Firefox Takes the Next Step Towards Rolling Out Multi-Process To Everyone (arstechnica.com)

An anonymous reader quotes a report from Ars Technica: With Firefox 50, Mozilla has rolled out the first major piece of its new multi-process architecture. Edge, Internet Explorer, Chrome, and Safari all have a multiple process design that separates their rendering engine -- the part of the browser that reads and interprets HTML, CSS, and JavaScript -- from the browser frame. They do this for stability reasons (if the rendering process crashes, it doesn't kill the entire browser) and security reasons (the rendering process can be run in a low-privilege sandbox, so exploitable flaws in the rendering engine are harder to take advantage of). Moreover, these browsers can all create multiple rendering engine processes and use different processes for different tabs. This means that the scope of a crash is narrowed even further, typically to a single tab. Internet Explorer and Chrome both implemented this long ago, in 2009. Firefox, however, has not offered a similar design. Although work on a multi-process browser was started in 2009, under the codename Electrolysis, that work was suspended between 2011 and 2013 as priorities within the organization shifted. In response, Mozilla started switching to a new extension system in 2015 that opened the door to a multi-process design. The first stage of Firefox's move to multi-process involves separating the browser shell from a single rendering process that's used by every tab. In Firefox 48, that feature was enabled for a small number of users who used no extensions. Firefox 49 was rolled out to include users running a limited selection of extensions. Now, in Firefox 50, a separate renderer process is used for most users and most extensions. Developers are now able to mark their extensions as explicitly multi-process compatible. Firefox 51 will extend this even further to cover all extensions, except those that are explicitly marked as incompatible. Mozilla says that, even with the limited changes made in Firefox 50, responsiveness of the browser has improved by 400 percent due to the separation between the renderer and the browser shell. During page loads, responsiveness will increase to 700 percent.

Comments

Firefox...hmmm

[hyades1]

Isn't that the browser a lot of people used to use before the developers decided to wreck just about everything they liked in order to turn it into some kind of inferior imitation of Chrome?

Re:Firefox...hmmm

[mlts]

I wouldn't say splitting stuff into different processes is wrecking Firefox. In fact, the Edge browser will end up being split off into a completely separate Hyper-V VM (Windows Defender Application Guard.) The advantage of separating at this level is that malware has a layer between it and hardware, where it can't just flash bogus firmware to a keyboard because the keyboard input is virtualized, for example.

The more separation between the browser and the rest of the OS, the better. The ideal would be something like Qubes OS, with COW used to allow multiple browser instances that don't take up much RAM (although rowhammer protection will be needed.) We have firewalls to keep untrusted stuff from the Internet well away from internal machines; we should do the same with web browsers that are in constant contact with potentially malicious stuff.

Re:Firefox...hmmm

[Anonymous Coward]

And each update starts the infuriating process of finding a set of addons that give me my old browser back.

Re:Firefox...hmmm

[grcumb]

I wouldn't say splitting stuff into different processes is wrecking Firefox.

Every single grizzled, grey-haired veteran of the Browser Wars is putting on his best Comic Book Guy expression right now, and saying, "Multi-process vs multi-threaded, is it? You are discussing the virtues of a multi-process software application from a security perspective? How... interesting."[*]

--------------
[*] The decade-long turf-war between supporters of multi-process architectures and adherents of the multi-threaded heresy is legendary, for those who survived it. It consisted largely of a bunch of jaded old fucks warning each succeeding generation of new kids that multi-threaded sounds like a better optimisation, but if you just live with the extra few cycles required to spawn each new process, you get software that's way easier to manage. And successive generations of way-too-fucking-clever comp-sci grads yelling, 'Fine, gramps, but it won't matter how secure your apps are, 'cause no one's gonna buy them!!' and then rage-fragging a few more newbs before going back to coding their magnum opus. Then the SystemD singularity did for the Linux community what Trump did to the electorate, and all the old fucks quietly barred the door to their cabin in the woods, and cleaned their shotgun.[**]

[**] And then Netscap^H^H^H^H^H^HFirefox announces that multi-process is cool and more secure.

Re:Firefox...hmmm

[Motherfucking Shit]

It's coming back, albeit gradually. They're getting rid of bloatware like Hello, Pocket is easy to disable, and Classic Theme Restorer still works fine even though it shouldn't have to exist. Going from 47 to 50, memory usage has dropped for me. I still see some hiccups. Ghostery doesn't work properly for me in 50.x (its dialog is blank). According to the Add-on Compatibility Reporter, about half my addons are "Not compatible with multiprocess," so I won't turn that on. And something changed between 50.0.2 and 50.1 that causes brief freezes when I'm scrolling.

I at least feel like they're finally moving the project in a good direction, being a good browser instead of trying to be an entire OS.

Re:Firefox...hmmm

[JustAnotherOldGuy]

Isn't that the browser a lot of people used to use before the developers decided to wreck just about everything they liked in order to turn it into some kind of inferior imitation of Chrome?

Nicely put. Yes, that would be the one.

Re:Firefox...hmmm

[MSG]

No, it's the browser people use because they want to sync their bookmarks *and* have extensions on a mobile device.

Re: Firefox...hmmm

[Anonymous Coward]

Missing the point. The minimum requirements to use Chrome is 256MB RAM per tab. This includes things like nw.js (chromium webview) where as Firefox didn't do this and required 256MB for the entire browser.

The entire sand boxing , one-thread per tab, thing is killing the "mobile" experience, and destroying the desktop experience where things like Twitter and Slack require 64-bit processors because they are constantly increasing the size of the webview DOM with no Garbage collection. Like I do not know how anyone releasing a HTML5 app ever expects to have a usable universal experience when the goalposts are moving faster than the hardware revisions.

Re:Firefox...hmmm

[EzInKy]

What is wrong with "cp bookmarks.html"?

Re:Firefox...hmmm

[Anonymous Coward]

....[*] The decade-long turf-war between supporters of multi-process architectures and adherents of the multi-threaded heresy is legendary, for those who survived it. It consisted largely of a bunch of jaded old fucks warning each succeeding generation of new kids that multi-threaded sounds like a better optimisation, but if you just live with the extra few cycles required to spawn each new process, you get software that's way easier to manage..

Not quite gray haired here yet, as long as I shave at least, but anyway, I'd tend to argue that yes, full process separation works well if things are fairly separate. In this case you need it because the underlying web architecture has components that aren't particularly trustworthy so you need as much separation as possible, within reason, and it only gets worse from here as far as threats go.

For that matter I've got a piece of c# code that embeds an application in another application. It works poorly, but that could be fixed. (Putty embeds decently, as does powershell.) Realistically, for most browser type work, just running multiple instances does work. Windows/Linux/Whatever could put things in tab groups that you drag around. It wouldn't have to just be all firefox, or all chrome. It could be a group of tabs related to a task with firefox, chrome, excel, notepad++, and visual studio.

That being said, if your creating an app that needs higher performance that does have a lot of tightly interacting components, well shooting everything over network calls or similar is going to be expensive and will add a bit of latency. Sometimes that does not matter. Sometimes it does.

One area where efficiency does not usually matter is config time stuff. I used to come up with C/C++ APIs that got called in other languages, then kept changing the API's over and over again. Just use someone's json library. Pass in a really big json config string to C, then if you add a config option, it just passes right on through to where it is needed. I really wish I started on this way of doing things years ago. It would have saved considerable time and the memory used is still tiny. I do have a few areas where I pass through a binary blob for efficiency alongside, but that is more for things like raw data. This nice thing about this approach is by keeping the API mostly one big string, your close to being able to wrap a C++ dll and making it a generic service using separate processes.

