 So Thanks for being here for the last session of the safety critical software summit It's not that still a bunch of audience is here on a Friday afternoon. It's not too big like us so Yeah, before I jump in Normally I have a short shout out to my employer, which was We're not yet there. Not good enough. Okay, but I do it like this. This is better We take it in this way. Okay. So first of all is short shout out to my employer, so I'm at Bosch and I want to Say thanks because I'm working in an embedded Linux IoT project Well, we make use of a bunch of open source projects So I say once thanks to all the communities which making this possible and also thanks for Bosch for sending me here That I'm able to talk about some parts of my work what we do with all these projects is like we're bringing them in very different product areas and I also like to highlight this because not everybody is aware What Bosch is all about so something about the fridge or other parts some bought the automotive But we really go into very very different areas and all of these areas include Linux and Some of them also in safety critical context. Some of them have to run for 10 15 years and so on so long lifetime And yeah, though this also makes it attractive for us to work with Elisa and to give a few words about me, so I'm Philip I'm on this you could already see I am currently technical business development manager and It's a long title, but basically I take care that more Bosch engineers Can work on open source get in touch with open source that open source gets spread inside Bosch and try to foster Collaboration with communities inside Bosch and so on besides that. I'm the technical steering committee chair of the Elisa project So that's also why I try to place some update or outlook session and more generic part on this Not a lot. I just came from a Linux meeting But there's the Linux Foundation Europe advisory board where I'm also a member in Linux Foundation Europe founded last year during the last embedded open source summit or open source summit in Dublin. That's where I'm also a member Generally on Linux. I started my Linux journey not as early as some others Which may sit here just other on 2005 six time frame where I was setting up Linux PCs for exchange students and when I was studying in Sweden, they couldn't afford their own Laptop cat didn't have a notebook or so So we used old PCs from the university and brought them to the students so that they had something To serve and work from home at this time and we did this on a Ubuntu 610. I guess Right for automotive perspective That's a little more than 12 years. We brought Linux into the car with infotainment devices navigation device at first was a 2.628 kernel I guess Was nothing about secure boot at this time, right? So if you if you search good enough, you will find how to hack these devices these days It's different and we take a lot of care of this. I Put the apparatus on this is our district But it's not the one which is in the system at this point of time where to plain yoke to usage at this And now we come to the real topic the aliza things and I start typically was a little bit of disclaimer Saying what is Linux in safety critical systems? If you see some other slides I have seen a session before it's basically the house keeping things for aliza So to always mention it and what I want to shout out here is that if you see and look into system And whether to sass if a system is safe, this means you need to require the system sufficiently And in this system you may find a Linux part of it so the Linux itself is not the full system as you may judge by just in the system context of a wider system and You need to understand how this Linux system was part of the system is used in there And by this you will see there are certain parts which may be relevant Others are not so you need to like select which Linux components and Features are there and which are safety relevant which may not be safety relevant Which is the core of all the activities and when you look at this. I'm pretty sure we will identify Gaps because if there wouldn't be gaps for Linux and use in safety criticism system, we would have already done it What's makes it hard for Linux is that it differs quite a lot from traditional Safety critical operating system traditional safety critical operating system will be certified it was taken one subversion it comes with a very small food print It fulfills a lot of standards. It may have hard real-time capabilities The real time is making a good step forward not everybody says it's hard real-time and I agree So the Linux kernel are cheap. I just will not guarantee hard real-time in any case, but we have real-time in there But the traditional systems also face some challenges if we see a coming from automotive background We have these software defined vehicle environments if you like the word or not, but still it will happen that you have virtualization involved containers Arches next to each other multiple VMs running and if you want to put all this only with the certified arches You may have hard times and even just bringing art using an arches and hyper by then there to Scale on high performance computers It may be that your safety safety certified as which was prepared for chip environment ten years back Does not fulfill your memory demands your cash demand and all other things So you will face a lot of challenges and suddenly Linux becomes very attractive because you have a rich ecosystem You have strong community support. You will find drivers. You can discuss things open but you have to take the challenge of Well get somehow to a point of certification getting it robust and reliable enough and that's where we work in Elisa and to say this clear. We are not Safety certifying a Linux. We want to enable others To with the right tools and so we seal in the mission shortly to enable then so I come with Like the legal disclaimer. We have collaboration is a password collaboration, but we also have Things we cannot do and the things we cannot do is like well We don't make your system safe because we don't know your system Why you have your use case of mind you have your obligations You need to see where to use it and if