1,793,367 events recorded by gharchive.org of which 1,793,367 were push events containing 2,601,845 commit messages that amount to 191,925,722 characters filtered with words.py@e23d022007... to these 40 messages:
[READY] [KC13] Showing "The Derelict" some love: General updates, aesthetic changes and misc (#67696)
With this PR I aim to make KC13 (TheDerelict.dmm), or Russian Station (whatever you guys call it) a tad bit more flavorful with its environment as well as somethings on the mapping side (like adding area icons!). To preface, no, I'm not remapping anything here extensively. The general layout should be relatively the same (or should be in theory).
Halfway through naming the area icons I checked the wiki page and found out it was KC not KS, so, its KS13 internally.
Readability for turf icons are cool. Also just making the ruin more eye appealing would be better. General cleanup and changes will give new life to this rather.. loved? Hated? Loot pinata? Ruin. The ruin also now starts completely depowered, like Old Station (its a Derelict, it makes no sense for it to still be powered after so long). As for some mild compensation, a few more batteries were sprinkled in to offset any issues. If there is any concern of "But they'll open the vault faster!", there were always 5 batteries that people used to make the vault SMES. Lastly, giving it some "visual story telling" is cool, as mapping fluff goes.
I also added a subtle OOC hint that the SMES in the northern most solar room needs a terminal with the following:
SMES Jumpscare As an aside, I aim to try and keep the feel of this ruin being "dated" while at the same time having some of our newer things. With that, certain things I'll opt out of using in favor of more "generic" structures to give KC13 that true "Its old but not really" feel and look.
fix my stupid ass fucking code so other people's pc won't expolde
[clang] Implement ElaboratedType sugaring for types written bare
Without this patch, clang will not wrap in an ElaboratedType node types written without a keyword and nested name qualifier, which goes against the intent that we should produce an AST which retains enough details to recover how things are written.
The lack of this sugar is incompatible with the intent of the type printer default policy, which is to print types as written, but to fall back and print them fully qualified when they are desugared.
An ElaboratedTypeLoc without keyword / NNS uses no storage by itself, but still requires pointer alignment due to pre-existing bug in the TypeLoc buffer handling.
Troubleshooting list to deal with any breakage seen with this patch:
-
The most likely effect one would see by this patch is a change in how a type is printed. The type printer will, by design and default, print types as written. There are customization options there, but not that many, and they mainly apply to how to print a type that we somehow failed to track how it was written. This patch fixes a problem where we failed to distinguish between a type that was written without any elaborated-type qualifiers, such as a 'struct'/'class' tags and name spacifiers such as 'std::', and one that has been stripped of any 'metadata' that identifies such, the so called canonical types. Example:
namespace foo { struct A {}; A a; };If one were to print the type of
foo::a, prior to this patch, this would result infoo::A. This is how the type printer would have, by default, printed the canonical type of A as well. As soon as you add any name qualifiers to A, the type printer would suddenly start accurately printing the type as written. This patch will make it print it accurately even when written without qualifiers, so we will just printAfor the initial example, as the user did not really write thatfoo::namespace qualifier. -
This patch could expose a bug in some AST matcher. Matching types is harder to get right when there is sugar involved. For example, if you want to match a type against being a pointer to some type A, then you have to account for getting a type that is sugar for a pointer to A, or being a pointer to sugar to A, or both! Usually you would get the second part wrong, and this would work for a very simple test where you don't use any name qualifiers, but you would discover is broken when you do. The usual fix is to either use the matcher which strips sugar, which is annoying to use as for example if you match an N level pointer, you have to put N+1 such matchers in there, beginning to end and between all those levels. But in a lot of cases, if the property you want to match is present in the canonical type, it's easier and faster to just match on that... This goes with what is said in 1), if you want to match against the name of a type, and you want the name string to be something stable, perhaps matching on the name of the canonical type is the better choice.
-
This patch could exposed a bug in how you get the source range of some TypeLoc. For some reason, a lot of code is using getLocalSourceRange(), which only looks at the given TypeLoc node. This patch introduces a new, and more common TypeLoc node which contains no source locations on itself. This is not an inovation here, and some other, more rare TypeLoc nodes could also have this property, but if you use getLocalSourceRange on them, it's not going to return any valid locations, because it doesn't have any. The right fix here is to always use getSourceRange() or getBeginLoc/getEndLoc which will dive into the inner TypeLoc to get the source range if it doesn't find it on the top level one. You can use getLocalSourceRange if you are really into micro-optimizations and you have some outside knowledge that the TypeLocs you are dealing with will always include some source location.
-
Exposed a bug somewhere in the use of the normal clang type class API, where you have some type, you want to see if that type is some particular kind, you try a
dyn_castsuch asdyn_cast<TypedefType>and that fails because now you have an ElaboratedType which has a TypeDefType inside of it, which is what you wanted to match. Again, like 2), this would usually have been tested poorly with some simple tests with no qualifications, and would have been broken had there been any other kind of type sugar, be it an ElaboratedType or a TemplateSpecializationType or a SubstTemplateParmType. The usual fix here is to usegetAsinstead ofdyn_cast, which will look deeper into the type. Or usegetAsAdjustedwhen dealing with TypeLocs. For some reason the API is inconsistent there and on TypeLocs getAs behaves like a dyn_cast. -
It could be a bug in this patch perhaps.
Let me know if you need any help!
Signed-off-by: Matheus Izvekov mizvekov@gmail.com
Differential Revision: https://reviews.llvm.org/D112374
Merge #82440
82440: admission,storage: introduce flush tokens to constrain write admission r=tbg,irfansharif a=sumeerbhola
In addition to byte tokens for writes computed based on compaction rate out of L0, we now compute byte tokens based on how fast the system can flush memtables into L0. The motivation is that writing to the memtable, or creating memtables faster than the system can flush results in write stalls due to memtables, that create a latency hiccup for all write traffic. We have observed write stalls that lasted > 100ms.
The approach taken here for flush tokens is straightforward (there is justification based on experiments, mentioned in code comments):
- Measure and smooth the peak rate that the flush loop can operate on. This relies on the recently added pebble.InternalIntervalMetrics.
- The peak rate causes 100% utilization of the single flush thread,
and that is potentially too high to prevent write stalls (depending
on how long it takes to do a single flush). So we multiply the
smoothed peak rate by a utilization-target-fraction which is
dynamically adjusted and by default is constrained to the interval
[0.5, 1.5]. There is additive increase and decrease of this
fraction:
- High usage of tokens and no write stalls cause an additive increase
- Write stalls cause an additive decrease. A small multiplier is used if there are multiple write stalls, so that the probing falls more in the region where there are no write stalls.
Note that this probing scheme cannot eliminate all write stalls. For now we are ok with a reduction in write stalls.
For convenience, and some additional justification mentioned in a code comment, the scheme uses the minimum of the flush and compaction tokens for writes to L0. This means that sstable ingestion into L0 is also subject to such tokens. The periodic token computation continues to be done at 15s intervals. However, instead of giving out these tokens at 1s intervals, we now give them out at 250ms intervals. This is to reduce the burstiness, since that can cause write stalls.
There is a new metric, storage.write-stall-nanos, that measures the cumulative duration of write stalls, since it gives a more intuitive feel for how the system is behaving, compared to a write stall count.
The scheme can be disabled by increasing the cluster setting admission.min_flush_util_fraction, which defaults to 0.5 (corresponding to the 0.5 lower bound mentioned earluer), to a high value, say 10.
