Intel Laptop Users Should Avoid Linux 5.19.12 To Avoid Potentially Damaging The Display (phoronix.com) 48
Intel laptop users running Linux are being advised to avoid running the latest Linux 5.19.12 stable kernel point release as it can potentially damage the display. From a report: Intel Linux laptop users on Linux 5.19.12 have begun reporting "white flashing" display issues with one user describing it as "[the] laptop display starts to blink like lights in a 90's rave party." Intel Linux kernel engineer Ville Syrjal posted this week on the kernel mailing list: "After looking at some logs we do end up with potentially bogus panel power sequencing delays, which may harm the LCD panel."
just wait for hp or other to say linux = void warr (Score:1, Flamebait)
just wait for hp or other to say linux = void warranty
Re: (Score:2)
They didn't say so when Windows updates left the system unbootable, or damage the HDD, etc. so what makes you think they will do now?
Re: (Score:2)
Sorry, you install an incompatible OS, it is either $1000 to repair it or $1000 to get a new one.
Re: (Score:2)
I don't know if you're a troll or if you're genuinely ignorant.
Laptops have been outselling desktops by roughly 2:1 for years.
In my house, we have two desktops and 4-5 laptops (one is retired/spare... count that as you will). One is a gaming laptop, one is a mobile workstation, and two (plus spare) are ultraportables.
At work, all users are provisioned with laptops or VDI seats unless they specifically request something else. Very few do.
Re: (Score:2)
It's easier to sue a corporation like Microsoft for damages than a vague collection of open source kernel developers or various Linux distributions. Most businesses are risk adverse so it is a conceivable response. But I doubt it from HP, Lenovo, or Dell as they've been making real money on Linux in at least the data center markets but also have seen some uptick in the workstation and laptop market.
Re: (Score:2)
Because with Windows, it's happening with the OS that they supplied with the hardware and officially support.
Also, they don't want to piss off Microsoft. With Linux, there's nobody important they'd make angry.
Re: (Score:2)
I wouldn't count on it as HP is now selling AMD-based Linux laptops. /shrug
It must be about 7 or 8 years (Score:2, Troll)
since i tried arch on a laptop, which fried the display. Even before the bootloader handed off control to linux.
i got brutally ridiculed for daring to speak such vile accusations. No "huh... you sure? Did you do anything uncommon? Was the laptop otherwise ok?", just straight up "you're lying, that is not possible".
Re: (Score:2)
when software damages hardware in a matter of seconds where should the blame lie?
Re: (Score:1)
Usually the software. E.g. two industrial robot arms try to pass through the same space simultaneously. Or your fan quieting software doesn't monitor temperature and allows the CPU to overheat. (less of an issue now that "drop dead" thermal shutdown circuits are fairly common).
Unless there's some reason that the hardware should reasonably completely exclude a capability necessary to cause a failure, it falls upon software to ensure that the *combination* of conditions necessary to cause damage never occu
Re: (Score:2)
Unless there's some reason that the hardware should reasonably completely exclude a capability necessary to cause a failure
The display has a controller. That hardware should do sanity checks.
You could certainly try to blame the hardware manufacturers for not publishing such failure modes - but they may not even know it exists.
In that case, I would blame them twice as hard. It's their hardware.
Re: (Score:2)
The display has a controller. That hardware should do sanity checks.
I'm on the fence.
It would prevent things like this, but hardware has *never* done sanity checks.
Overclocking, overscan are a few examples of a lack of restriction to running devices out of tolerance.
In that case, I would blame them twice as hard. It's their hardware.
Eyeroll. You can destroy an AMD processor as easily as an Intel processor. You just crank the vcore until it pops. All the protections for that are in software. There's no overvolt protection on the CPU.
Again, you might be right that maybe it's time to revisit the current paradigm, but it has always been pos
Re: (Score:2)
here on Earth we use electro-mechanical stops to prevent software from overrunning equipment and damaging it.
Re:It must be about 7 or 8 years (Score:4, Insightful)
Usually the software. E.g. two industrial robot arms try to pass through the same space simultaneously. Or your fan quieting software doesn't monitor temperature and allows the CPU to overheat. (less of an issue now that "drop dead" thermal shutdown circuits are fairly common).