I take it like a security part You need to if one security perspective on Linux kernel if you have C groups name spaces other things in there You need to know how to configure it just because it's in there. It doesn't mean that the feature is used properly and When we describe processes guidelines and so on that's also something that you have to take care on on your own We cannot directly because we deliver something to the open where we don't get any or sometimes not any feedback on this back We will not have an out-of-three module We are straightly try to keep up with the mainline speed because there are continuous improvements This may be a breach to the traditional way of thinking about safety about certification But there are so many things going better and just because they are more boxed in this kernel than in a 4.14 It doesn't mean that the 4.14 kernel is better than the one now Yeah, and you remain responsible for a lot of things Nevertheless, we find enough partners even if we have this limitations which support us in this mission Which bring things in and we have some premium members Which are like from aerospace recently, but also redhead was really strong and driving activities so also redhead in vehicle as a lot of automotive parts in there and Spread all over the world. We get a little bit of associate members This basically mean they don't pay for the membership, but they support us in the one or the other way Like for automotive grid Linux We have collaboration on the use cases we use some of their work and also with CIP Which has long-time support stability robustness we all cooperate to certain extent with them as well Osada is mentioned because of the history. There was a still to Linux MP project in there And at least I basically continued this with a little bit chain scope, but the idea remains right all the different members subscribe to a mission and This mission basically concentrate on the element processes tools documentation to make it amenable for Safety certification, so we make life easier than we get things forward. We do this by said of horizontal working groups The safety architecture basically looks into the Linux kernel So more deeply looks at the subsystems whole subsystems are interoperating and so on try to understand how Linux is used for use cases and This is a wide approach not very bound to use cases sometimes was a use case focus and in a similar part There's also the Linux features work group and I mentioned the part of example with C group namespaces And so on this group really tries to figure out what is there what could be Supporting features for safety was in the Linux system or what could be something which breaks with safety So if you switch this on or off or configured in the wrong way This could harm or benefit for safety, but it still remains on the product which you build And you will bring expertise as an integrator to just bring this in The tool investigation code improvement work group has a little bit different focus. It's the smallest group more or less and they set up Some tools like syscaller code checker and so on run it on the kernel It could benefit from a little bit more Contribution use case focus so we if someone thinks you're very good in tools reach out to me later So here we are really actively looking for a larger path in there And yeah, the old set open source engineering process This is more really on the process side like how can you argue? Things how which analysis tools are possible within the Linux kernel We use heavily the system theoretic process analysis There was a raffia talk from coding this morning, which also goes in this direction my portrait would he presented the work Yeah, and the last is the systems work group. That's where which is the group which I'm leading also here We try to bring all the different pieces together So we put the things into a system and into a system context which fits into the original mission which we had so to say here's something you need to understand the system context and Normally this would be something which you put in a safety element out of context If you take the automotive language, which means you have something which is not in context But it doesn't work whenever you have an out-of-context you always have an assumed context our assumed contacts Are prepared by verticals There's the newly founded aerospace work group. They don't have the use case yet They just align with partners looking for members try to see where's linux already news find commonalities Before they really drive into use case and they're really open for everything from their thing From drones or so up to going into space passenger flights and so on Automotive you see the AGL old style instrument cluster in there That's where we base on in the automotive work group We work on the little danger sign there to look for warning signs tell tales and see that they operate properly That's our main use case. That's also what the architecture work group puts up and the second one in there is the Open APS This is from the medical device work group. It's a very nice one. It's an open artificial pantry system This was developed in the open. It uses standard components, but as the control of insulin is a very critical sensitive sim and highly safety regulated in the field of medical these equipments are very expensive and Many people suffer from it that they don't kind of force such a system because of the High amount of cost for the certification. So what someone came up with was Dana Lewis. She put in a raspberry pi in the middle and Let's script run on it and this script controls basically the insulin pump and the glucose monitor things And by this she gets a much better work or better much better live quality So she she was telling about improved sleeping quality because she's not waking up by an alarm at night and so on and Here the nice thing is is really that everything is in the open We could do analysis we had outreach to community and this is a bit different to automotive Which is more a clocked and closed space Still as we were talking about the system before if we see a typical architecture. It's not that every Element fits into all the parts. So mainly these two use cases