The scheme was evaluated using a single node cluster with the node having a high CPU count, such that CPU was not a bottleneck, even with max compaction concurrency set to 8. A kv0 workload with high concurrency and 4KB writes was used to overload the store. Due to the high compaction concurrency, L0 stayed below the unhealthy thresholds, and the resource bottleneck became the total bandwidth provisioned for the disk. This setup was evaluated under both:
- early-life: when the store had 10-20GB of data, when the compaction backlog was not very heavy, so there was less queueing for the limited disk bandwidth (it was still usually saturated).
- later-life: when the store had around 150GB of data.
In both cases, turning off flush tokens increased the duration of write stalls by > 5x. For the early-life case, ~750ms per second was spent in a write stall with flush-tokens off. The later-life case had ~200ms per second of write stalls with flush-tokens off. The lower value of the latter is paradoxically due to the worse bandwidth saturation: fsync latency rose from 2-4ms with flush-tokens on, to 11-20ms with flush-tokens off. This increase imposed a natural backpressure on writes due to the need to sync the WAL. In contrast the fsync latency was low in the early-life case, though it did increase from 0.125ms to 0.25ms when flush-tokens were turned off.
In both cases, the admission throughput did not increase when turning off flush-tokens. That is, the system cannot sustain more throughput, but by turning on flush tokens, we shift queueing from the disk layer the admission control layer (where we have the capability to reorder work).
Screenshots for early-life: Flush tokens were turned off at 22:32:30. Prior to that the flush utilization-target-fraction was 0.625.
Screenshots for later-life: Flush tokens were turned off at 22:03:20. Prior to that the flush utilization-target-fraction was 0.875.
These results were produced by running
roachprod create -n 2 --clouds aws --aws-machine-type=c5d.9xlarge --local-ssd=false --aws-ebs-volume-type=gp2 sumeer-io
roachprod put sumeer-io:1 cockroach ./cockroach
roachprod put sumeer-io:2 workload ./workload
roachprod start sumeer-io --env "COCKROACH_ROCKSDB_CONCURRENCY=8"
roachprod run sumeer-io:2 -- ./workload run kv --init --histograms=perf/stats.json --concurrency=1024 --splits=1000 --duration=30m0s --read-percent=0 --min-block-bytes=4096 --max-block-bytes=4096 {pgurl:1-1}
Fixes #77357
Release note (ops change): Write tokens are now also limited based on flush throughput, so as to reduce storage layer write stalls. This behavior is enabled by default. The cluster setting admission.min_flush_util_fraction, defaulting to 0.5, can be used to disable or tune flush throughput based admission tokens, for writes to a store. Setting to a value much greater than 1, say 10, will disable flush based tokens.
Co-authored-by: sumeerbhola sumeer@cockroachlabs.com
Cancel fixed, filecrypt notation added, some other shit, fuck dis bitch im out.
Fix grammatical mistakes and style (DE)
Explanations for some of the corrections:
The dash is wrong between two words when the first word has a 'Fugen-S' which implies it to be genitive. Thus I removed it.
'URL' as in 'Website-URL' can have either female or male grammatical gender as per the Duden. Male, however, is far more uncommon, thus the change of the article from 'Der' to 'Die'.
The apostroph is not necessary in the imperative when there's no letter missing.
'zum Bankkonto gehörig' is a participle construction which makes it generally a bit hard to read and understand. The genitive will already be enough to show posessiveness.
Attempts on an InteractableTile event workflow
...It didn't work with my current understanding of Unity's tilemaps. My problem is, how can I verify what Interactable Tile a listener is listening for. Normally, we could send this using an Observer Pattern to some script, but I have no way of sending an event from a SPECIFIC tile on the tilemap to some listeners because I cannot check for who sent the event. Because my current system has InteractableTiles firing off events from the LevelTileGO class on an instantiated RUNTIME-ONLY prefab, I have no way of knowing of it without hard-coding and magic numbers - all things I HATE!
I'm thinking of separating this idea from the RuleTile instantiated prefab, and creating a separate set of interactable objects in the editor, per-map, that we can use and hook up to one another...
I'll have to see if this is really the best approach though.
🐫
5 reasons to hire a professional plumbing company in Perth
Plumbing is undeniably one of the most important systems of your home that needs to function optimally for smooth functioning inside the home. Whether you are planning a home renovation or plumbing maintenance and repair, hiring a professional plumbing company in Perth is important to save your money and time. In this article, we will tell you about the five reasons to hire professional plumbing company services. Plumbers know their job Plumbers in Plumbing Company in Perth are the ones who have got specific training, knowledge, and practical experience regarding plumbing. From minor issues like repairing leaks to other major issues, they are trained to deal with these issues. They are licensed and have the required experience and skills to know they are precisely doing the job and get you the permanent results. Plumbers have the right tools and materials Many people at home think that they can easily handle the job of a plumber by following different DIY methods. But they might not be aware of the fact that plumbers are the ones who have all the tools from major to minor that are required to perform the job seamlessly. They save a lot of your time in running back and forth from different hardware stores to buy different tools that you may even not probably use ever again! The plumbing company covers multiple services The plumbing company offers a wide variety of services that can help you to install, repair, or maintain your plumbing system easily. Whether it's your hot water system, drain camera inspection, block drainage, water leaks, water filter supplies, or even bathroom or renovation, plumbing services cover all of it under them. So you don't have to worry about it and run here and there for one separate service. Time-saving In our daily busy lifestyle, we hardly have time to do the plumbing job by ourselves. Juggling between so many responsibilities like handling a job, work, family, and hobbies we hardly get time to do such kind of detailed and hard work as plumbing. Thus in such situations, it's always best to hire plumbing company services that will help you to save a lot of time and hassles on your part. High-quality work Professional Perth Plumber is trained in such a way that he can provide you with a high degree of work. They will address all the issues regarding your plumbing and even the issues that you are not able to see but are there in your plumbing system. They will recognize all those loopholes and with the help of their skills and the latest equipment, they will give you the best services possible. Gives you a permanent solution Hiring professional plumbers will help you in giving a permanent solution to all your plumbing problems. The repair or the installation that a professional plumber will do for you will give you a permanent solution. And it is not a temporary fix that you have to get rectify again and again. So if you are changing your home or renovating it or renovating specific rooms of your home like the kitchen and bathroom then it's best always best to have professional plumbers. Which is the best plumbing company in Perth? We are sure that after knowing the benefits of professional plumbing services you must be considering hiring one. But which company to choose? As there are already a lot of services present in the market for plumbing. Well, the answer to this question is straightforward as NDA Plumbing and Gas services are one of the leading plumbing service providers in Perth. We cover a wide range of services ranging from backflow devices and testing to commercial and domestic plumbing services. Apart from plumbing we also cover gas services. And to be rest assured of our services you can easily check our testimonials on our website or even on the internet and a lot of clients are loving our work and we will sure that you will love it too! So what's the wait for? Visit our website now! For more Information Click on the Link: https://ndaplumbingperth.com.au/
Phone Number: 0478 757 622 For any Query Contact Us: info@ndaplumbingperth.com.au Address: 23 Gingerale Cir, Byford WA 6122, Australia.
power: Introduce OnePlus 3 fingerprintd thaw hack Taken from Oneplus 3, this hack will make fingerprintd recover from suspend quickly.
Small fixes for newer kernels since we're coming from 3.10.108..
