (wizard-houseoffire) wrote :
Revision history for this message
David (wizard-houseoffire) wrote :
Revision history for this message
Timo Aaltonen (tjaalton) wrote :

When you kill the browser the memory is released, right?

Changed in xorg:
status: New → Incomplete
Revision history for this message
David (wizard-houseoffire) wrote :

yes, it is, which is why I hesitated to label this with xorg, but with the apparent widespread nature of what I'm seeing, I didn't know where else to put it.

Revision history for this message
Timo Aaltonen (tjaalton) wrote :

The xserver just shows all the mapped pixmem, so the leakage is from the app. Note that flashplugin is very keen on memory, and won't release it until the browser is killed.

Changed in xorg:
status: Incomplete → New
Revision history for this message
David (wizard-houseoffire) wrote :

As I noted, this doesn't just occur in Firefox, but all browsers (including opera) whether the page had flash or not, and also Nautilus file browser.

Revision history for this message
Alexander Sack (asac) wrote : Re: [Bug 177680] Re: leaking pixmaps

On Sat, Dec 22, 2007 at 10:31:14PM -0000, Timo Aaltonen wrote:
> The xserver just shows all the mapped pixmem, so the leakage is from the
> app. Note that flashplugin is very keen on memory, and won't release it
> until the browser is killed.
>
> ** Changed in: firefox (Ubuntu)
> Sourcepackagename: xorg => firefox
> Status: Incomplete => New
>

moving to flashplugin-nonfree

 affects ubuntu/flashplugin-nonfree

 affects ubuntu/firefox
 status invalid

Thanks,

 - Alexander

Changed in firefox:
status: New → Invalid
Revision history for this message
David (wizard-houseoffire) wrote :

Attached is a screenshot I took after about a week of usage. Note the memory usage.
I surfed several websites over the week, web comics, youtube, regular surfing, opening and closing several hundred tabs. I closed all but the one tab to take this screen shot.

Revision history for this message
David (wizard-houseoffire) wrote :

I then closed firefox. See attached screenshot.
Some memory was returned, but Xorg still seems to be taking a lot. Or am I mistaken in my interpretation?

Revision history for this message
David (wizard-houseoffire) wrote :

I immediately restarted. I notice how little Xorg takes compared to my last screen shot with nothing running.

Revision history for this message
David (wizard-houseoffire) wrote :

I restarted firefox, and loaded up the same single website kingdomofloathing.com. Very little mark in the xorg memory usage.

Revision history for this message
David (wizard-houseoffire) wrote :

I then removed the flash plugin, restarted, and brought firefox back up. The attached screen shot is almost 2 days of usage. Progression suggests that if I'd kept it open another 4 days, memory would continue to grow, but I felt this was enough time to demonstrate the issue, and I was getting annoyed by the warnings of missing plugins.

Revision history for this message
David (wizard-houseoffire) wrote :

This is the last screen shot, after I closed firefox.
While flash may have exacerbated the issue, in these last few shots I had completely removed the flash plugins from firefox. Flash is not the issue here.
It's entirely possible that I'm totally misinterpreting the output here, and if someone more knowledgeable than I wants to prove me wrong, I welcome it. All I know is that I have to restart my computer about every two weeks of normal use to reclaim memory, else it grinds to a halt and (on two occasions) crashes.

Revision history for this message
jack (benbenny) wrote :

Dear David, Please send NO More Bugs Reporting Mail. Thanks, Ben
 -------------- Original message ----------------------
From: David <email address hidden>
> I then removed the flash plugin, restarted, and brought firefox back up.
> The attached screen shot is almost 2 days of usage. Progression
> suggests that if I'd kept it open another 4 days, memory would continue
> to grow, but I felt this was enough time to demonstrate the issue, and I
> was getting annoyed by the warnings of missing plugins.
>
>
> ** Attachment added: "Firefox minus flash after 2 days"
> http://launchpadlibrarian.net/11579207/memory5.png
>
> --
> leaking pixmaps
> https://bugs.launchpad.net/bugs/177680
> You received this bug notification because you are a bug contact for
> firefox in ubuntu.

Revision history for this message
David (wizard-houseoffire) wrote :

My apologies. I was only trying to help, when it looked like no one was even looking at the problem. This will be the last message I send. I'll find some way of dealing with it.
Thank you.

Revision history for this message
Alexander Sack (asac) wrote :

On Sun, Jan 27, 2008 at 06:40:29AM -0000, jack wrote:
> Dear David, Please send NO More Bugs Reporting Mail. Thanks, Ben

Pleaes don't complain, unsubscribe on your own. visit the bug webpage
and you find an option to unsubscribe on the left.

 - Alexander

Daniel T Chen (crimsun)
Changed in flashplugin-nonfree:
status: New → Invalid
Changed in fglrx-installer:
status: Invalid → New
Revision history for this message
Bryce Harrington (bryce) wrote :

[No explanation why this was moved back to xorg was given; please see comments #5 and #7 which suggest the problem is in flashplugin-nonfree, and provide an explanation why those are incorrect before reassigning.]

Revision history for this message
Joel Goguen (jgoguen) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue for you. Can you try with the latest Ubuntu release? Thanks in advance.

Changed in flashplugin-nonfree:
status: New → Incomplete
Revision history for this message
Joel Goguen (jgoguen) wrote :

We'd like to figure out what's causing this bug for you, but we haven't heard back from you in a while. Could you please provide the requested information? Thanks!

Revision history for this message
David (wizard-houseoffire) wrote :

Sorry.
I've tried, but haven't been able to reproduce the error exactly as it was at the time. I no longer have the system that was affected, having long since replaced it with a newer one with a more recent video card and more memory.
The only thing I can confirm at the moment is that Firefox doesn't quite return to the same memory state when all tabs are closed, though it's not nearly as bad as it was when I originally posted this bug. I have suspicions (from further searches around the net) that this very last bit of memory that never seems to get returned is due to memory fragmentation, or such is my minimal understanding of the problem.
Nautilus still continues to grow with every thumbnail it displays. I'm not sure if the memory is ever reclaimed. I've got 8 times as much now, so I've not noticed any instabilities. I'll watch it over the next week and report back.

Revision history for this message
David (wizard-houseoffire) wrote :

I've attached my memory summary and interpretation at various points in time over the last week. It's entirely possible that I have no idea what I'm talking about, and would gladly be told such, but when I was running 7.04 with only 512 megs of ram, this behavior was enough to cripple my machine, requiring a weekly restart to reclaim the memory and get things back up to speed. I'm still seeing similar memory usage, but because I have so much more, it's not noticeable.
It's also possible that I'm seeing two different bugs here and am just ignorant enough to not know the difference.
Anything else I can provide that would help?

Revision history for this message
Joel Goguen (jgoguen) wrote :

Hi David,

Could you please attempt to produce this on another browser, such as Epiphany? Epiphany can be installed by installing the epiphany-browser package. If you are able to produce this, please provide output similar to what you provided for Firefox.

Thanks in advance!

Revision history for this message
Launchpad Janitor (janitor) wrote :

[Expired for flashplugin-nonfree (Ubuntu) because there has been no activity for 60 days.]

Changed in flashplugin-nonfree (Ubuntu):
status: Incomplete → Expired
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.