Which you have seen are in this more reddish middle part about the Linux and we need to see how we get it further and Here also then the architecture work the features were combining the kernel the code improvement also in their Tools engineering process will more or more on the surroundings of it It's for the whole system environment fits into every scale and the use case just basically tailor her whole system We do this work also in collaboration with other communities, which is quite cool So we reached out to Xen and sapphire community because both have a safety critical pass They have same challenges. They have different challenges But at least it's a good room for collaboration because the room of open source and safety is a little limited Especially for the automotive part. We see that the architectures are repeating and Sophie eclipse SDV Have also architectures which have of contain involvement virtualization wolf and are taught in there So we want to reach out to them as well and we reach out already and we have also related communities like the Octo project Which participate in the work? We have a special interest group for it was an SPDX for safety so The call is working on it. She's here Kate also and we also reached out to in our already because also here a lot of the I make a secret just fits in and it always remained the idea of yeah, I Like this if you have an apple I have an apple if we change apple then we still both have one apple But if you change an next idea and with me and I have an idea changing with you Then we both have two ideas, right? So that's basically the overall concept of this collaboration to just get a knowledge spread get a wider open All these things and we bring this into artifacts These artifacts are split into various elements is based on the mission. We have elements processes tools software parts the working groups Work with different extent on the set of different areas So not everybody works on processes because it's more generic thing and not every working group need to go in there But we also have a bunch of activities and The meter Eliza, I will shortly mention the rate of status which we have About the reproducible system. I give you a short insight also workload tracing which has been recently upstreamed and the latest activities arts of Linux Other things are the STP and all those code checker Cultury and if you want to know about this just check the Eliza YouTube channel For example, then you will find all recordings and get more insight on this All this is also documented So it's on github gdocs and in this presentation there are clickable links So you just if you download them you can click on them and should be guided to the different I first Few words about the automotive part just that you hooked up from the use case. We've selected this warning signs of an instrument cluster which could be the air bake sign or Oil pressure and so on because it was a nice and simple use case, but complex enough to put it under analysis It has for example less real-time constraints and it's easily reproducible. You can make a simulation out of it so this is where we started and Therefore we had an extension which is Still a work in progress. So there's the meter Eliza layer. It's not integrated into a GL. We maintain it locally in our github We share regular with a GL folks but it's more there and this is like seen as a demo set up Where we can improve where we can trace workloads and so on this Was always a challenge for us to prove this and get this forward to different working groups So or for just also for people joining newly So for this work, we really come to a flow where You can either start from a plain document taking a step-by-step guide forward Or if they I go with a document rather than building things a full yacht on my own I can either enable it as state or not. I can even take a pre-made image of the QEMO and just boot a QEMO and But this you can see and play around with the potential system This is made in a very dependable way, which means Each step of this queue depends on the previous one. So the docker file depends on the github documentation It's creating the image the git lab pulled the image which in the end means Wherever there's some break in this queue. There would be an issue. We would see it and can trace it back to its origin and We see for example the git lab See I more like being a user through the steps, which are in the github are basically the steps Which are documented and pulls so we hope by this we get more stable on the things But this was all QEMO and I said something about the reference system This is something where it comes to real hardware So we want to bring the work onto reference hardware and for those who have been there on Wednesday there was Talk by Thomas colleague of mine and me. We were talking about our setup with Xen Zafire and Linux We currently bring it on the Xilinx XU. We tried with renaissance before And the idea is really that you can use a git hub Documentation later on get all the artifacts build artifacts download the things and reproduce something on your own the XU hardware is Quite expensive. So therefore we still look for an alternative in there We have basically selected this hardware as the reference because of a good community support within the Xen community and We will work on this first of all forward until someone comes with a better board because the idea is really to figure out Something where I have at least two hardware. I have two different kind of Linux and they're like it's mentioned The pita Linux and a purchase, but we also have Agl considered so we at least have a yok to an a devion version in there later on since we're conservation and we start with Zafire, but we're also looking for an alternative artist so that you get a skeleton at the end where you can put whatever feature you develop Into context let it be safety or not such as you just have something to learn about right So if you missed this on Wednesday the slides already online and they can you can access the recordings So take the chance. It was quite good presentation, which Thomas did there The other thing which I'm very proud of is that we have from the medical devices where