Re:Firefox...hmmm

[hyades1]

I did the sucky thing by just sidestepping the whole problem...I switched to Pale Moon. They just quit supporting XP, though, so my oldest computer will have to move over to Opera.

Re:Firefox...hmmm

[hyades1]

That's what actually drove me away. I simply couldn't be bothered. Apparently a lot of people agreed, because the number of people using Firefox has crashed since they decided productivity-oriented people needed to spend time reconfiguring a basic internet tool every few weeks instead of just...using it.

Re:Firefox...hmmm

[fisted]

Netscap^H^H^H^H^H^HFirefox announces that [...]

What is this NFirefox? I hadn't heard about it yet. New firefox? Next firefox?

That said, am I the only one who does NOT want multi-process simply because it's annoying as fuck already that firefox hogs ONE core? I don't want it to hog EVERY core, for resource waste, noise and temperature reasons.

Re:Firefox...hmmm

[Anonymous Coward]

That said, am I the only one who does NOT want multi-process simply because it's annoying as fuck already that firefox hogs ONE core? I don't want it to hog EVERY core, for resource waste, noise and temperature reasons.

$ taskset -c 0 firefox

Re:Firefox...hmmm

[Anonymous Coward]

cp? That's soooooooo 1970!

Re:Firefox...hmmm

[Bongo]

Heh. I'll always remember Jerry Pournelle writing in Byte: "I'm not sure I want to share cycles even with myself" back when preemptive multi tasking was a strange new fangled thing in PCs.

Re:Firefox...hmmm

[XparXnoiaX]

Is systemd insecure? Oh yes, yes it is. It can be crashed in one tweet.

Re:Firefox...hmmm

[fisted]

$ taskset
-sh: taskset: command not found
$ ls -ld /usr/pkgsrc/*/*taskset*
ls: /usr/pkgsrc/*/*taskset*: No such file or directory
$ locate taskset
/mnt/huge/cmport/repo/cm/external/busybox/miscutils/taskset.c
/mnt/huge/cmport/repo/cm/external/busybox/testsuite/taskset.tests
$
 
:(

Re:Firefox...hmmm

[helga the viking]

Exactly. Multi-Process means nothing if its coded poorly. Insofar as saying multi-process equals extra security, speed, robust interace but without the right goals (and actually building it well) will just complicate matters. 30 years of SW engineering says this is an ongoing fad that keeps coming back.

Re:Firefox...hmmm

[Anonymous Coward]

I switched to Pale Moon. It's Firefox with less bloat.

Re:Firefox...hmmm

[Anonymous Coward]

In case you didn't know, they replaced bookmarks.html with an SQL database several versions ago. Doing so made Firefox even slower than it was (though, for some strange reason, it still doesn't run backwards), and none of the updates since has fixed the slowdowns that introduced.

Re:Firefox...hmmm

[QuietLagoon]

...Classic Theme Restorer still works fine even though it shouldn't have to exist....

"...even though it shouldn't have to exist...." --- this

.
So long as statements such as the one I quoted need to exist, Firefox is not coming back.

The developers still appear to think they know better, as they slowly turn Firefox into a Chrome clone.

Mozilla had the golden ring, but they dropped it.

Re:Firefox...hmmm

[Anonymous Coward]

And each update starts the infuriating process of finding a set of addons that give me my old browser back.

Which is why many of us went to the Pale Moon fork, haven't had a theme or extension break in a few years now.

Re:Firefox...hmmm

[Anonymous Coward]

Bug has nothing to do with 'tweets'.
Bug existed for a short time.
Bug was fixed quickly.

Just more pathetic anti-systemd trolling.

Re:Firefox...hmmm

[sremick]

Chrome? Isn't that the trojan that gets bundled with every unrelated download one finds on the internet, and then once it gets sneaks onto a user's computer and tricks the user into making it their default browser, arrogantly takes over the computer spawning a dozen background processes that exhaust all available CPU and RAM, bringing everything to a crawl?

I've been in IT for decades, and have been rather ambivalent about Firefox-vs-Chrome until just recently, mainly just being content as long as users weren't using IE. However, despite any influence from me, I've seen countless longtime Chrome fans abandoning it and coming back to Firefox because of the background process/CPU/RAM issue I mention above. This is also being seen on both Windows and Mac platforms... it's not specific to either OS. At my work, Firefox continues to be the deployed browser of choice to all the many thousands of computers we manage because Chrome is an ass and thinks you don't want to use your computer for anything but web browsing.

Re:Firefox...hmmm

[bad-badtz-maru]

On a project I work on, we have a few methods that accept XML as parameters. If you think about it long enough, you'll realize (for better or worse) that you've recreated the annoying part of Perl functions, the lack of a formalized parameter declaration in the function signature.

Re:Firefox...hmmm

[Anonymous Coward]

Have you tried automating it for syncing between multiple devices? A purely one-way sync isn't difficult to set up, but it is a bit more complicated when you are syncing 2 or more ways.

Re:Firefox...hmmm

[Anonymous Coward]

Say I bookmark 5 new sites/pages on my PC. Now you want me to cp the bookmarks.html to my tablet AND phone. when I don't even have a decent access to directory/file structure without 3rd party apps (like ES File Explorer). on top of that the OS isolates app data folders so that third party cannot read or write data in the folder. So I have to now root the device. I'd rather store that information in the cloud and sync automatically.

I should probably just ask, "what's wrong with wget?"

Re:Firefox...hmmm

[Anonymous Coward]

"And something changed between 50.0.2 and 50.1 that causes brief freezes when I'm scrolling."

I had this issue to - try turning off 'Use hardware acceleration when available' under advanced>browsing until they figure out the bug.

Re:Firefox...hmmm

[fbobraga]

Isn't that the browser a lot of people used to use before the developers decided to wreck just about everything they liked in order to turn it into some kind of inferior imitation of Chrome?

Yeap! (it's projected to be an PERFECT Chrome clone as of version 7517576215751257!)

Re:Firefox...hmmm

[m.dillon]

I wouldn't really characterize it as any sort of turf war... insofar as I know, there was never any turf war. Each has its uses.

In terms of security, don't forget user account separation. Why trust the application to secure the environment for you? I don't. I segregate browser use cases into separate dummy user accounts and start independent instances of the browser with a simple ssh localhost -l dummy1 -n chrome .... (or whatever, all bound up into a GUI button).

After all, the browser is only part of the problem. I want to be able to open up PDF documents (which runs xpdf), or other things (which might run open-office) from the browser. Even if I trusted the browser's security, I don't trust xpdf or open office. If I were really rabid, I'd segregate xpdf and OO execution into a sub-sub user account (but I'm not quite that rabid).

-Matt

Re:Firefox...hmmm

[postbigbang]

Firefox 50.1 has taken its trickle of rendering and tab cleanup to be measured in furlongs per fortnight. It has become horrifically slow, not the reverse using common add-ins.

Worse, some pages simply die or are rendered incorrectly if at all.

While I used Firefox loyally, especially for its debug qualities, I might actually be forced to change because it is stunningly slow even on fast hardware. It's like I leaped back to 1998.

Re:Firefox...hmmm

[Anonymous Coward]

Splitting it into separate processes is horrible for performance and resource usage. The result is something that is ultimately detrimental to most users.

That's why I stick with Pale Moon. None of this garbage that is promoted as a good thing just because it's new.