Change-Id: I0166e82d51a07439d15b41dbc03d7e751bfa783b Co-authored-by: Cyber Knight cyberknight755@gmail.com [cyberknight777: forwardport and adapt to 4.14] Signed-off-by: Shreyansh Lodha slodha96@gmail.com Signed-off-by: Pierre2324 pdbbasketball@gmail.com Signed-off-by: PainKiller3 ninadpatil100@gmail.com Signed-off-by: Dhruv dhruvgera61@gmail.com Signed-off-by: Cyber Knight cyberknight755@gmail.com Signed-off-by: officialputuid officialputuid@hack.id
parallel.h refactoring and add support for TBB
Hide nearly all of the implementation of the parallel_for family of methods rather than expose them as inline. This gives us more freedom to change the implementation in the future without breaking ABI.
Deprecate the parallel methods that take task functions whose first parameter is the thread ID. The only reason they existed is because our cobbled together thread_pool used that in its inner workings. But there are good reasons why we should not expose that:
-
We never used it anyway.
-
It was not very useful, since sometimes it was a real thread ID, but other times it was -1 in cases where the calling thread executed the task directly.
-
It is inconsistent with other thread pool implementations that we may wish to try in the future.
So better just to not expose those thread IDs at all. Mark those versions of the parallel loops as deprecated and schedule them for removal in OIIO 3.0.
If TBB is detected at build time, support will be built in that allows either the old OIIO built-in pool, or TBB, depending on the setting of a new global OIIO attribute, "use_tbb" (default 0), if set to nonzero will use the TBB thread pool.
In my experiments, the TBB thread pool seems slightly better -- I think because there is less overhead, and maybe the clever work-stealing it does elps to load balance. It's not used by default, set the attribute if you want to use it (assuming the build included. After we've had a chance to test more thoroughly, we may change to use TBB by default. I'm interested to hear people's thoughts on the matter.
One case where you almost certainly will want to use the TBB thread pool is if you're using libOpenImageIO from within an application that uses TBB for its own threding -- that way you're using one pool for everything, rather than have two separate thread pools that don't know about each other (and probably twice as many threads as you have cores)..
feat: Add support for fields, attr_* and funcs in method calls (#21)
This patch rewrites the handling for aliases to largely consistently handle classes etc., instead of having one-off hacks. (TBF, I didn't fully understand how aliases worked earlier.)
Thanks to this, constants also seem to mostly work now.
Additionally, adding support for attr_* forms required that we support emitting references for functions in method invocations properly. So I fixed that too. The earlier implementation of trying recursive lookup seems very silly in hindsight. 🤦🏽
However thanks to this change, we also get support for occurrences from the stdlib for "free."
i386 testsuite: cope with --enable-default-pie
Running the testsuite on a toolchain build with --enable-default-pie had some unexpected fails. Adjust the tests to tolerate the effects of this configuration option on x86_64-linux-gnu and i686-linux-gnu.
The cet-sjlj* tests get offsets before the base symbol name with PIC or PIE. A single pattern covering both alternatives somehow triggered two matches rather than the single expected match, thus my narrowing the '.*' to not skip line breaks, but that was not enough. Still puzzled, I separated the patterns into nonpic and !nonpic, and we get the expected matchcounts this way.
Tests for -mfentry require an mfentry effective target, which excludes 32-bit x86 with PIC or PIE enabled, that's why the patterns that accept the PIC sym@RELOC annotations only cover x86_64. mvc7 is getting regexps extended to cover PIC reloc annotatios and all of the named variants, and tightened to avoid unexpected '.' matches.
The pr24414 test stores in an unadorned named variable in an old-style asm statement, to check that such asm statements get an implicit memory clobber. Rewriting the asm into a GCC extended asm with the variable as an output would remove the regression it checks against. Problem is, the literal reference to the variable is not PIC, so it's rejected by the elf64 linker with an error, and flagged with a warning by the elf32 one. We could presumably make the variable references PIC-friendly with #ifdefs, but I doubt that's worth the trouble. I'm just arranging for the test to be skipped if PIC or PIE are enabled by default.
for gcc/testsuite/ChangeLog
* gcc.target/i386/cet-sjlj-6a.c: Cope with --enable-default-pie.
* gcc.target/i386/cet-sjlj-6b.c: Likewise.
* gcc.target/i386/fentryname3.c: Likewise.
* gcc.target/i386/mvc7.c: Likewise.
* gcc.target/i386/pr24414.c: Likewise.
* gcc.target/i386/pr93492-3.c: Likewise.
* gcc.target/i386/pr93492-5.c: Likewise.
* gcc.target/i386/pr98482-1.c: Likewise.
drm: Handle dim for udfps
-
Apparently, los fod impl is better than udfps cuz it has onShow/HideFodView hook, which allows us to toggle dimlayer seamlessly.
Since udfps only partially supports the former one, we'd better kill dim in kernel. This is kinda a hack but it works well, bringing perfect fod experience back to us.
Co-authored-by: Art_Chen Chenxy0201@qq.com Signed-off-by: alk3pInjection webmaster@raspii.tech Change-Id: I80bfd508dacac5db89f4fff0283529c256fb30ce
Update on onionsan's "y'hear" sprites under Doru's request
So I (19M) was sitting in the Living room watching Morbius. My parents(senior citizens) visited to come check on me. They caught me watching Morbius and they got very upset because I have been through phases like this before. I convinced my mother (55F) that Morbius is great, but my father (56M) disagreed. So I yelled ITS MORBIN TIME and may have beat the living shit out my father. My mother(idiotic) then said YOU DID THIS LAST TIME DURING YOUR NFT PHASE so then I plunged her to the depths of Tartarus. My sister (19F) and her friend (18F) saw me and then tried to run so I yelled OH NO YOU MORB'T and morbed so much they died. I at this point, had almost become God. I had blown up my house and set on fire to get rid of the evidence. At this point I, (20M) have been running from the CIA, the FBI, the police, the Secret Service, Interpol, the UN security unit, Disney S&P, and Google Maps. I have almost reached the final phase of morb. Soon my original being will die off and I will morb in the final morb. I morbing regret morbing Morbius. Oh morb. Im gonna morbing morb myself. Morb this. Morb it all. TL;DR I killed my dad over a movie and ended up internationally wanted.
apollo: Enable thermal-engine service
- fuck you xiaomi
Signed-off-by: Tushar Mahajan mahajant99@gmail.com
SIG Monitoring at 11 instead of 12
The current time (12:05) for SIG Monitoing is not optimal:
- Putting the Meeting back-to-back with the hacking session creates one big block with no interlude to regenerate focus. This is problematic, especially in connection with the following:
- 12:00 is regular lunch time for many people working at the office or coworking spaces. So by missing lunch by an hour the focus of participants suffers.
Personal touch:
- The lack of "social regeneration time" poses a problem for me personally as a neurodivergent person. I think meeting at 11:05 and then having some time for myself would benefit my ability to participate in the hacking session starting at 13:00 in a focused way.
I am sure other people have additional reasons, be they based in time constraints and/or other personal preferences and necessities.
Thus, I would like to propose changing the time from 12:05 to 11:00 or 11:05.
Signed-off-by: linwalth 30180336+linwalth@users.noreply.github.com
"8:15am. My stress is pretty high over this, so I am getting up early, but that is not bad. I don't dislike stress. At least it is better than boredom. Yesterday, I said I wanted a challenge in BG2, but I can't focus on it at all. I am just not into it. I miss when I was a kid and I could immerse myself properly. Any mail?