the sdp I was done it reached to a certain level So it's more like a systematic approach drawing some boxes and then at some point of time the group figured out Oh, we don't get deep enough into this topic We need to understand what's really happened in the subsystems of the Linux kernel how things are how does it work? look like and they're Sure and Alice did a great job of upstreaming later on this documentation. I think was also she finally she did this She wasn't meant he was in the Lisa project and dated upstreaming of This guidelines into the user and admin guide That's something which really tried to do more that you as potential user awful in a system in safety criticality can Make use of these kind of things the upstream part is more than the generic case with stretch and G while The work of the record of course goes on the low level of the real workload use case the continuation of this for the Linux features work group also here in cooperation between Ilana. She's leading the work group and Sure and some other members of color work group To bring more RT documentation upstream because we see that the RT patches get mainline It's close to all being done, but How do you really use it hard what to do what to avoid what not to do? What could cause problems? This is really documentation which is needed and they therefore have all the presentation this morning Maybe a bunch of you have attended if not all their slides again later on and your recordings and here They look into various kind of updates, and this is really achievement which will go on and end up then in the kernel documentation But we also have upcoming this is quite an eye thing is the usage of Yachto all Linux in aerospace So Boeing was presenting some work during the open source summon of America in May and as we had so great feedback on this and Large audience there was a request to also have an extension of this presentation So we plan in the middle of July a webinar So dates Maybe around the 18th of July, but there is no direct invitation or QR code, which I could share here I know so maybe watch out the Lisa social or the mailing list So it will be all announced there. So we'll make some promotion when it's there somewhere in the next weeks right and Yeah, like a pre-check you can just search on YouTube to find the recording of it there old recordings Or all life. It's a very nice one to see also how Linux gets some responsibility in aerospace Although it's mainly not on a full safety level. It's more on the lower level of these standards rights Another thing which comes which is not mentioned here. I guess this goes more towards September. We will also have Neighbor community webinar so to say so we will have some parts of the extent project also in this explaining their past our safety How they deal with misra how? their work differs from what we are doing how the virtualization environment differs from the plain Linux. So this is something which will come up in September again here watch out To see it coming. Yeah One latest news, I guess that's a really new slide which are Nobody most likely has seen because I've just prepared a last beginning of the week So there I know there was the deadline to submit everything on Monday, right? This slide came in a little bit later We had a workshop in Berlin Last week was a bunch of new participants new companies in there And this was a really great chance to discuss where should the project go? Which problems do we have where do we see challenges and One point which was mentioned that we want to concentrate on the core of the core How I call it so we know everybody has a Linux kernel, but everybody configures it in a different way Still there are similarities in almost every system and we want to go on and see what are these core components of A kernel which all have in common because there is most likely a strong safety criticality in there then From our mission you can see that there's elements processes and tools But we also want to phrase a little bit clear what this means. We don't have a good name for this yet So if you're coming in automotive space, you know, maybe about a safety manual which gives you an instruction on how to use your System and we cannot write a safety manual So it's a question if it's an unsafety manual or a safety how to or whatever by at least what should go in there It's like what are certain Issues in there. What are certain capabilities? What do we have to consider so that you get? Most likely it starts with a little bit of input and then it grows over the time that you get a guide document Which helps you in understanding the usage of Linux and I'm pretty sure a bunch of this will also be in the user and admin guide then so these are both related and What we also figured out that we will work better on our big picture document the big picture document should help all Those which are in the community to judge if the work fits So if we're doing in the right direction, but also for those joining newly to the project To find a part of what is really safety critical part and what is the ELISA project just doing To get operating. So for example, we need to do certain processes in order to have good Reviews guidelines and so on internally But this is nothing which you would need to care of as an integrator or a provider of the system because you will bring your own processes and There are other things which you may want to use like our analysis techniques But also this you may bring instead of an SDPA and FME a FTA whatever is your preferred method And we want to clearly differentiate this and say these are the core parts of it. This is really from the kernel part This is how we differ from Sapphire from sand for example, this is what we have in common with the other communities This will be in a big picture document because a lot of the people were asking for having it and in order to Prepare this also. We see that it's very good to interact with the other we Decided because we had as TV involvement in there and a little bit of Sophie involvement in the workshop as well So it looks as TV and Sophie we will take a meeting together with a GL and also line more on the work of systems Workgroup that elements from the