Re:Firefox...hmmm

[Anonymous Coward]

If you say so... I tried using FF on android for the purpose of using NoScript or uMatrix on mobile, but FF wouldn't allow me to install them on mobile...

Re: Firefox...hmmm

[Anonymous Coward]

2 or more ways? You can sync *to* something, and you can sync *from* something. Unless you live in some strange 4-D world those are your only choices.

Re:Firefox...hmmm

[Anonymous Coward]

Even when the background processes weren't resource hogs, they were still very sleazy about resisting removal. On Macs, Google's "updater" would be installed in several places and would repopulate all of these locations if any of them were removed, malware-style.

Re:Firefox...hmmm

[Anonymous Coward]

I would like to sync my bookmarks with Firefox, but now that they require me to establish identity (email address) and make my browsers trackable in order to use Sync, I'd rather not.

Re:Firefox...hmmm

[Motherfucking Shit]

Thanks, that didn't seem to help matters any, but I upgraded to 51.0b10 tonight and the problem went away. Switching to the beta turned on multiprocess, and the only incompatible add-on was Fireshot, so I'm leaving multiprocess enabled. So far so good.

How many DNS queries can it launch

[Anonymous Coward]

The main question is how many DNS queries can it launch at once to pull in all the crazy bits that advertisers use. This goes back to the "how fast can it render" question. Well first it needs to pull in all the stuff and badly written web code, especially advertisements, pull in stuff from all over the place that often require multiple fetches from a myriad of different sites to get all the elements needed to render a page,

((browsers should give each page a score for how easy it was to assemble))

Re:How many DNS queries can it launch

[Anonymous Coward]

responsiveness of the browser has improved by 400 percent due to the separation between the renderer and the browser shell.

Unfortunately, the amount of crap on web pages has increased by 500 percent.

Re:How many DNS queries can it launch

[0100010001010011]

Unbound or PiHole and send those to a nginx server that returns a 204.

Re:How many DNS queries can it launch

[Anonymous Coward]

Here's a real HCF page that will burn your computer to the ground. It plays nice in Edge but Firefox completely grinds to a Halt.

https://learn.adafruit.com/adafruit-neopixel-uberguide/arduino-library-installation

Re:How many DNS queries can it launch

[Immerman]

I'd love to see the browser do an end-run around all that - I wouldn't think it would be too hard for at least 80% of cases: just load everything hosted on the same server / subnet as the page itself first. That should quickly give you enough information to render the page reasonably well unless the page designer specifically attempted to avoid it, and the rest can be loaded in at a more leisurely pace. That just leaves the challenge of avoiding constantly reflowing the page as new components load in. Worst case you could just do two reflows: once after the subnet content is all loaded, and again after everything else is done.

Re: How many DNS queries can it launch

[Anonymous Coward]

Works fine in iOS Safari. Zerohedge can crash that sometimesâ"ZH is a real torture test of bad code and excessive ads/trackers.

Re:How many DNS queries can it launch

[DonaId Trump]

We're going to fix that problem in my first 100 days in office, believe me. DNS is low energy and high latency, SAD! Did you know there are DNS servers in Mexico and CHY-NA, folks? I'm going to make it ILLEGAL and force every link on the cyber to go through my new URL shortening service, HTTP://BIG.LY. I'm appointing Barron to run it so you know it's going to be the best!

Re:How many DNS queries can it launch

[Anonymous Coward]

I have to say that was kind of awesome

Re:How many DNS queries can it launch

[Anonymous Coward]

you mean HTTP://YU.GE.

Re:How many DNS queries can it launch

[Anonymous Coward]

Page opens fine, and immediately, for me in Firefox 50.1.0 on Linux Mint 17.3 with another 20 tabs open too, uBlock Origin, Ghostery, Privacy Badger and NoScript extensions - perhaps the problem is more with Windows - if its Windows you're using? (on i3, 4Gb Asus laptop)

Re:How many DNS queries can it launch

[Anonymous Coward]

Isn't that dependent of the max number of TCP connections? Just adjust the thing until the servers puke blood on your client.

Re:How many DNS queries can it launch

[Anonymous Coward]

Opens fine in Firefox 50 on Windows 7.

Re:How many DNS queries can it launch

[Merk42]

There's also the issue of CDNs.

Re:How many DNS queries can it launch

[Yvan256]

No problem at all in Safari on an old Core 2 Duo Mac mini.

Re:How many DNS queries can it launch

[Immerman]

Is there really though?

If embedded content is identified by relative address, or an address on the same server name, it's logically on the same server, regardless of where it's physically located. That much should be trivial to load first.

Of course once you've finished loading all the "local" content, then if ads are served from the same CDN, that would make them harder to prescreen by subnet for delayed loading, but by the same token they should be served much more quickly as well, so it's probably not an issue. And if the ads then link to content outside the CDN, so what? Those components will then get screened for delayed loading.

And while I'm not well versed in CDN details, it seems likely that the same physical server will end up handling all the requests for "local" content on a given web page, or at the very least it will look as though that's the case from the browser's perspective. Pre-screening could also be done on a domain name based proximity as well, which would prioritize content from "legitimate" support servers: e.g. when loading a page on www.domain.com, you'd consider content from database.domain.com as "almost local", likewise from x4m3.contentserver.domain.com.

Re:How many DNS queries can it launch

[Merk42]

Is there really though?

...if ads are served from the same CDN, that would make them harder to prescreen by subnet for delayed loading...

That's what I'm saying. The content from the same server is one thing.
What if there is some CDN, (e.g. Akamai), that is used for both serving images AND ads? How would the browser know the difference if it's only looking at the server?

Re:How many DNS queries can it launch

[Immerman]

Why would it need to? The point is not to avoid loading ads, but to get the important content loaded as quickly as possible so that the page can be rendered nicely, and ad networks often have ugly delays that interfere with that goal. If the ads and images are served by the same CDN, then most likely either both will be served promptly, or both will be delayed. Either way there's limited advantage to distinguishing between the two.

I also suspect that that it's common practice to link images by relative address (i.e "/images/logo.png"), and likewise that it's common for ads to be addressed by absolute address ("www.someserver.com/x/y/z/images/logo.png"), making it trivial to distinguish between the two. We're not looking for a 100% solution here, just a good heuristic that will tend to speed up loading a page's important content.

Re:How many DNS queries can it launch

[plover]

Nope, this was certainly not a Windows problem. I'm running the same suite of extensions that you are, on Firefox 50.1.0, on Windows 10 (on a 3 year old tablet with only 4GB RAM and over OpenVPN, no less.) The page loaded instantly for me. I had no problem scrolling to the bottom and back to the top.

Of course now my battery is dead... :-)

Re:How many DNS queries can it launch

[m.dillon]

Kinda of a funny question. Don't people realize that they can just run a caching DNS service on their workstation or in-home server? I do! It takes care of the problem just like that. No way I would *ever* trust my ISP's DNS service. Or anyone else's for that matter.

-Matt

Re: How many DNS queries can it launch

[Anonymous Coward]

Hillary didn't trust her ISP's servers, and look where it got her.

Stopping processes of background tabs?

[Midnight Thunder]

One of the issues today is sleeping or stopping of the processes of tabs that haven't been used in a while. I don't believe Chrome does this and would suspect it is one of the causes of making it heavy in terms of resources?

Re:Stopping processes of background tabs?

[Anonymous Coward]

It's actually something they're investigating for Firefox right now, and one of their focuses for the whole Project Quantum thing seems to be to try to reduce the amount of CPU usage that background tabs get (unless it's somehow clear that they're supposed to be working in the background without being slowed down too much). Hard to say what the results will be at this early stage, but there's a lot of interesting stuff that seems possible now that they've rewritten most of the browser to support e10s.