Thanks so much for applying to Generally Intelligent! We would love to set up a 30 min chat to learn more about you and tell you about our company. Would you mind finding some time with me here? Let me know if you have a hard time finding a slot that works.
Kek. It is just what I thought would happen. Of course, I'll do the interview.
Also TK is offering 33% equity and my desired salary when they raise seed money in 2 months. This is a good offer, but cultivating a company from the ground up is not what I am into, but cold hard cash. At this point I have both trading skills and a long term outlook needed to make proper use of it.
33% equity does mean being a huge part of the company.
Also this is risky. Since he said nothing about how much he is offering now, I can only infer that he is trying to hook me with future promises, but does not have much to offer at the moment.
8:20am. Anyway, yesterday's cloud videos as basic as they are have given me inspiration.
Let me check out if it is real...
https://aws.amazon.com/braket/
Amazon Braket is a fully managed quantum computing service designed to help speed up scientific research and software development for quantum computing.
Yeah, it is real. This nails it for me. If Amazon has quantum computing services, then it will definitely have neurochips in the future.
Rather than focus on getting my hands on just one of them, why not a 1,000? But if it is a 1,000 I am not going to be putting them into my house. I already hate training on my personal GPU. It is not that I am that hung up about owning it personally.
I need to aim big. Cloud services might be expensive to train on, but if I have a good model, it should be able to cover the 0.1/h cost of a single instance.
Current hardware is not appropritate, but if there are big memristor breakthroughs in the future, I can be that big cloud providers will have priority on getting that over myself.
https://semianalysis.substack.com/p/cxl-enables-microsoft-azure-to-cut
Datacenters are an incredibly expensive affair. Microsoft stated that up to 50% of their server costs is from DRAM alone.
I actually wasn't aware up to now how much DRAM cost them. In a PC DRAM is not that expensive, but in a datacenter things are different. Non volatile memory would change this.
9:15am. Anyway, I am going to turn TK down.
My preferences are job of >150k and if I can't get those I'll consider >100k. If I can't get that then I'd consider taking risks and developing a startup such as his.
Damn it, I want to take a look at the GI page on Glassdoor so I can see its salary, but it is asking me to leave a review.
https://angel.co/company/generally-intelligent
It has an AngelList page. Yeah, the compensation of 140-200k is quite good for remote work. This is my preference. It says the company size is only 1-10 people though. But that does not matter to me too much, hard compensation does since I will be able to convert that into trading capital. This would not be an option for ordinary programmers, they'd be liable to gamble it away on Bitcoin, but I have a system now.
9:20am. Let me take a short break and then I will get started. First I'll reply to TK telling him as such and then focus on the AWS course. I want to get through it today. Actually before that I'll also grab a slot with GI.
9:50am. I am back. I've had more time to think about it. I won't touch the GI interview slots until I feel ready.
If I go there, my job will definitely be to take the ML researcher's models, optimize and then distribute them on the cloud. Knowing how to do that as well as use multi GPU setups will be essential, so I want to make sure I am done with my tour of AWS before that.
Having my trick work essentially validates my view, but I am just a tad short to enter the field at the moment. I can afford to spend a few weeks studying.
I put some thought into your offer and it is not too bad, but my preference is to look for a paid job in the >150k range, and failing that at least >100k. If that is not working out for me, I'd be open to taking risks with a startup like yours instead of accepting low pay. It is not impossible that I'd work with you at some point, but not at the moment. Once again, thank you for reaching out to me.
Diplomacy check pass.
10am. Now let me focus on skilling up. Over the next few days I should get replies from the other places I applied to. When I feel ready, I'll grab the slots and try going through the challenges.
I am going to go through this today. I do not want to waste time.
10:10am. AWS has a service where it can install an outpost directly inside one's own building. It would be owned and operated by the AWS, but the customer would have the physical proximity he needs.
This video is useful as I wouldn't even be aware that such options exist otherwise. I need to skill up, especially if I am going to be working at small startups where I'd need to do everything on my own.
I was wrong - instead of being obsessed about AI chips, I should be obsessed about the cloud. If I made a mistake last year, it would be this.
10:15am. > Lesson 17 - In AWS everything is an API call.
This is what I'd expect at any cloud provider. You wouldn't call up the rep to reserve an instance, instead you'd use an API to provision resources.
AWS Management Console is likely the first place you will go when you are learning about AWS.
I guess that is where I'll start then.
10:35am. Let me go through module 4.
10:45am. Let me make little change to the C backend. I realized that the way I am doing string lengths is wrong.
| TyArrayLength(_,b) -> return' $"{tup_data b}->len"
| TyStringLength(_,b) -> return' $"{tup_data b}->len-1"Because of that null char, the string length will be 1 larger than I'd prefer. Let me make the change.
Oh, right. I forgot to update the changelog.
11am. Published the 2.2.1 version and even tested it. Let me commit here and I will go back to the tutorial. I want to finish it today."
[MDB Ignore] Shifts all (sane) varedited signs to directionals (#68004)
- [MDB Ignore] Shifts all (sane) varedited signs to directionals
Hey there,
So we have these cool new sign directionals now, but we still have all of the old pixel-shifted pre-fabrications lying around. So, I added an UpdatePaths (as well as Updated the Paths) to be a bit better at using directionals, because directionals are pretty neato.
This should update every single var_edit that used the proper 32 pixelshift (some of them used 28, and I'm unable to account for that automatically with current tooling) into a proper subtype. Mappers tend to learn by looking at well established maps, so it's always important to ensure that the well-established maps use the most recent tooling (i.e.: bring them up to the surface) and avoid needless excess lines in maps.
- The Commit With All The Maps
OH GOD OH FUCK
- Renames the UpdatePaths
kmod: add init function to usermodehelper
About 6 months ago, I made a set of changes to how the core-dump-to-a-pipe feature in the kernel works. We had reports of several races, including some reports of apps bypassing our recursion check so that a process that was forked as part of a core_pattern setup could infinitely crash and refork until the system crashed.
We fixed those by improving our recursion checks. The new check basically refuses to fork a process if its core limit is zero, which works well.
Unfortunately, I've been getting grief from maintainer of user space programs that are inserted as the forked process of core_pattern. They contend that in order for their programs (such as abrt and apport) to work, all the running processes in a system must have their core limits set to a non-zero value, to which I say 'yes'. I did this by design, and think thats the right way to do things.
But I've been asked to ease this burden on user space enough times that I thought I would take a look at it. The first suggestion was to make the recursion check fail on a non-zero 'special' number, like one. That way the core collector process could set its core size ulimit to 1, and enable the kernel's recursion detection. This isn't a bad idea on the surface, but I don't like it since its opt-in, in that if a program like abrt or apport has a bug and fails to set such a core limit, we're left with a recursively crashing system again.
So I've come up with this. What I've done is modify the call_usermodehelper api such that an extra parameter is added, a function pointer which will be called by the user helper task, after it forks, but before it exec's the required process. This will give the caller the opportunity to get a call back in the processes context, allowing it to do whatever it needs to to the process in the kernel prior to exec-ing the user space code. In the case of do_coredump, this callback is ues to set the core ulimit of the helper process to 1. This elimnates the opt-in problem that I had above, as it allows the ulimit for core sizes to be set to the value of 1, which is what the recursion check looks for in do_coredump.