Never the software when you're talking about a general-purpose computer that's designed to run various operating systems, etc. The hardware vendor can't control the software, so it has to be designed to accommodate or reject bad input, in much the same way as a function or method has to recognize invalid inputs and fail in a reasonable way (which could be an error message, an exception, etc.).
In fact, I would argue that the same is also true for your industrial robot arms. They should have sensors to detect that they are about to collide with something, and they should stop. If they don't, then that's a design flaw, because that also means they won't stop when, for example, they hit a person who is in the wrong place at the wrong time, and suddenly someone is dead.
It is *always* the responsibility of hardware to reject input that would cause physical damage to the hardware (or other hardware or people or...). Software is responsible for correct functioning of the software. Hardware is responsible for correct functioning of the hardware. Expecting software people to not break your hardware is how you end up bankrupt after a virus bricks your entire product line and you have to recall millions of devices.
Were I an HP laptop user, I'd be inclined to upgrade the OS, get it repaired, upgrade again, get it repaired, etc. until lemon law kicks in and they are forced to replace the hardware with a laptop that doesn't have that flaw. There's really no excuse for such a critical vulnerability in hardware.
I'm guessing this is some MIPI display where where turning on a flat panel involves sending a series of I2C commands (IIRC) to tell some bespoke ARM chip to power up the panel in a certain way. Displaying things on a screen really should not be that hard, nor that device-specific. If the software has to do so much work to turn on the panel that it is possible to do things in the wrong order and destroy the panel, then the panel is designed wrong. If an I2C write can fail and cause the panel to destroy itself, then the panel is designed wrong. And so on.
Re: (Score:3)
Usually the software. E.g. two industrial robot arms try to pass through the same space simultaneously. Or your fan quieting software doesn't monitor temperature and allows the CPU to overheat. (less of an issue now that "drop dead" thermal shutdown circuits are fairly common).
Never the software when you're talking about a general-purpose computer that's designed to run various operating systems, etc. The hardware vendor can't control the software, so it has to be designed to accommodate or reject bad input, in much the same way as a function or method has to recognize invalid inputs and fail in a reasonable way (which could be an error message, an exception, etc.).
That's my opinion as well, I thought we learned from the killer poke [wikipedia.org] many years ago...
Re: (Score:2)
Using software to sequence events to avoid damage is very common. Often it's done in firmware, but to reduce cost it can be done by the host software.
Cost sensitive consumer hardware is only designed to be used with the Windows drivers. It's been that way for decades. Many Core era laptops with Intel CPUs need a model specific driver that had all the thermal and power limit data. Without it, at best the CPU would run extremely hot.
Re: (Score:3)
Usually the software.
No never the software. Software doesn't talk to hardware, it talks to pre-written firmware. It's the job of that firmware to have fail safes.
Or your fan quieting software doesn't monitor temperature and allows the CPU to overheat.
Your fan software doesn't monitor temperature full stop. It's the job of a fan controller firmware. When your software fails, locks up, or was written by the morons at Gigabyte who can't seem to code hello world without screwing it up the fan controller will none the less keep the fans going the way they always do.
But the firmware on the controller is user adjustable o
Re: (Score:2)
No never the software. Software doesn't talk to hardware, it talks to pre-written firmware. It's the job of that firmware to have fail safes.
Generally correct.
Generally the EC firmware is where things like this would be prevented.
Your fan software doesn't monitor temperature full stop. It's the job of a fan controller firmware. When your software fails, locks up, or was written by the morons at Gigabyte who can't seem to code hello world without screwing it up the fan controller will none the less keep the fans going the way they always do.
Correct and not quite correct.
Fans, again, are nearly always controlled by the EC firmware. You can, however, directly write to them, and you can turn them off. And doing so, you can damage components in your system.
But the firmware on the controller is user adjustable over the SMBus you say? You're absolutely right, and yet it would be the hardware's fault if the CPU then overheated since something you can't control is the CPU's own protective throttling.
Something you can control is the CPUs vcore.
And the CPU is protected by overvoltage by nothing but firmware.