Re:Stopping processes of background tabs?

[0ld_d0g]

How do you define a background tab? I (and I suspect many others) have spotify or youtube or netflix or some other media playing in the b/g. Allocating CPU cycles to threads and processes is the job of the OS, not the browser.

Re:Stopping processes of background tabs?

[Midnight Thunder]

In this case, the browser design is likely to provide a hint to how the process are being used. The OS doesn't have that contextual information, so for it is just another process. While not perfect, there are likely to be hints that can be used. For example a browser that just relaunched with 30 tabs open (I have seen that), may want to default to only spawning processes, representing tab that have been visited by the user in the current session.

Good job

[Anonymous Coward]

The light at the end of the tunnel is only the light of an oncoming train

Surprisingly my plugin still works

[rsilvergun]

last I heard they were going to break it with version 40 sometime in August. There's a ton of work that'll have to be done to make it work properly in a multi-process environment. More or less a full re-write. To say nothing of having to redo all my UI components. Here's hoping they get cross compatibility with Chrome so I can at least get some code reuse out of it.

Fix the popup blocker

[Anonymous Coward]

I'd say the most important thing they need to fix is the popup blocker. Whether one tab crashes or they all crash in one go, isn't really a problem I'm bothered with. What I am bothered with is the way I still get popups.

Re:Fix the popup blocker

[tepples]

Firefox's pop-up blocker requires pop-ups to be initiated by a user event, in particular a click. If you don't want pop-ups, then don't click on the page.

Not all rosy

[markdavis]

Unfortunately there are some users who are going to be greatly adversely affected by Firefox moving to full multi-process. I am one of those. At work we make use of Firefox under Linux thin client environments, running on a centralized server. This type of change opens the door for Firefox to completely overrun all our CPU resources, no matter what we throw on it. Why? Because I am certain the Firefox developers won't give us an "off" switch for this new functionality (or if they do, it will be taken away eventually). See, if you are running on your own dedicated computer, it might be perfectly OK to unleash the browser to take over all your resources. But when it is a shared environment on a multiuser system, this can be horrible. Believe it or not, we do lots and lots of other things on that system at the same time. There aren't enough OS tricks available to deal with it. "Nice" just isn't enough (for a variety of reasons... none of which I have the time to explain right now). We are already having problems just trying to limit memory usage in Firefox (it would be the same with any browser) because ulimits don't apply correctly with certain types of threaded applications, just causing either no effect or crashing; and Firefox doesn't have good settings to deal with that either).

Yes, I know what we do is a little unusual and old-fashioned, but it actually works pretty good and has a lot of advantages. I certainly don't want to hold Firefox back, but I do wish that the developers would consider the fringe cases and give us persistent options to prevent Firefox from taking over more than X number of cores. Choice, options, and settings are key, that way, everyone could "win".

Re:Not all rosy

[Anonymous Coward]

What's nice is that Mozilla does not plan to use processes without discretion. All the discussions I've read so far imply that they want to only use as many as makes sense for the number of cores on your system (they also plan to continue using threads as much as possible). This implies that you'll be able to control how many processes Firefox tries to use, even if it's more than it was before e10s. So really it's quite early to sound the doom sirens.

Re:Not all rosy

[trparky]

You can control how many sub-processes are created by the browser using an entry at the about:config screen. If you set it to one it will only spawn one sub-process. This will of course negate the whole idea of having sub-processes since all tabs would be running in one rendering sub-process but it would fix your issue.

Re:Not all rosy

[Anonymous Coward]

"Firefox to completely overrun all our CPU resources" -- IMHO, all browsers have been doing that since before the dinosaurs. It not a new problem. Deal with it.

Re:Not all rosy

[markdavis]

>"All the discussions I've read so far imply that they want to only use as many as makes sense for the number of cores on your system (they also plan to continue using threads as much as possible). This implies that you'll be able to control how many processes Firefox tries to use, even if it's more than it was before e10s."

What scares me is that it is what THEY think it appropriate for the system, depending on what THEY think is the use case. They simply can't know that and will make assumptions that could be wrong. As long as they give the admin some control through settings, it will probably be OK. One of the most powerful would be something that could could simply limit ALL use to no more than X cores.... so setting it to "1" would allow old behavior, when necessary, "2" could expand it, "0" could be automatic of some sort.

>" So really it's quite early to sound the doom sirens."

It is never too early to be concerned. Lots of things have happened in Firefox that made life miserable for me in our unusual environment in the past. I don't see that as changing. As my Mom says "planning for the worst but hoping for the best".

Re:Not all rosy

[Anonymous Coward]

On a multiple processor system, a single process can't eat up all the CPU time.

Re:Not all rosy

[Anonymous Coward]

Mozilla has made it crystal clear for many years now that they simply do not want companies to use their browser, do not want their browser in shared environments, and intend to provide 100% capability for users to make any mistake they want and utilize the system it is running on as if it was theirs including when that is demonstrably not the case.

There is no sense complaining about this now. From the sounds of your setup, there is no possible way you haven't seen this for yourself already.

Re:Not all rosy

[tepples]

I've been using Firefox 51 from the firefox-next PPA in Xubuntu on a 1-core, 2-thread Atom N450 laptop for a couple weeks now. Even if all tabs run in one process, scrolling still runs smoothly because it's in a separate process, though occasionally I outrun the rendered portion and there's a delay before the blank part fills in.

Re:Not all rosy

[subreality]

I suggest you look at cgroups. Instead of relying on the processes to play nice with resources, you can specify the resources you want allocated to each user. For instance, if each user has cpu.shares=1024, then it'll fully balance - if user A starts an old firefox and it's running singlethreaded, and user B starts a new firefox and it spawns 50 processes, you'll see user A's process consuming 50% of CPU, and user B's flock consuming 1% CPU each.

In this way you specify what you want to achieve (user B doesn't steal all the CPU making life suck for user A), instead of how to achieve it (singlethreaded software, and relying on users to not run more than one copy). It's easier and more efficient.

The other cool part: this may already be set up for you. On my Ubuntu 14.04 system this is all done by default when I log in.

Re:Not all rosy

[Anonymous Coward]

It can. I think you are confusing the idea of a "process" with a "processor".

Re:Not all rosy

[Anonymous Coward]

Surely there are some interfaces for you to write a little C-program or a script to set the process affinity, priority level and maybe even set limits for the created process in Linux too?

Re:Not all rosy

[munch117]

I'm curious, in what way does turning threads into processes overrun CPU resources? Threads can peg the CPU meter on multiple cores just as easily as processes can. Is it a page table thing or an OS scheduler thing or what?

It seems to me that you have an operating-system level problem. If nice and ulimit and the OS scheduler aren't cutting it, then they need to be improved or replaced.

Re:Not all rosy

[Anonymous Coward]

Vote parent up, way up. UNIX has its roots as a fine-grained resource accounting and limiting system (memory and CPU time used to be a thing you could charge a lot for!). Linux inherits that tradition and has layered stuff like cgroups on top of it to make it easy to manage.

I mean -honestly-, how did GPP think decent multi-tenant VPSs hosts handled resource constraints between competing clients? Cron jobs that pinned VMs to specific cores and killed "runaway" processes?

Re:Not all rosy

[Billly Gates]

Oh please can the hardware run IE 8 and Chrome 1.0 circa 2009?