This patch:
Create new function call_usermodehelper_fns() and allow it to assign both an init and cleanup function, as we'll as arbitrary data.
The init function is called from the context of the forked process and allows for customization of the helper process prior to calling exec. Its return code gates the continuation of the process, or causes its exit. Also add an arbitrary data pointer to the subprocess_info struct allowing for data to be passed from the caller to the new process, and the subsequent cleanup process
Also, use this patch to cleanup the cleanup function. It currently takes an argp and envp pointer for freeing, which is ugly. Lets instead just make the subprocess_info structure public, and pass that to the cleanup and init routines
Signed-off-by: Neil Horman nhorman@tuxdriver.com Reviewed-by: Oleg Nesterov oleg@redhat.com Cc: Andi Kleen andi@firstfloor.org Signed-off-by: Andrew Morton akpm@linux-foundation.org Signed-off-by: Linus Torvalds torvalds@linux-foundation.org
Add SPM plugin
SPM is providing new capabilities for tools to expose themselves as plugins which allows them to generate build commands. This allows us to create a plugin for SwiftProtobuf which invokes the protoc compiler and generates the Swift code. Creating such a plugin is greatly wanted since it improves the usage of the protoc-gen-swift plugin dramatically. Fixes apple/swift-protobuf#1207
This PR adds a new SPM plugin which generates build commands for generating Swift files from proto files. Since users of the plugin might have complex setups, I introduced a new swift-protobuf-config.json file that adopters have to put into the root of their target which wants to use the new plugin. The format of this configuration file is quite simple:
{
"invocations": [
{
"protoFiles": [
"Foo.proto"
],
"visibility": "internal"
}
]
}
It allows you to configure multiple invocations to the protoc compiler. For each invocation you have to pass the relative path from the target source to the proto file. Additionally, you have to decide which visibility the generated symbols should have. In my opinion, this configuration files gives us the most flexibility and more complex setups to be covered as well.
# Open topics
Hosting of the protoc binary is the last open thing to figure out before we can release a plugin for SwiftProtobuf. From my point of view, there are three possible routes we can take:
- Include the
artifactbundleinside theSwiftProtobufrepository - Include the
artifactebundleas an artifact on the GH releases in the protobuf repo - Extend the the artifact bundle manifest to allow inclusion of URLs. This would require a Swift evolution pitch most likely.
However, with all three of the above we would still need to release a new version of SwiftProtobuf with every new release of protoc.
With the current shape of the PR one can already use dependencies between proto files inside a single target. However, it might be desirable to be able to build dependency chains of Swift targets where each target includes proto files which depend on protoc files from the dependencies of the Swift target. I punted this from the initial plugin because this requires a bit more work and thinking. Firstly, how would you even spell such an import? Secondly, the current way of doing ProtoPathModuleMapping files is not super ideal for this approach. It might make sense to introduce a proto option to set the Swift module name inside the proto files already.
We now have a SPM plugin that can generate Swift code from proto files. To use it, it provides a configuration file format to declare different invocations to the protoc compiler.
macsec: avoid heap overflow in skb_to_sgvec
While this may appear as a humdrum one line change, it's actually quite important. An sk_buff stores data in three places:
- A linear chunk of allocated memory in skb->data. This is the easiest one to work with, but it precludes using scatterdata since the memory must be linear.
- The array skb_shinfo(skb)->frags, which is of maximum length MAX_SKB_FRAGS. This is nice for scattergather, since these fragments can point to different pages.
- skb_shinfo(skb)->frag_list, which is a pointer to another sk_buff, which in turn can have data in either (1) or (2).
The first two are rather easy to deal with, since they're of a fixed maximum length, while the third one is not, since there can be potentially limitless chains of fragments. Fortunately dealing with frag_list is opt-in for drivers, so drivers don't actually have to deal with this mess. For whatever reason, macsec decided it wanted pain, and so it explicitly specified NETIF_F_FRAGLIST.
Because dealing with (1), (2), and (3) is insane, most users of sk_buff doing any sort of crypto or paging operation calls a convenient function called skb_to_sgvec (which happens to be recursive if (3) is in use!). This takes a sk_buff as input, and writes into its output pointer an array of scattergather list items. Sometimes people like to declare a fixed size scattergather list on the stack; othertimes people like to allocate a fixed size scattergather list on the heap. However, if you're doing it in a fixed-size fashion, you really shouldn't be using NETIF_F_FRAGLIST too (unless you're also ensuring the sk_buff and its frag_list children arent't shared and then you check the number of fragments in total required.)
Macsec specifically does this:
size += sizeof(struct scatterlist) * (MAX_SKB_FRAGS + 1);
tmp = kmalloc(size, GFP_ATOMIC);
*sg = (struct scatterlist *)(tmp + sg_offset);
...
sg_init_table(sg, MAX_SKB_FRAGS + 1);
skb_to_sgvec(skb, sg, 0, skb->len);
Specifying MAX_SKB_FRAGS + 1 is the right answer usually, but not if you're using NETIF_F_FRAGLIST, in which case the call to skb_to_sgvec will overflow the heap, and disaster ensues.
Signed-off-by: Jason A. Donenfeld Jason@zx2c4.com Cc: stable@vger.kernel.org Cc: security@kernel.org Signed-off-by: David S. Miller davem@davemloft.net
update the shortcut-manager.sh to work better with journaling
Now, when I press Alt+n, a ton of things happen:
- The script checks what time of day it is Morning: 6 am - 12 pm Afternoon: 12 pm - 5 pm Evening: 5 pm - 10 pm Night: 10 pm - 6 am
- Then, it checks to see if we already wrote down a journal note for
the day type in the
note.texjournal file. If we didn't, it will output a template for us to fill in. Here's the morning template:Here's the afternoon template:\morning \begin{goals} \end{goals}Here's the evening template:\afternoon \begin{stats} \end{stats}Here's the night template:\evening \begin{stats} \end{stats}\night \begin{results} \end{results} \time{03:42:52 AM} \contentment{} \finalnote{} - Finally, it will just output
\time{03:42:52 AM}for example in thenote.texjournal file.
MGMT-10973: Enable DNS validations if coredns or keepalived static pod manifests in day-2 connectivity-check ignition
For context, this is a service-side follow-up to:
openshift/assisted-installer-agent#388
and also this small agent fix openshift/assisted-installer-agent#403
During day-2 installations, we want the service to optionally perform DNS validations when the worker attempts to join none-platform clusters.
When the cluster is imported via our .../v2/clusters/import endpoint,
we have no way to know whether that cluster is a none-platform cluster
or cluster with managed networking (e.g. baremetal), so we have no way
to know whether we should perform the DNS validation or not. We wouldn't
like to have that validation on all the time, because it's unnecessarily
annoying for clusters with managed networking.
As part of our existing API connectivity-check, the service asks the agent to download the worker.ign file from the to-be-joined-cluster's machine-config-server, then send that back to the service.
The contents of that file include information that will allow the service to make an educated guess about whether the ignition originated from a cluster with managed networking or not.
Also, a new "imported" column has been added to clusters, to indicate whether they were created through a conversion or through being imported. This is important because the solution should only be applied for imported clusters, and this will provide a good way to tell apart imported from non-imported clusters.
Also, when a user imports a cluster, we will inspect the params.NewImportClusterParams.APIVipDnsname
parameter and extract:
- The cluster name
- The cluster base domain
The cluster name will override the name given in params.NewImportClusterParams.Name,
see diff comment for more information on why.