You have always been able to destroy hardware with software.
I absolutely would blame the car manufacturer if the engine failed when send data to it via software on some openly published bus.
Well, a court wouldn't agree with
Re: (Score:3)
Considering its linux we are talking about its obviously the fault of the noob moron user
Re: (Score:1)
For those of you who got the answer wrong (Score:2)
The correct answer to my semi-rhetorical question is: The engineers
Re: (Score:2)
i never expected my point to be proven so beautifuly
it is almost poetic
linux is amazing, but its fans are mind-numbingly shortsighted
Intel make laptops? Who knew?! (Score:2)
Re: (Score:3)
Quoting from the linked article, something you had not bothered reading
My 2013 Acer Travelmate has Intel graphics so this problem could potentially affect it, but it's running a highly modified vendo
Re: (Score:2)
They need to get marketing!
They already have. For decades laptops had "Intel inside" stickers on them.
Intel GPU != Intel CPU (Score:5, Insightful)
Re:Intel GPU != Intel CPU (Score:5, Informative)
Most laptops that do not have a separate graphics chipset, have an Intel CPU with an embedded GPU. So this can potentially affect a whole lot of devices.
Re: (Score:3)
I vaguely remember there being some really obscure soft-launch parts that were quietly dumped into trash tier Lenovos exclusively for the Chinese market, or something equally obscure, with the iGP disabled back when Intel has having really awful 10nm yield problems but needed to state that they'd indeed shipped 10nm parts; but aside from that I think you pretty much have to buy a Xeon(and
Re: (Score:2)
Are there even any Intel CPU offerings that would both fit in a laptop and not have integrated graphics of some sort?
Probably correct. However there are (I think) gaming laptops with gives users the option to switch to a separate graphics chip, e.g. nVidia. So maybe that's a workaround until some kernel hack or firmware fix is released? .
Re: (Score:2)
Re: (Score:2)
LCDs typically have many power rails - you have AC drive for the pixel themselves (AC because it flips polarity to avoid polarizing the display with DC), the transistor in the pixel needs a drive voltage, then there's the line drivers and pixel driver chips which have their own supplies, and the interface chip power. A lot of these rails can be common, but
Re: (Score:2)
Clearly I'm wrong in some respect: do you know if the video interface works at a lower level than I un
Re: (Score:2)
Until recently Intel only offered GPUs integrated with the CPU. So essentially any laptop with an Intel CPU is potentially affected, unless you're running with the discrete GPU enabled.
Re: (Score:2)
On notebooks, even with a discrete GPU, the integrated GPU may be actually driving the monitor. NVIDIA Optimus works this way - when using the GeForce GPU, the rendering is done on the GeForce and it does a fast DMA transfer to send the pixels to the Intel GPU's frame buffer for output to the screen.
bad hardware design (Score:1)
fairly common at this point.
let the hardware do whatever it wants and let the software control it.
Software does the wrong thing ? oh, well, we get to sell another laptop ?
Really, somebody should be embarrassed about this. Power supply sequencing should be hardwired. A good hardware designer would NEVER allow something to power up in some fucked up state that could damage it.
Re: (Score:2)
Really, somebody should be embarrassed about this. Power supply sequencing should be hardwired. A good hardware designer would NEVER allow something to power up in some fucked up state that could damage it.
Pure nonsense.
From your computer today, you have the direct capability to turn your CPU into a smoldering pile of blown silicon junctions.
It has always been the responsibility of software to drive hardware correctly.
LG CD-ROM bug (Score:2)
I borrowed my brother's PC for doing something with a live Linux CD and it happened to
HALT AND START RAVE (Score:2)
Bum tst, Bum tst, Bum tst, Bum tst, Bum tst, woohoohoohoohoohoo, woohoohoohoohoohoo, woohoohoohoohoohoo, ...
I heard (Score:2)
Yeah I heard 5.19.12 is "too thick" and may crack the display when the lid is closed. :)
(Like on a Mac with their legendary 0.001" clearance spec. Don't close the lid with a business card in there or bad things can happen.)
So what are todays rave partry lights like? (Score:2)
n/t