If yes then you're fine. Firefox is now catching up to IE 8 security and performance which is why I left it

Re:Not all rosy

[Billly Gates]

It most certainly is appropriate. Can your mom run IE 8 from 2009? If so then it can handle Firefox. No you did not misread that?! Firefox is so far behind and all everyone does is whine about change when they try to do catch-up.

Now Firefox can run in lowrights mode in %apodata% in your mom's profile and have this new cutting edge technology called a sandbox like what every other browser had since last decade.

Re: Not all rosy

[Anonymous Coward]

about:config

dom.ipc.processCount

Does what you're asking for

Re:Not all rosy

[AmiMoJo]

They can't win. If they don't do it, Firefox's performance and security will continue to fall behind and the vast majority of users will complain. If they do it, a small number of users with unusual use-cases will complain that it breaks their system. If they try to accommodate both it will create much more work and delay the feature.

Oblig. XKCD.

Re:Not all rosy

[markdavis]

I used affinity to try and tame/control Java, since it has the same problem. It worked, but using affinity, you can only lock a process to a particular cpu core... you can't limit it to just 1 of any core. And this leads to the system being unbalanced. The best I could do was a script that chose a core at random.

Re:Not all rosy

[markdavis]

>"Oh please can the hardware run IE 8 and Chrome 1.0 circa 2009? If yes then you're fine. Firefox is now catching up to IE 8 security and performance which is why I left it"

That is ridiculous.

1) Running IE under Linux is not realistic.
2) Saying Firefox security and performance is at 2009 levels is just completely false.
3) Implying that I could use Chrome to do what Firefox does is equally false... especially Chrome 1.

Re: Not all rosy

[markdavis]

>"dom.ipc.processCount Does what you're asking for"

yes it does... for now. Until they disable or remove it. They have done that before with other things we rely on, often without real discussion and sometimes with no warning. That is what worries me.

Re: Not all rosy

[Anonymous Coward]

So now you're whining about an imaginary problem you MIGHT have sometime in the future?

Even first-world problems are worse than that.

Another update?

[Anonymous Coward]

PLEASE. I am tired of the minutely update that happens without my permission or consent and forces me to restart my browser once every TEN SECONDS. Hate Chrome. Hate Firesux. Any GOOD ones out there?

Re:Another update?

[Anonymous Coward]

> minutely update that happens without my permission or consent

Firefox -> Options -> Advanced -> Update

Choose either option 2 or 3 if you don't want updates

Re:Another update?

[Anonymous Coward]

Exactly.

Or use the Extended version and update just once a year.

PEBKAC

Re:Another update?

[Anonymous Coward]

Pale Moon.

Haiku

[Anonymous Coward]

I don't take that software project seriously anymore. They have no real multiplatform strategy. If they would be serious about multi task they would code the main branch for haiku.

Re:Haiku

[Anonymous Coward]

I don't take that software project seriously anymore. They have no real multiplatform strategy. If they would be serious about multi task they would code the main branch for haiku.

That's not a Haiku
Even though you have three lines
Syllables are wrong



Before pedantry
I know they don't need to be
It was just a joke

Re: Haiku

[Anonymous Coward]

The transgender boy
Stared in utter confusion
At choice of two loos

Depends on if you are running a windows or linux

[Anonymous Coward]

backend, but it should be possible to set per-user limits on memory, cpu time, i/o access and other features to help reduce the chances of that happening.

Furthermore firefox SHOULD still have options in about:config to limit the number of open tabs alowed which will help you keep both memory and cpu utilization under control.

Stop being lazy with your administrative duties and it should be plenty possible to make even the latest version of firefox deal with the contraints you need it set against.

no problems with ff here

[Anonymous Coward]

we run it at work on about 60 machines, with 3-4 add-ons (adblock, https everywhere, etc.). seems to work just fine. thanks. edge is stoopid. chrome is opaque.

Really ... FFox separates wheat from chaff ??

[Anonymous Coward]

then howcomes every-other *.midi download I try crashes my FFox 50 browser under Ubuntu-16.04 ?? Mebby they mean REAL wheat kernels and real M$ chaff ! But, THAT could get separated by an 1818 ginmill !!

Re:Depends on if you are running a windows or linu

[markdavis]

>"Furthermore firefox SHOULD still have options in about:config to limit the number of open tabs alowed which will help you keep both memory and cpu utilization under control."

I expect it does and will for a while, but it is my experience they like to deprecate such settings after a while, and often with no notice or documentation.

>"Stop being lazy with your administrative duties and it should be plenty possible to make even the latest version of firefox deal with the contraints you need it set against."

Please don't presume to know what my duties are, how much time I have spent on it, or what we have had to deal with in the past. Thanks

Surf the Web with lightning speed

[pigsycyberbully]

http://www.opera.com/download/...

Surf the Web with lightning speed.
Presto/2.12.388 Version/12.16

https://duckduckgo.com/

Search Engines in all countries in the world
http://searchenginesindex.com/

Re:Surf the Web with lightning speed

[EzInKy]

You are seriously suggesting the Chinese State browser, are you?

Re:Surf the Web with lightning speed

[sirber]

Opera 12.16 doesn't work well for most recent websites.

Re:Surf the Web with lightning speed

[Anonymous Coward]

No. He's suggesting the Norwegian Opera 12.

That's nice

[JustAnotherOldGuy]

"With Firefox 50, Mozilla has rolled out the first major piece of its new multi-process architecture."

That's nice, but I'd rather you jackasses just fix yer fuckin' memory leaks.

Pro Tip: Stop "improving" shit until it turns into a steaming pile of donkey shit. Oh, wait, we passed that point back around Firefox 30 or so.

Re:That's nice

[Anonymous Coward]

> That's nice, but I'd rather you jackasses just fix yer fuckin' memory leaks.

Which leaks are you referring to?
Specifically, which bug numbers.
Oh, that's right. You are just virtue-signaling to all the other bitchy whiners but haven't lifted a finger to help yourself.
Go ahead, use chrome. Or one of those bug-ridden security nightmares of a barely maintained years old firefox fork.
Ain't nothing stopping you.

Re:That's nice

[Anonymous Coward]

That's nice, but I'd rather you jackasses just fix yer fuckin' memory leaks.

I'd rather you shut the fuck up as long as other browsers are far worse. Like Chrome. Friggin' memory hog.

Pro Tip: Stop "improving" shit until it turns into a steaming pile of donkey shit. Oh, wait, we passed that point back around Firefox 30 or so.

Pro Tip: Shut the fuck up and make yourself look like slightly less of a fool.

Re:That's nice

[AmiMoJo]

Firefox doesn't leak memory these days. Some add-ons do, but that's the add-on developer's fault. Part of the trade-off for having powerful add-ons with deep ties into the browser is the risk of memory leaks and security issues. Chrome is much more restrictive and aggressively manages add-ons, but even there you can get memory leaks with Javascript heavy sites like GMail.

Re:That's nice

[Anonymous Coward]

I run Firefox without add-ons and it does leak.
I have to restart it every week or so because it apparently exhausts its address space and starts to fail in weird ways.
I've seen it use 1.3 GB of private memory when there was only a single, blank tab open.

Re:That's nice

[l20502]

Shut the fuck up, they've been designing electrolysis to use a sane number of processes in order to reduce memory usage, do you know any other modern browser doing that?

Re:That's nice

[sirber]

Ain't seen memory leaks for years... are you with us in 2016?

Re:That's nice

[Anonymous Coward]

Firefox doesn't leak memory these days.