The cluster base domain will be used to set the BaseDNSDomain of the
cluster, because up until now we didn't set it for imported cluster.
The reason BaseDNSDomain and Name have to be correctly set is
because the DNS validations use them to construct the domain names
that are being validated.
Also updated some validation messages for the API connectivity check and DNS validations.
Also, the clusterdeployments_controller can now import SNO clusters, it was an oversight that should have been done as part of 4cda26533d377f453f68783e8b2eae438734555d (#3404)
The ignition files we currently look at are:
/etc/kubernetes/manifests/coredns.yaml
/etc/kubernetes/manifests/keepalived.yaml
This is a hack - since we have no official way to know whether a worker ignition file originated from a cluster with managed networking or not, we instead rely on the presence of coredns and keepalived pod manifests to indicate that. We only expect those to be present in clusters with managed networking. To be a bit more robust, the service will consider the presence of any one of them to mean that the cluster has managed networking. This gives us better forwards compatibility if one of them gets renamed / replaced with other technologies in future OCP versions.
Another way in which this is hacky is that users could manually create static pods with the same name as part of their machine-configs, in which case we would have a false-positive detection. But that is admittedly very unlikely.
Hopefully we can negotiate with the relevant OCP teams to have a more official, stable way to have this detection - like a magic empty file placed somewhere in the ignition that we can check for the presence of. Once we have such file, we can slowly deprecate this detection mechanism and fully move to the new one by inspecting that file instead.
signal: add pidfd_send_signal() syscall
The kill() syscall operates on process identifiers (pid). After a process has exited its pid can be reused by another process. If a caller sends a signal to a reused pid it will end up signaling the wrong process. This issue has often surfaced and there has been a push to address this problem [1].
This patch uses file descriptors (fd) from proc/ as stable handles on struct pid. Even if a pid is recycled the handle will not change. The fd can be used to send signals to the process it refers to. Thus, the new syscall pidfd_send_signal() is introduced to solve this problem. Instead of pids it operates on process fds (pidfd).
/* prototype and argument /* long pidfd_send_signal(int pidfd, int sig, siginfo_t *info, unsigned int flags);
/* syscall number 424 */ The syscall number was chosen to be 424 to align with Arnd's rework in his y2038 to minimize merge conflicts (cf. [25]).
In addition to the pidfd and signal argument it takes an additional siginfo_t and flags argument. If the siginfo_t argument is NULL then pidfd_send_signal() is equivalent to kill(, ). If it is not NULL pidfd_send_signal() is equivalent to rt_sigqueueinfo(). The flags argument is added to allow for future extensions of this syscall. It currently needs to be passed as 0. Failing to do so will cause EINVAL.
/* pidfd_send_signal() replaces multiple pid-based syscalls */ The pidfd_send_signal() syscall currently takes on the job of rt_sigqueueinfo(2) and parts of the functionality of kill(2), Namely, when a positive pid is passed to kill(2). It will however be possible to also replace tgkill(2) and rt_tgsigqueueinfo(2) if this syscall is extended.
/* sending signals to threads (tid) and process groups (pgid) */ Specifically, the pidfd_send_signal() syscall does currently not operate on process groups or threads. This is left for future extensions. In order to extend the syscall to allow sending signal to threads and process groups appropriately named flags (e.g. PIDFD_TYPE_PGID, and PIDFD_TYPE_TID) should be added. This implies that the flags argument will determine what is signaled and not the file descriptor itself. Put in other words, grouping in this api is a property of the flags argument not a property of the file descriptor (cf. [13]). Clarification for this has been requested by Eric (cf. [19]). When appropriate extensions through the flags argument are added then pidfd_send_signal() can additionally replace the part of kill(2) which operates on process groups as well as the tgkill(2) and rt_tgsigqueueinfo(2) syscalls. How such an extension could be implemented has been very roughly sketched in [14], [15], and [16]. However, this should not be taken as a commitment to a particular implementation. There might be better ways to do it. Right now this is intentionally left out to keep this patchset as simple as possible (cf. [4]).
/* naming */ The syscall had various names throughout iterations of this patchset:
- procfd_signal()
- procfd_send_signal()
- taskfd_send_signal() In the last round of reviews it was pointed out that given that if the flags argument decides the scope of the signal instead of different types of fds it might make sense to either settle for "procfd_" or "pidfd_" as prefix. The community was willing to accept either (cf. [17] and [18]). Given that one developer expressed strong preference for the "pidfd_" prefix (cf. [13]) and with other developers less opinionated about the name we should settle for "pidfd_" to avoid further bikeshedding.
The "_send_signal" suffix was chosen to reflect the fact that the syscall takes on the job of multiple syscalls. It is therefore intentional that the name is not reminiscent of neither kill(2) nor rt_sigqueueinfo(2). Not the fomer because it might imply that pidfd_send_signal() is a replacement for kill(2), and not the latter because it is a hassle to remember the correct spelling - especially for non-native speakers - and because it is not descriptive enough of what the syscall actually does. The name "pidfd_send_signal" makes it very clear that its job is to send signals.
/* zombies */ Zombies can be signaled just as any other process. No special error will be reported since a zombie state is an unreliable state (cf. [3]). However, this can be added as an extension through the @flags argument if the need ever arises.
/* cross-namespace signals */ The patch currently enforces that the signaler and signalee either are in the same pid namespace or that the signaler's pid namespace is an ancestor of the signalee's pid namespace. This is done for the sake of simplicity and because it is unclear to what values certain members of struct siginfo_t would need to be set to (cf. [5], [6]).
/* compat syscalls */ It became clear that we would like to avoid adding compat syscalls (cf. [7]). The compat syscall handling is now done in kernel/signal.c itself by adding __copy_siginfo_from_user_generic() which lets us avoid compat syscalls (cf. [8]). It should be noted that the addition of __copy_siginfo_from_user_any() is caused by a bug in the original implementation of rt_sigqueueinfo(2) (cf. 12). With upcoming rework for syscall handling things might improve significantly (cf. [11]) and __copy_siginfo_from_user_any() will not gain any additional callers.
/* testing */ This patch was tested on x64 and x86.
/* userspace usage */ An asciinema recording for the basic functionality can be found under [9]. With this patch a process can be killed via:
#define _GNU_SOURCE #include <errno.h> #include <fcntl.h> #include <signal.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <sys/stat.h> #include <sys/syscall.h> #include <sys/types.h> #include <unistd.h>
static inline int do_pidfd_send_signal(int pidfd, int sig, siginfo_t *info, unsigned int flags) { #ifdef __NR_pidfd_send_signal return syscall(__NR_pidfd_send_signal, pidfd, sig, info, flags); #else return -ENOSYS; #endif }
int main(int argc, char *argv[]) { int fd, ret, saved_errno, sig;
if (argc < 3)
exit(EXIT_FAILURE);
fd = open(argv[1], O_DIRECTORY | O_CLOEXEC);
if (fd < 0) {
printf("%s - Failed to open \"%s\"\n", strerror(errno), argv[1]);
exit(EXIT_FAILURE);
}
sig = atoi(argv[2]);
printf("Sending signal %d to process %s\n", sig, argv[1]);
ret = do_pidfd_send_signal(fd, sig, NULL, 0);
saved_errno = errno;
close(fd);
errno = saved_errno;
if (ret < 0) {
printf("%s - Failed to send signal %d to process %s\n",
strerror(errno), sig, argv[1]);
exit(EXIT_FAILURE);
}
exit(EXIT_SUCCESS);
}
/* Q&A
- Given that it seems the same questions get asked again by people who are
- late to the party it makes sense to add a Q&A section to the commit
- message so it's hopefully easier to avoid duplicate threads.
