If @reC didn’t noticed (sure he did), I remember: ‘Every pot finds its own lid.’ And, as talking about two people here (in this thread) and there (in this forum) let us leave @user1 to enjoy her own peace (and weekend), as her “apparent view” or question, IMO, was based on current Linux phone options, where to invest (and how much).
I was just making an observation. It seems to me that the response to “what information do you have to back up your claims?” being “go find it yourself, I can’t be bothered to provide that information” is a terrible way to make one’s case. As for “putting my guns down” you seem to be way more fires up than I ever got. Do you also need something else to fill your time?
3 Likes
how about i put my pot on your lid ? 
1 Like
It was just some kind of dialog with you. Somehow related to this one:
Why am I a women now 
edit: thanks?
1 Like
The question is: what keeps the PF from FSF certification, and does it matter?
The reason the L5 has firmware blobs stored on secondary flash chips is so the user can’t change them, because the RYF certification prohibits user-changeable closed source blobs. Ostensibly, this is because the state of such physical flash chips is fixed (a pattern of electrical potentials), and so is hardware instead of software. This is silly. First, because your flash controller might well decide to load level your flash chip (probably not in the case of the L5, but that really doesn’t matter when considering the RYF requirements in abstract), so it’s not really fixed hardware. Second, because this would still just be a “hardware” configuration if it was stored on the same physical chip as something else, or was user “writeable”. In fact, all being user “writeable” means is the user can change their hardware configuration. And that is the fundamental problem with it.
The Respect your Freedom certification requires the hardware manufacturer to take steps to prevent the owner of the device from modifying some physical aspect of the device. When put that way, it sounds more like the Restricts your Freedom certification.
Purism isn’t to blame for this, the FSF is. But, if the only thing missing from the PF to get the RYF certification is keeping the owner from replacing the binary firmware (say if some anti-trust lawsuit or similar allows the creation of an open source version), that’s a point in favor of the PF, not the L5.
3 Likes
are you refering speciffically to the closed-firmware that runs the modem ?
here it seems that you - were - in fact, refering to that …
Better than a Master Card:sweat_smile:
To the best of my knowledge, that would be the primary closed source firmware blob (there might be a couple others lurking inside disk controllers or similar).
Which is actually something I forgot to mention, your hard drive has firmware. That firmware is closed source. It also doesn’t run on the host CPU, and doesn’t interfere with the RYF certification. It’s also user updateable.
3 Likes
but not the L5 firmware modem … as far as i know that is illegal to be made free-software … unless someone reverse-engineers it … what ? they would knock your door down ?
1 Like
To me restriction on loadable blobs makes perfect sense as to me it means I really get a hardware which must work out of the box. I don’t need to feed it with activation code, license agreement, bootstrapping procedure… just supply the power and it must work - according to the bus specification. That also means you can put on it whatever OS you can think of - which speaks the bus protocol - and it must work. So it’s just another dimension of freedom.
On the other hand I’d love this restriction to be spelled this way so that it maintains your freedom from the blobs while still giving the freedom to put the blobs (if you really wish doing so).
1 Like
Although I suspect a large portion of the cost is for the hardware, I’ve been wondering for awhile if they should fund the hardware and software separately. That would maybe keep the hardware costs a bit lower and allow people using a different device to still support the development of Phosh. And perhaps they could have an optional subscription to help with the continuing software development costs. I think it’s important to the phone’s long-term success that it is affordable for the average person and somewhat competitively priced. I believe we really do need more than one successful Linux phone for the ecosystem to thrive.
There are a number of advantages of this phone for me and reasons why the cost will always be higher than the Pinephone.
- Higher specs as has already been mentioned.
- Removable/replaceable modem and wifi cards.
- The wifi card has the firmware built in which was an extra cost but means there is no need for a blob to be transferred from the OS to use it.
- The screen uses the LM36922 to drive the LED backlight, hopefully meaning at the very least significantly reduced screen flicker when the brightness is lower than 100%. I’m not sure how this compares to the Pinephone but it’s important for me at least.
- Much more convenient kill switches than the Pinephone. Aesthetically, I kind of prefer the kill switches being hidden under the back cover, however, I honestly can’t see myself using them much on the Pinephone because of how inconvenient it is to switch them. I think it’s awesome they both have them though.
The Librem 5 adheres as strictly as possible to the FSF philosophy which is great, but it does mean some trade-offs. A bit off topic and I think someone mentioned this before but I would like to see someone (maybe System 76?) create a Linux phone at the other end of the spectrum from Purism. Performance over FSF and privacy/security ideals. I think that would round off the Linux phone market nicely.