Sounds good, except we've heard that ever since Firefox 2.2, and every time it was a case of the poster "forgetting" to qualify it with "for me".

Personally I solved the "Firefox leaks memory" problem by putting more memory in my PC than Firefox could address. Unfortunately, now people are telling me to switch to 64-bit, and doing the same on 64-bit costs about as much as a large hadron collider.

Re:That's nice

[PvtVoid]

That's nice, but I'd rather you jackasses just fix yer fuckin' memory leaks.

Pro Tip: Stop "improving" shit until it turns into a steaming pile of donkey shit.

Dude, maintenance is boring.

Re:That's nice

[JustAnotherOldGuy]

Firefox doesn't leak memory these days.

WRONG.

Re:That's nice

[JustAnotherOldGuy]

Shut the fuck up, they've been designing electrolysis to use a sane number of processes in order to reduce memory usage, do you know any other modern browser doing that?

Shut the fuck up, I don't care. You're the kind of dick who brags about his car's awesome paint job when it's been up on cinder blocks for the last 5 years.

Re:That's nice

[JustAnotherOldGuy]

Ain't seen memory leaks for years... are you with us in 2016?

Obviously because you don't personally see them, they don't exist.

Re:That's nice

[Waccoon]

Based on my tests with both Pale Moon and Firefox, the problem isn't memory leaks, but the fact the caching policy doesn't want to release the memory at all. Firefox caches WAY too much damn stuff for no good reason. Eventually the memory can be released, but you have to do all kinds of stupid tricks, like use the back and forth buttons a few dozen times and load lots of blank pages to clear out the crap.

I'm desperately trying to find out why Pale Moon 26.5 worked terrific usually hovering around 300MB for hours, and now version 27.0 has the same awful memory behavior as Firefox, sucking down 1.6GB of memory within 5 minutes and staying there forever. I'm pretty sure it's a configuration/policy issue, and not an engine bug. I'm also pretty sure it's not a user-configurable setting, though, so it's going to be real fun to track down.

Re:That's nice

[JustAnotherOldGuy]

FWIW, it seems that the latest FF build is more capable at memory management. I've only had it up about 12 ~ 15 hours and it's hovering around 1.4Gb. We'll see where it is after a day or two.

If it stays there and doesn't keep climbing then I'll be very happy...I've been dealing with FF memory issues for the last 20 releases or so.

Re:That's nice

[Anonymous Coward]

It got better, but still leaks resources all over the place.

The fastest way to get it to "leak" on my laptop is to use a site like google images for a while. It seems to keep a lot of images cached, even after closing all tabs again and leave a blank tab open. Even the garbage collection in about:memory doesn't seem to free it.

Even faster is to use a site with a lot of embedded video. After a while, this causes FF to run out of image buffer or something, causing it to "freeze" and render portions of the screen black, until it unfreezes again. Keep using it after that, and it will invariably crash.

And don't get me started about their recent change in the video tag behavior where they just start loading the whole (often 1GB) video because the tag attribute said so (even when the video is not playing). This very much should be decided by the user, instead of the fsking web designer.

So, yeah, I'm also one of the people that'd rather see them fix that stuff first. Still, as long as I can hack it to do what I want, I prefer it over Chrome or IE.

Re:That's nice

[Waccoon]

Yeah, the trouble is that when the browser gets to 800MB the pauses become really noticeable, and at full tilt (on my system that's 1.6GB), the browsing is really annoying. I'm on the latest version, BTW. I have FF and Pale Moon installed side-by-side, along with other browsers.

Vulgar as it may be, I'm sad to see your comment has been modded as flamebait. It's a genuine problem that FF fans have been downplaying for way too many years.

Re:That's nice

[JustAnotherOldGuy]

Vulgar as it may be, I'm sad to see your comment has been modded as flamebait. It's a genuine problem that FF fans have been downplaying for way too many years.

It would have been modded down even if I'd said "pretty please". It's a sign of the downvoter's irrational defense and denial of a genuine problem.

Some of the numpty's here deny it because they haven't personally seen it, so "obviously" the problem must not exist.

I think their downvoting just proves my point, otherwise why would they care enough to waste a mod point? :)

Toyyoda

[tepples]

So if Trump wants us to buy American and hire American, and Toyota is Japanese, then why is the featured product on big.ly's front page a toy Yoda? I guess even though Toyota is Japanese, the company hires Americans to build vehicles right here in America.

Re:Toyyoda

[Anonymous Coward]

Toyota is as American as GM is Chinese. Final assembly is a minority of the overall jobs within an automaker, and placement of those final assembly jobs doesn't determine the company's nationality.

Jeb is a mess

[Anonymous Coward]

The amount of time (years) it is taking them to implement this feature demonstrates how horrid and agony-inducing its code is.

Donations

[gringer]

Hey, maybe we should give Mozilla more money, so that they can spend it on community education projects and saving hungry children instead of fixing their browser.

Re:Donations

[Anonymous Coward]

They will spend the money on renaming "Master/Slave" and other "offensive" terminology.

Re:Depends on if you are running a windows or linu

[Anonymous Coward]

> Please don't presume to know what my duties are, how much time I have spent on it, or what we have had to deal with in the past. Thanks

You are right. All we can know is what you've posted. And what you've posted gives every indication that either you are not fit for the job, or you are running firefox on some obscure platform that is not fit for the job. That's because no modern OS lacks the ability to control processes the way you say you want to control processes.

If you prefer not to be judged by what you post, there is an easy solution. Don't post.

Nightly user here.

[AbRASiON]

It's still atrociously slow, atrociously. Infact there's a either a bug with one of my plugins or something but I've deliberately endured it for nearly 8 weeks, hoping my feedback data is going in the bloody logs somewhere to help.

I run about 6 or 7 plugins which are all fairly common and I have run the same ones for 3 to 5 years. They worked fine on my last profile rebuild but this most recent one was broken /out of the gate/

It's so bad i'm using chrome primarily. Firefox performance frustrates me absoloutely no end and I've been running e10s, 64bit nightlys now for over 6 months.
Infact, except for the last 4 weeks, the ironic (?) / odd thing is, nightly has been RIDICULOUSLY stable for a piece of nightly code. It's just plain stable. It's just slow as heck.
I don't know what idiot legacy code is in Firefox but they seriously need a scrum or party or a bonus or something and the whole fucking team *STOPS* for 3 weeks solid and focuses on NOTHING but performance, performance, performance.
If you check my slashdot post history, I've been whining about Firefox performance for over 3 years.

It's ridiculous, stop adding features, just stop. Make the damn thing fast. Because Chrome destroys firefox and I'm now running a HEAP of Chrome addons (which I never knew existed) to replicate my Firefox functionality and I'm very, very close to full replication now (Firefox: Tab Mix Plus is bloody hard to replicate, requires multiple poor addons to be configured right)

Seriously, just stop whatever is going on Mozilla devs and bring in "The Carmack" or the equivalent at coding. Scrap something old, I don't care but something is fundamentally wrong when a 24GB memory, quad core machine can run so poorly with a browser.

Re:Nightly user here.

[narcc]

they seriously need a scrum or party or a bonus or something and the whole fucking team *STOPS* for 3 weeks solid and focuses on NOTHING but performance, performance, performance.

Take a look at Servo, Stylo, and Quantum. Here's a longer talk that goes over some of the you can expect.

Re:Nightly user here.

[yuvcifjt]

I run about 6 or 7 plugins which are all fairly common and I have run the same ones for 3 to 5 years.

Plugins or addons?