- For the sake of progress please consider these arguments settled unless
- there is a new point that desperately needs to be addressed. Please make
- sure to check the links to the threads in this commit message whether
- this has not already been covered. */ Q-01: (Florian Weimer [20], Andrew Morton [21]) What happens when the target process has exited? A-01: Sending the signal will fail with ESRCH (cf. [22]).
Q-02: (Andrew Morton [21]) Is the task_struct pinned by the fd? A-02: No. A reference to struct pid is kept. struct pid - as far as I understand - was created exactly for the reason to not require to pin struct task_struct (cf. [22]).
Q-03: (Andrew Morton [21]) Does the entire procfs directory remain visible? Just one entry within it? A-03: The same thing that happens right now when you hold a file descriptor to /proc/ open (cf. [22]).
Q-04: (Andrew Morton [21]) Does the pid remain reserved? A-04: No. This patchset guarantees a stable handle not that pids are not recycled (cf. [22]).
Q-05: (Andrew Morton [21]) Do attempts to signal that fd return errors? A-05: See {Q,A}-01.
Q-06: (Andrew Morton [22]) Is there a cleaner way of obtaining the fd? Another syscall perhaps. A-06: Userspace can already trivially retrieve file descriptors from procfs so this is something that we will need to support anyway. Hence, there's no immediate need to add another syscalls just to make pidfd_send_signal() not dependent on the presence of procfs. However, adding a syscalls to get such file descriptors is planned for a future patchset (cf. [22]).
Q-07: (Andrew Morton [21] and others) This fd-for-a-process sounds like a handy thing and people may well think up other uses for it in the future, probably unrelated to signals. Are the code and the interface designed to permit such future applications? A-07: Yes (cf. [22]).
Q-08: (Andrew Morton [21] and others) Now I think about it, why a new syscall? This thing is looking rather like an ioctl? A-08: This has been extensively discussed. It was agreed that a syscall is preferred for a variety or reasons. Here are just a few taken from prior threads. Syscalls are safer than ioctl()s especially when signaling to fds. Processes are a core kernel concept so a syscall seems more appropriate. The layout of the syscall with its four arguments would require the addition of a custom struct for the ioctl() thereby causing at least the same amount or even more complexity for userspace than a simple syscall. The new syscall will replace multiple other pid-based syscalls (see description above). The file-descriptors-for-processes concept introduced with this syscall will be extended with other syscalls in the future. See also [22], [23] and various other threads already linked in here.
Q-09: (Florian Weimer [24]) What happens if you use the new interface with an O_PATH descriptor? A-09: pidfds opened as O_PATH fds cannot be used to send signals to a process (cf. [2]). Signaling processes through pidfds is the equivalent of writing to a file. Thus, this is not an operation that operates "purely at the file descriptor level" as required by the open(2) manpage. See also [4].
/* References */ [1]: https://lore.kernel.org/lkml/20181029221037.87724-1-dancol@google.com/ [2]: https://lore.kernel.org/lkml/874lbtjvtd.fsf@oldenburg2.str.redhat.com/ [3]: https://lore.kernel.org/lkml/20181204132604.aspfupwjgjx6fhva@brauner.io/ [4]: https://lore.kernel.org/lkml/20181203180224.fkvw4kajtbvru2ku@brauner.io/ [5]: https://lore.kernel.org/lkml/20181121213946.GA10795@mail.hallyn.com/ [6]: https://lore.kernel.org/lkml/20181120103111.etlqp7zop34v6nv4@brauner.io/ [7]: https://lore.kernel.org/lkml/36323361-90BD-41AF-AB5B-EE0D7BA02C21@amacapital.net/ [8]: https://lore.kernel.org/lkml/87tvjxp8pc.fsf@xmission.com/ [9]: https://asciinema.org/a/IQjuCHew6bnq1cr78yuMv16cy [11]: https://lore.kernel.org/lkml/F53D6D38-3521-4C20-9034-5AF447DF62FF@amacapital.net/ [12]: https://lore.kernel.org/lkml/87zhtjn8ck.fsf@xmission.com/ [13]: https://lore.kernel.org/lkml/871s6u9z6u.fsf@xmission.com/ [14]: https://lore.kernel.org/lkml/20181206231742.xxi4ghn24z4h2qki@brauner.io/ [15]: https://lore.kernel.org/lkml/20181207003124.GA11160@mail.hallyn.com/ [16]: https://lore.kernel.org/lkml/20181207015423.4miorx43l3qhppfz@brauner.io/ [17]: https://lore.kernel.org/lkml/CAGXu5jL8PciZAXvOvCeCU3wKUEB_dU-O3q0tDw4uB_ojMvDEew@mail.gmail.com/ [18]: https://lore.kernel.org/lkml/20181206222746.GB9224@mail.hallyn.com/ [19]: https://lore.kernel.org/lkml/20181208054059.19813-1-christian@brauner.io/ [20]: https://lore.kernel.org/lkml/8736rebl9s.fsf@oldenburg.str.redhat.com/ [21]: https://lore.kernel.org/lkml/20181228152012.dbf0508c2508138efc5f2bbe@linux-foundation.org/ [22]: https://lore.kernel.org/lkml/20181228233725.722tdfgijxcssg76@brauner.io/ [23]: https://lwn.net/Articles/773459/ [24]: https://lore.kernel.org/lkml/8736rebl9s.fsf@oldenburg.str.redhat.com/ [25]: https://lore.kernel.org/lkml/CAK8P3a0ej9NcJM8wXNPbcGUyOUZYX+VLoDFdbenW3s3114oQZw@mail.gmail.com/
Cc: "Eric W. Biederman" ebiederm@xmission.com Cc: Jann Horn jannh@google.com Cc: Andy Lutomirsky luto@kernel.org Cc: Andrew Morton akpm@linux-foundation.org Cc: Oleg Nesterov oleg@redhat.com Cc: Al Viro viro@zeniv.linux.org.uk Cc: Florian Weimer fweimer@redhat.com Signed-off-by: Christian Brauner christian@brauner.io Reviewed-by: Tycho Andersen tycho@tycho.ws Reviewed-by: Kees Cook keescook@chromium.org Reviewed-by: David Howells dhowells@redhat.com Acked-by: Arnd Bergmann arnd@arndb.de Acked-by: Thomas Gleixner tglx@linutronix.de Acked-by: Serge Hallyn serge@hallyn.com Acked-by: Aleksa Sarai cyphar@cyphar.com
msm_thermal: simplified thermal driver
Thermal driver by franco. This is a combination of 9 commits:
msm: thermal: add my simplified thermal driver. Stock thermal-engine-hh goes crazy too soon and too often and offers no way for userland to tweak its parameters
Signed-off-by: franciscofranco franciscofranco.1990@gmail.com
msm: thermal: moar magic
Added a sample time between heat levels. The hotter it is, the longer it should stay throttled in that same freq level therefore cooling this down effectively. Also due to this change freqs have been slightly adjusted. Now the driver will start a bit earlier on boot. Few cosmetic changes too because why the fuck not.