systems workgroup get into the project and that the other project bring in value in there So that we're not all reinventing the wheel or it was the outcomes from the Berlin workshop as key takeaways We have some more activities going on What's in there? For example, is the s-bomb generation for Linux? We have enabled already the s-bomb Generations so in our CI setup, which I mentioned earlier for the automotive use case where Sudeep was doing a very great job. So Thanks a lot Sudeep or We will continue with it because it's not Linux in the system And we know that system bill often or suffer bill of material and the system bill of material is very important We want to get this in there and see how does it work with that fire? How does it work with send? How do we get the things together really properly automatically generated? Then we have this reproduction of sense effort Linux But it's not in a state that it's really robust and that you can have it in a github PR We have a github PR for the renaissance one, but we most likely don't get it into a CI at the end we want to get a first on the exo further and then have see a comparison of both guides how things go on Therefore the documentation is something which will come up in the next months directly Was once as this achieved also the s-bomb part will move on. So basically the orders swapped here It should be moving on from a pita Linux from just a Linux to really have all the chance to exchange your Linux into the preferred version of Linux So that you can say I'll say I'm a Davian guy or I'm an yok2 guy Or I would like to go in between with some part of one or the other But at least the system should work with all of them and also this opens new domains because the Agl and apparatus are more in the automotive space while Cip is in the civil infrastructure and as a leader is also cooperating with these With this project in LF we want to see that we also get their Linux in there Here always the topic you cannot mention often off at the upstream part the upstream work because we want to get as many things upstream as Possible and to increase the documentation. So we see that this is really demand there. There's a lot of improvement possible Here and I guess the last Point was all that we need to improve our internal development process We have a lot of things written down, but they are not really structured their cross tracing is missing and especially for onboarding new people We got quite a bunch of new people joining the working groups. They ask questions. We see repeating questions So we know okay, let's see here How do we align that we have really a streamlined process and that people have an idea way into the community Right by this these were the main activities. I can say we have a bunch to subscribe So we have the mailing list We have weekly meetings or by our fortnightly meetings. They are all open So as soon as you subscribe you get the invitation even the technical steering committee meeting is completely open But you can also bring a concerns discussion points ask where to go best where to put your use case we also have the chance now to ask and Yeah, we I love to hear your questions because I Reach already the end a little bit earlier slightly ahead on time But this gives us more room to ask some questions Or get early into the weekend Okay, thank you. I have the microphone here. If you have any questions I'll pass the microphone to you. If there's no question. I have questions. I just want to ask who is from automotive in The room who brings an automotive background? That's of you. What about medical a Little less industrial We want to have more inputs from the industrial work group actually and we try to find to Found one because we see especially due to the RT patches getting in they're getting new use cases right where There's the first time and we can say oh, we don't need the artist anymore We could bring it in Linux and combine certain functionalities. So if you have a statement on this Let's say, okay. I have some use cases which I'm able to share or I have an interest I have a certain concern on this we treat things completely different though That's why we will never use a safety critical Linux or have no demand on this. It would be all interesting input because Yeah, so far we look in the IC650 or 8 and to ISO 26262 mainly as safety standards by no industrial also uses Understanders you may most likely share the same requirements and maybe not aware Maybe some things are different, but this is interesting typically requirements traceability design Implementation traceable use all the elephant all are the same and If you do a lot of things this helps at least so we also want to reach out to more Industrial actually the systems work group was intended to be industrial work group, but we put it on a wider footprint because there was not the quarrel and not the Yet the all we should say oh, yeah, let's have drivers from the industry industrial and support Who's already using Linux already? from the industrial space Yeah, it's much less so Maybe it's a good chance to give a little extra chance Anyway comes up with a question One in front Cool talk. Thank you. And you said you're generating s bombs. Yes. It's that using like Yachto's features Yeah, it's using the as pedigts inherits Or inherited as pedigts and as pedigts pretty and the report is on the git lab so you can all download the It on a daily base That's something which I haven't mentioned We delete the images every day, but you can just go to like the artifacts was in the git lab pain There's something to just download we keep it fresh So you get a fresh release every day So as soon as a new feature and you functionally gets in it thing You will experience and we also rebuilt Basically from scratch once a week if I'm correct that we also do the s-state. We don't do Every night. Yeah, we built every night But then we use the s-state and we cleaner the s-state the shared state every week because we said well Maybe something happens. So we want to also make sure it works still and The benefit of this we could already see and this may be nice to see We our CI did not broke actually but we did was detecting errors twice and One I think the