See if you have perf issues running firefox in "safe mode"...
https://support.mozilla.org/en...

Otherwise, try a full browser refresh, which will reset it to "factory defaults".

I'm on a really old Athlon X2 cpu with less than 4gig of ram (WinXP), and firefox runs beautifully...
But what I tend to do is unusual, I run a bunch of different sites under different Firefox profiles.
So mail would be under a separate profile (and process), and Google sites (like the memory hogging Google Maps) would be under a completely different profile (and process), and another profile for shopping sites, and default for all other browsing.

command line: firefox.exe -profilemanager -no-remote

But to start firefox under default, remove the "no-remote" flag.

Re:Nightly user here.

[Anonymous Coward]

I don't know what idiot legacy code is in Firefox but they seriously need a scrum or party or a bonus or something and the whole fucking team *STOPS* for 3 weeks solid and focuses on NOTHING but performance, performance, performance.

Don't give them any ideas. It's a known fact that the only two things they ever do is remove useful stuff and add broken stuff. And of those, the only one that improves performance is removing stuff.

After those three weeks they would announce that they've found two pieces with performance problems - the UI and the rendering engine - and the next version will remove both of them.

Re:Nightly user here.

[Anonymous Coward]

See how your Nightly interacts with your anti-virus solution. I'm not able to use the 32-bit version simply because of the constant core-hogging virus checking after every single operation, while the 64-bit version works like an angel in comparison as the anti-virus software doesn't interfere similarly.

Noticed this - 250% CPU!

[GNious]

Noticed this - in the past, the browser would idle at 80-115% CPU usage, now it see it rest at 180-250% CPU usage.

Giving it the Restart command also seems faster - shutting down and coming back up is now less than a minute most times!
(hey, it frees up 4-7 gigabytes of physical ram, and several gigabytes of diskspace, and makes pretty much everything run faster for a while)

Re:Noticed this - 250% CPU!

[Anonymous Coward]

My cpu load is no where near 250%, but firefox is a freaking RAM destroyer for me. Currently 6.4GB virtual and 4.5GB resident! Thankfully this little laptop has 16GB ram.

Re:Firefox...hmmm - can someone explain

[Anonymous Coward]

Rather worryingly I can't understand what @grcumb is saying. Isn't putting stuff into separate processes an extra layer in adding security? And a major one? If he's a point then I misunderstand something fundamental.
I'd appreciate any explanation, thanks.

Multi-Window, Multi-Tab

[CanEHdian]

I use Firefox in several windows, each having several tabs. When I need to do a restart, all I do is kill firefox.exe, do stuff & restart, start firefox and all windows and tabs come back after I click restore. Now that's useful. Great things never last long, I guess.

Re:Multi-Window, Multi-Tab

[Anonymous Coward]

(U) File -> Exit. Closes all your windows cleanly (killing the process is riskier, though Firefox seems to be fairly resilient against it nowadays), and will preserve the session for next launch, exactly the way you like it to.

Hosts avoid excess DNS via hardcodes

[Anonymous Coward]

See subject: By placing your fav. sites @ top of hosts properly reverse DNS resolved it resolves faster vs. calling remote DNS!

* Easiest & best custom hosts file creator (only 1 that does what's noted above) APK Hosts File Engine 9.0++ SR-4 32/64-bit https://www.google.com/search?hl=en&source=hp&biw=&bih=&q=%22APK+Hosts+File+Engine%22+and+%22start64%22&btnG=Google+Search&gbv=1/

(Best part is the other 1/2 of what it does is block ads!)

APK

P.S.=> Safe & 100% FREE https://www.virustotal.com/en/file/e01211ca36aa02e923f20adee0a3c4f5d5187dc65bdf1c997b3da3c2b0745425/analysis/1433430542/ Verified by Malwarebytes' S. Burn "seen the code & it's safe" http://forum.hosts-file.net/viewtopic.php?f=5&t=4290/

One word: security

[Anonymous Coward]

The reason why we have this is because the attacks against web browsers are getting more and more sophisticated. Combine that with the "wink, wink, nudge, nudge" relationship between ad sites and criminal organizations distributing malicious software, it isn't a matter of design that the browser has to be separated from everything; it is a matter of necessity.

SystemD Haiku - Fixed

[Anonymous Coward]

Bug nothing to do with 'tweets'.
Bug exist short time.
Bug was fixed very quickly.

Listening to Community

[bill_mcgonigle]

I make no pretensions that the current work isn't mere Chrome-catch-up, but if they had listened to community in 2007 they might not have had to catch up. In the days of $350M revenue from Google, "patches welcome" was an obsolete response - management simply failed to properly manage the engineering process.

from #392073:

L A Walsh 2007-08-13 11:41:59 PDT
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.6) Gecko/20070725 Firefox/2.0.0.6
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.6) Gecko/20070725 Firefox/2.0.0.6

This isn't a new problem, but has come to stand out as a fairly major "booboo" as Firefox has more windows open and has become slower and less responsive.
But firefox seems to be limited to 1 Core on a multi-way machine. I have a 2Px2Core setup (4 cores). Firefox's threading is broken in that it is constrained to 1 Core.
Given that cores are increasing in number but CPU GHz is not, this is a "major bug". While it should be a goal, overall to decrease all resource usages, when it becomes necessary to use more resources, FF should be able to take advantage of *parallel* processing (running on more than one processor at a time).

It looks like they'll have this rolled out right about the ten-year mark.

Re:Firefox...hmmm - can someone explain

[justthinkit]

You are responsible for a pack of pitbulls, and are equipped with Milkbone and "Here doggie, doggie, doggie". Then someone has the idea of multiple people in charge of a single pitbull, each person equipped with Milkbone and "Here doggie, doggie, doggie". Should anyone feel safer?

Firefox bad!

[Merk42]

I didn't read TFS but since this is Slashdot and the subject is Firefox. One of two things is wrong:

Change happened, and as we all know, any change is bad. That's why people leave Firefox for Chrome! Stupid Mozilla!
Change didn't happen, why won't you change ${thing}? That's why people leave Firefox for Chrome! Stupid Mozilla!

Of course if I were calling the shots it'd be the most perfect browser ever, but I won't even so much as file a bug report even though it is FOSS. That way I don't have to come to the realization that things are more complicated than my own personal desires that may not even be reflective of the average user.

Separate processing per tab?

[Anonymous Coward]

Is there a browser that sandboxes each tab individually? I run into a lot of problems in Chrome with HTML5 causing all tabs to pause rendering when another tab is busy.

OFF-TOPIC: WHY!!

[higuita]

WHY are you still using windows XP?

no updates, no security and many apps already fail to run on it...

If the machine is old, switch to linux... if you really want windows, at least use the windows 7

Generally speaking a good thing

[m.dillon]

I'm glad they're finally giving each tab its own process. Of course, if they didn't they'd wind up in the dust bin of history... its about the minimum of work they need to do just to keep Firefox relevant. There is much more they need to do in addition. Honestly, Firefox's biggest competition here, after fixing the tab problem, is going to be chrome 55 with its significantly improved (reduced) memory footprint.

The interesting thing about giving each tab its own process is that although this increases the total amount of memory used by the browser, it also has the effect of reducing the memory fragmentation that forces the OS to keep almost every byte of it in core (Firefox is best known for this effect). With the process separation, the OS will have a much easier time paging out unused memory without nominal browser operation forcing every single last page back in. THAT is a big deal, and is one of the reasons why chrome is so much more usable than firefox.