2 Likes
I don’t understand the RYF and FSF certifications requiring that the user can not audit nor modify some of the firmware code. This sounds like marketing doublespeak that might come from Google or Microsoft to justify their invasive practices. Either the code is auditable and user-changable or it isn’t. I can understand perhaps Purism maybe making compromises to make the device be possible by maybe accepting some blobs and building-in measures to keep them benign through hardware. I can understand government regulations. But I can’t understand RYF or FSF saying that ‘we need to lock you out’ for any reasons that are legal or ethical per their respective stated purposes for existing. Free means free, ‘not like free beer but more like freedom of speech’ (paraphrasing Richard Stallman). That pretty much defines all we need to know. The firmware is either free or not free. It’s an easy binary choice. Coming from either of these organizations, the justifiable reasons to lock anyone out have to be non-existant. Coming from Purism, perhaps a compromise justification could be appropriate.
It would then be a 0 but through their effort of locking it kind of out through hardware, they say they are 1. To quote a show I admire:
Mr. Robot : Tell me one thing, Elliot. Are you a one or a zero? That’s the question you have to ask yourself. Are you a yes or a no? Are you going to act or not?
Elliot : Yo… You’ve been staring at a computer screen way too long, homie. Life’s not that binary.
Edit: I remember being linked to a talk they did on stage about these blobs and you could see the frustration in their face when people were scrutinising them with questions so I think they won’t just try to do their best which means nothing but actually do a good job and just segregate through hardware. The whole purpose of the phone is to be a phone that is FOSS and if they do not even keep to that, what do they even represent? A ‘we tried’
Edit2: Maybe I just remember wrong but here is the video if you want to double check because my memory isn’t always persistent: https://youtu.be/rGHqSPnWEDw
Edit3: 37:06 is a timeslot I found and I think there may be more.
Yes, it is illegal, but if not free-software and therefore I’m not allowed (not free) to update/upgrade my (future) non-free modem firmware, is this somewhat weird? And, we must be talking here about BroadMobi 4G module BM818 M.2 or any other similar (like Neoway 4G module N720 Mini PCIe) that are based on Qualcomm MDM9607 chip**:
Or eventually, download and upgrade particular Linux modem with approved (and current) firmware that is allowed as part of non-free Linux Sources, through some Purism-customer-only logon download-site or reachable just through official regional (USA or EU, etc.) OTA update images, from BroadMobi or Thales Group or some method like LVFS from https://fwupd.org/, other method not known to me, as I’m not an expert.
My above food for thought was based on:
**Quectel EC20 4G module internal structure. Here is link, for example, to available EC20 firmware revisions. And if not having/using Windows PC is there another way to update modem firmware? Still, similar method might be promising, see EC20 link, under tab Files.
That’s a non sequitur. You aren’t likely to update the firmware on your hard drive to an open source version. You’re likely to update it to a newer vendor-supplied version (to fix power management or longevity issues). Similarly, the L5 modem vendor might come out with new firmware, which might improve idle power usage, fix security issues, or otherwise improve the modem’s behavior. Some of us might like to be able to use that updated firmware without needing a hot-air station.
As for not being able to make it free software legally… That’s not exactly true. It would only require the agreement of Broadcom and several other hostile-to-open-source large companies, and possibly the permission of the FCC and other hostile-to-freedom government agencies. Not likely to happen anytime soon, but remember, Microsoft itself released the source code for an ancient version of DOS, so stranger things have happened.
3 Likes
maybe but that’s a really fringe example … if only it were that obvious on how to do it …
a funny thought: i’ve set-up a linux box recently for a neighbour and i’ve used a WD blue 512GB m2 SATA 3 SSD and slapped ubuntu 20.04 LTS RC on it because of the newer APU from AMD under x470 platform. so MESA it is … pft. 
the thing is, no samsung drive has shown the temperature reading inside the GNOME-disks front-end for NVME (as that’s all i’ve used from samsung so far) … but now i’m thinking it’s because WD uses an FPGA microcontroler on the board and that’s what caused it to be loved by the Linux-kernel … maybe it’s just me but does only SATA act this way or is it ok for NVME as well ?
What would happen if a silicon manufacturer approached Purism and said this? “Give us your wish list for a cell phone SOC. Include every feature you want. We’ll build it and publish everything necessary for the open source community to make full use of all available features”. Would anything else stand in the way? Eventually, someone is going to steal Apple and Google’s lunch.
It’s pretty rare to update firmware in a device that’s been around a while. Usually, the manufacturer gets to where they’re reasonably happy with the firmware for the product, then moves on. When it’s new, it’s pretty common for there to be firmware updates, as people find serious bugs (for example, the Radeon VII shipped without UEFI support).
Since the L5 modem is old, it’s not likely to be an issue in practice.
I don’t know what Gnome-disks uses to pull temperatures. I use nvme smart-log $device | grep temperature | sed "s_.*: __" to extract the temperature (which can go into an X widget or similar (I use a curses display)).
As for why you have hardware monitoring easily with the WD drives. It might be that it’s SATA… WD also actually cares about Linux, as their drives get used in lots of high end Linux servers, so it may be that they actually bothered to provide drivers for reading statistics through the unified interface… I know nvme support is still in progress on Linux. atop only recently added support, for example.
1 Like
and what’s even funnier is that sata ssd drives don’t normally get hot at ALL so it’s just something that stuck to my mind due to the fact that it simply WAS there to be gawk-ed at 