one we've put back to a GL as an error and the second one was also Interesting because what everything was building the system was booting just the danger sign didn't show up because the certificate expired for the can stack This is something to keep in mind and this is something which you can experience This is the I am where it also makes sense to we come to a booting state doing test cases and we do this every night Then and it's not at all a safety critical Work which we have done or can be judged as safety critical at short show best practices on how to do it There's one also in the back and one and there Okay. Thank you for your talk You only asked about industrial. What about home and building automation? It's an interesting topic, right? I guess for home and building automation. They're also the hobby projects in there, right? Actually fire safety. Yeah, it's our industry fire safety. Yeah. Yes Yeah, but I guess I don't do you make use of Linux already? Yes. We use your go. Oh, cool And you're do you have certain you but do you have use really safe the integrity center? Or do you have actually also safety standard which tell you what? Conditions your system has to work on I know this for twice for example, they say don't work basically both we have like our own European standard the EN 54 but for on an international escape. We also do safety integrity level So for example for fire safety, usually you have sills we so we need to comply Yeah, and you comply with this work completely worse Linux or you have additional elements in the mind currently we have like also other elements which are compliant and Linux yachter Linux is just basically our Cloud connectivity right by ourselves, but it would be nice to have like everything on Linux Yeah, I think it's it and this also shows like there's a way of You have a heterogeneous system with other things involved and linux is part of it So it shows basically what we draw with the architecture and maybe also how you can get more responsibility to the system Like say if you enhance feature reduce complexity by getting certain things out There so it's a very good use case if you want to discuss more Feel free. I guess I was another one a little bit more back. I Don't know where was yeah, I can Comment on the previous one for homes and passage systems I've done work with the leading company for in Sweden for four years and Made them start the jocto and they've done a great job but for industry I work with I Won't mention it, but How they didn't want any security at all they didn't even want encryption on the devices Because it wasn't connected to the internet. How do you make them realize that? I guess here. It's once at least one statement is we need to split between security and the safety part right because the security is the Like intrusion part The safety part goes more on where it could cause harm to people Still I guess So within our use case, it's more like the people are concerned that Linux is not safe enough yet to go and take over certain Responsibility especially for complex use cases, but I think you bring a very good point because this is something where you also need to Consider how your environment change and just to say well I don't care about all these kind of things because it's not in my primary use case You never know when someone connect the USB printer to the router and it gets suddenly and post to the internet because you just have an open Connection in the router because there's some firmware mismatch and so when you get Suddenly your spam may it comes out of the printer. I get there are some prominent examples if you shut the web So this is something to consider as well And I guess it's just making it public making things visible and point to these kind of things and say there are Examples which show exactly this and that's where we came from right ten years back when we did our first device in infotainment and from Bosch site there was not that high view on Security and we were not connected at all. It was just the device But someone put the device out found the pins hacked the whole thing and was sharing this report on github to show Okay here. I can get into the root shell. I Said very good. You had it a lot of effort actually cannot do this because all the relevant functionalities hooked with an Artus and as soon as you change and flush a new Linux, it doesn't work anymore, but that's another story And I have a meeting on Tuesday with the automotive Part company it's not Yeah, it's automotive and they already signaled that we don't need Safety or Cutting corners everyone So It really heavily depends also on the use case, right? So it depends which device you do and if you if you see our infotainment devices which have a review camera in there But normally you could say well the review camera is something safety critical because you need to see that the camera still operates If a person is standing in the back of the car a kid a child or whatever you could cause harm to the person Damaged a person Not only damage your car. That's just I guess is the least problem by what then come in this like argumentation This is not safety relevant for some people building it because they just make the stating I know a few have ever seen this some of you. Please always look in the mirror and Don't trust the camera. This is basically there to say I Have less responsibility for safety still even if it's mentioned like this It reduces the safety level but the quality which you have to do still there So for example, we build and mechanisms to still identify that the camera is working there to do what it should not with a 100 point zero zero zero percentage But to a sufficient amount and adding this as a safety belt like additional safety net for us to give just this information and so there could be even Some argumentation from other manufacturers say yeah the safety relevant others say it's not safe to run heavily depends on the use case It's really heavy equipment, especially the first way so it's It should be Critical if something goes wrong. Yeah Okay, we are the last minutes. We would have chance for one more question otherwise Which is close a session Thanks a lot to all of you staying until the end