I regularly leave my browser(s) running for weeks. Process sizes generally bloat up during that time, to the point where the browser is consuming ~8GB+ of ram. With Firefox the horrid memory management in-browser forces most of that to stay in core. With chrome, most of it gets paged out and stays out. This makes chrome far more usable, particularly considering the fact that my workstation runs from a SSD and has a relatively large (~60G) swap partition configured.

But these days I'm a chrome user. Firefox has been too buggy for at least the last 4 years. It crashes on all sorts of things, taking the whole browser out with it. And after all this time they *STILL* can't fix the idiotic pop-up windows. Disable popups only disables some of them. That and the bugs is the main reason why I stopped using Firefox.

In terms of sandboxing... also a good thing. In addition to the work the browser does, I also segregate my browser instances into multiple dummy user accounts (that my GUI buttons can just ssh into from my main account), and run multiple instances of the browser from those. One for unsecure browsing, one for browsing important accounts, and one with the most bulletproof setup I can think of (no video group access, no direct X server access)... which is slow, but about as safe as its possible to be in an X environment.

People often forget about user account separation. It's a bit sad.

-Matt

Confusion about processes and cores

[m.dillon]

There seems to be a lot of confusion about processes and cores and cpu use. I'm going to clarify some of this for people (I should konw after all!).

A process can be thought of as a memory management context. The memory image used by the program. But a process is not limited to just one cpu. A process can be multi-threaded. In a multi-threaded process, all threads share the same memory context but each is scheduled independently by the schedluer, so one process can in fact easily use all available cpu resources.

When a process fork()s new processes, each new process has its own independent memory mangement context. However, any data from the parent will be entirely shared by the child until one or the other modifies it (then it because separate). So simply creating new processes does not necessarily eat all that much memory. It depends what the new processes do.

Browsers tend to use a combination of processes and threads:

office1:/home/dillon> ps ax | fgrep chrome | wc -l
            23
office1:/home/dillon> ps axH | fgrep chrome | wc -l
          188

Firefox also uses both processes and threads:

office1:/home/dillon> ps ax | fgrep firefox | wc -l
              4
office1:/home/dillon> ps axH | fgrep firefox | wc -l
            69

In fact, most complex GUI programs probably uses both processes and threads. However, firefox, until now, does not segregate tabs into processes. The processes it uses are to mange other aspects of browser operation.

In terms of cpu utilization, one process thread can use up to 100% of one cpu thread. One process with N threads can eat up all of your cpu resources. Multilple processes won't eat up any more cpu resources than one process with multiple threads.

In terms of memory use, firefox has *HORRIBLE* memory fragmentation problems (and always had). This means that if Firefox has a VSZ of 5GB, it will probably be forcing 4GB of that into core even when idle or even if you are only messing with one out of many tabs. This has been a serious problem in Firefox for ages. One advantage of giving each tab its own process is that now the OS can take care of cleaning up after Firefox's stupid memory mangement (even if firefox doesn't fix it), because the process context is tracking the per-tab memory use and the modified memory used in that tab is not fragmenting the memory used by other tabs. So in a per-tab process mechanism, when you close the tab the OS can scrap all of that memory and that is a good thing.

So going multi-process won't make memory use any worse. In fact, it will help the OS separate the VM pages out and give the OS a chance to page idle memory to swap, whereas the memory fragmentation that exists in a single-process-many-tabs setup generally prevents the OS from being able to swap out idle memory resources.

In terms of idle cpu use, there are many issues here. 100%+ idle cpu use on idle pages can usually be attributed to three things:

(1) Bad interaction between the browser and the sound device (the intermediate streaming libraries have been known to cause problems in the past).

(2) The open tabs are running lots of ads with animations or video. AdBlock+ helps a lot here.

(3) The javascript on many (most) sites gets *very* unhappy when it is not given any cpu at all, so the browser has to give each tab some cpu so the tab's javascript from that site can run at least a little in order to stay happy. Also, a great deal of site javascript is badly written and consumes cpu resources even when idle.

Finally, in terms of cpu use, the operating system's scheduler usually does a good job but if the browser is causing problems for other work you do on the machine you can always nice +5 or nice +10 the browser. Or (in Linux) run it in a scheduler-constrained container. However, either of these actions will reduce the responsiveness of the browser. Most people don't do it.

-Matt

Re:Confusion about processes and cores

[epine]

(1) Bad interaction between the browser and the sound device (the intermediate streaming libraries have been known to cause problems in the past).

I'm running Firefox on PC-BSD and I can pretty much confirm your observations on resident set and the weird audio bug.

I experience terrible slowdowns after visiting various multimedia sites that take a browser restart to clear up. I've long suspected the audio, because audio has been a weird duck in other regards, as well. (I've got a current problem that any YouTube speed other than normal plays the audio at system rate limit until the audio runs out, usually within a few seconds even for a long video, and then the video freezes, too. This only started to happen sometime in the summer. Previously, it had worked fine. And this YouTube feature still does, under Chromium.)

Do you have a way to spot stuck audio using any tool up to and including dtrace? I'm curious about what's broken here.

The memory management fiasco forces me to restart FF about once per day (after a 2 or 3 GB working set, very sluggish). I'm presently at 2.6 res / 3.3 size with 17.7 hours of accumulated CPU, on a FF instance probably less than 48 hours old, which I hardly touched at all yesterday due to xmas errands.

Another "blessing" is that I have three screens (landscape, portrait, portrait) and a FF restart never puts my active windows back on the same screens (much less the same desktops, so I pretty much quit using desktops). FireTitle helps a lot by allowing me to attach name to my windows so I can quickly return them to their proper places manually.

My FF is pretty tricked out, and I still like it plenty, but the frequency of stone-age moments completely boggles the mind.

I'd like to add one final tip. Contrary to the love of all things sacred, Google Maps does not classify PC-BSD as "supported" (no matter your build) and defaults to "lite" mode, which makes the stone age seem positively neolithic. You can't even break out of the lite-mode prison by clicking away your unborn on the lite mode icon (bottom right corner of map).

But then I discovered that you can escape lite-mode prison by clicking on "your contributions" in the side-bar, which then tells me "you haven't written any reviews yet" and offers to switch me into "full maps". Cue the Blazing Saddles hallelujah chorus (no, Google employees may not leave the auditorium until the chorus concludes).

"Edge" = "new Internet Explorer"

[fbobraga]

please use the proper wording here on /., to avoid unnecessary slashvertisements...

Preliminary impression: It speeds things up a lot

[Mike Van Pelt]

I checked about:support and saw that multiprocess windows was not enabled. Installed the Add-on Compatibility Reporter and saw that WoT (yeah, privacy, blah blah spyware blah blah) and Ubuntu Modifications add-ons were incompatible with multiprocess. Disabled those two and ... wow. The speedup and responsiveness improvement is substantial. I also disabled WoT on my Windows 10 laptop, and the speedup there was also improvement there.

Part of it could be "Yeah, running fewer add-ons, of course it'll be faster", but I've got a whole bunch of other add-ons; Tree Style Tab, NoScript, LastPass, etc. I think multi-process is a win.

With Firefox 50, Firefox has screwed the pooch

[drinkypoo]

This (50.1) is the flakiest release in a long, long time for me. I have to quit the browser and restart it several times a day because it begins to choke for no apparent reason.

Replacement for the Ghostery add-on?

[Futurepower(R)]

The newest version of Pale Moon does not allow the Ghostery Add-on.

Is there a substitute?

Tor

[Anonymous Coward]

In the process, they probably created dozens of vulnerabilities that will be exploitable against the Tor Browser Bundle.