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
msm: thermal: reduce throttle point to 60C
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
msm: thermal: rework previous patches
The changes on previous patches didn't really work out. Either the device just reboots for infinity during boot because if it reaches high temperatures and fails to throttle down fast enough to mitigate it, or it usually just crashes the device while benchmarking on Geekbench or Antutu again because when there's a big ass temp spike this doesn't mitigate fast enough.
These changes are confirmed working after testing in all scenarios previously described.
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
msm: thermal: work faster with more thrust
Last commit was not enough, it mitigated most of the issues, but some users were still having weird shits because temperature wasn't going down as fast as it should. So now queue it every fucking 100ms in a dedicated high prio workqueue. It's my last stance!
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
msm: thermal: offline cpu2 and cpu3 if things get REALLY rough
Just for safe measure put cpu2 and cpu3 to sleep if the heat gets way bad. Also the polling time gets back to default stock 250ms since the earlier 100ms change was just a band-aid for a nastier bug that got fixed in the mean time.
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
msm_thermal: send OFF/ONLINE uevent in hotplug cases
Send the correct uevent after setting a CPU core online or offline. This allows ueventd to set correct SELinux labels for newly created sysfs CPU device nodes.
Bug: 28887345 Change-Id: If31b8529b31de9544914e27514aca571039abb60 Signed-off-by: Siqi Lin siqilin@google.com Signed-off-by: Thierry Strudel tstrudel@google.com [Francisco: slightly adapted from Qcom's original patch to apply] Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
Revert "msm_thermal: send OFF/ONLINE uevent in hotplug cases"
Crashes everything if during early early boot the device is hot and starts to throttle. It's madness!
This reverts commit 80e38963f8080c3c9d26374693dd0f0a88f8060b.
msm: thermal: return to original simplified driver
Some users still had a weird issue that I was unable to reproduce which either consisted in cpu2 and cpu3 getting stuck in offline mode or after a gaming session while charging the device would crash with "hw_reset" and then the device would loop in bootloader -> boot animation forever until the device cooled down.
My test was leaving the device charging during the night, brightness close to max and running Stability Test app with the CPU+GPU suite. I woke up and the device was still running it flawlessly. Rebooted while hot and it booted just fine.
Since I was unable to reproduce the issue and @osm0sis flashed back to <r92 and was unable to reproduce it anymore here we go back to that stage.
Only change I made compared to that original driver was simply queue things into a dedicated high prio wq for faster thermal mitigation. Rest is unchanged.
Signed-off-by: Francisco Franco franciscofranco.1990@gmail.com
Created Text For URL [www.express.co.uk/news/science/1641028/energy-crisis-hell-shell-ceo-europe-really-tough-winter-rationing-putin-ukraine-war]
refactor(genshin): rewrite dot creation again?! i hate my life
Project Kasumi 1.3 "PoPiPa"
After a long time, another major update. This one doesn't have anything available to user base just yet, but we do have these changes as overview;
1- Picked PixelPropsUtils from Fork LineageOS. Even though doesn't have a good use on my personal tests at Rosemary, it at least has newer and more capable property sets now. 🤷♀️
2- Synced missing repos from LineageOS. Self-explanatory.
3- Some little touches and fixes here and there, for example charged percentage not appearing when unplugged.
This and that, we're still keeping up with our stuff, and have been dealing with some BS going on around my environment still...
Ah speaking about environment, I have completely discontinued both my main and Shorts YouTube channels. Don't blame me, it's my family's fault for causing headaches just because I'm minding my own business. Especially discovering my credit card was just some addon card that's tied to my mom's credit card already broke my trust against anyone.
Sorry for the little venting here, but I'm just not in the mood of continuing ROM development, you know..? I'm doing this just for you all, giving away from my life.
Signed-off-by: Beru Hinode windowz414@1337.lgbt
"11:20am. Let me have breakfast here.
12pm. Let me resume.
12:05pm. I've quietly removed the two posts on the TensTorrent and Groq subs. I should look into where their chips could be accessed over the clouds.
https://tenstorrent.com/devcloud/
We developed Tenstorrent DevCloud to give people the opportunity to try their models on our servers without purchasing our hardware.
I'll look into this later. Let me get back to AWS. The cloud is inevitable if you want to do more than play games on your desktop it seems.
12:20pm. Almost done with module 4. This is pretty boring and I have 7 modules left. I am not going to bother focusing too closely. Let me just deal with this and I will learn things hands-on.
12:30pm. Let me get started with module 5.
You can upload any type of file to Amazon S3, such as images, videos, text files, and so on. For example, you might use Amazon S3 to store backup files, media files for a website, or archived documents. Amazon S3 offers unlimited storage space. The maximum file size for an object in Amazon S3 is 5 TB.
Unlimited storage space? That could be abusable. I doubt it is really that, it probably just has a high limit.
This reminds me, how do SSD vs HDDs compare now?
While SSDs are much cheaper than they used to be, there is still a significant price difference. A 1 TB internal HDD costs roughly $60, whereas a 1 TB internal SSD averages around $150.
This is pretty cheap, I think. I forgot what the gap was 7.5 years ago, but it was more than 2.5.
https://www.tomshardware.com/features/best-deals-on-ssds
This is literally from 1 day ago.
Capacity: 1TB seems to be the sweet spot for price and performance, with decent NVMe drives going for around $100 or less and high-performance ones in the sub $150 range. You can save money with a 512GB drive or, for an older PC with limited needs, a 256GB unit can be extremely cheap.
1Tb SSDs go for 100$? That can't be right, he has to mean HDDs.
...Wow, he is right. I have a 500mb SSD and it was 5x more expensive than that when I got it. Moore's Law is alive and well for SSDs. It used to be the case that CPUs were getting better like that in the past.
1pm. Forget this. Let me get back to the tutorials. The storage talk just caught my attention.
1:15pm. Databases are not something I am familiar with, but AWS is a good way to get some of that benefit if I ever need it.
1:20pm. I had no idea what relational databases have performance issues under stress.
1:30pm. I am not even halfway done yet. I will pull through this.
1:45pm. Focus me. As boring as this is, let me finish module 5.
1:55pm. Mhhh, time for module 6. This one is on security.
2:35pm. Let me take a break here, the boredom of this is starting to get to me.
3:05pm. Let me resume. Security, analytics. Pricing at least will be something I am interested in. Let me just get it over with.
3:25pm. > Security groups operate at the AWS network level, not the EC2 instance level.
Had he not mentioned this I would have assumed the later simly due to the context EC2 was talked about.
3:45pm. Finally time for module 7. Why does the sidebar say 57% complete?
At any rate, let me just breeze through this. I was paying attention to the earlier modules, but through this I will just have to grind through. This is important for the sake of getting a basic familiarity with AWS.
4:05pm. Finally at module 8. This one is on pricing. This is something I need to know. After I am done with this today, tomorrow I will open a free tier account and start playing with it. The goal will be to built up to large distributed NN training. I won't actually train a real net on something, but I want to be able to do a single pass forward and backward on a net that couldn't be run on a single machine.
These offers are free for 12 months following your initial sign-up date to AWS. Examples include specific amounts of Amazon S3 Standard Storage, thresholds for monthly hours of Amazon EC2 compute time, and amounts of Amazon CloudFront data transfer out.
I guess I'll have some fun with that tomorrow.
4:20pm. Focus me, stop looking up Milky Holmes threads in the archive. Finish module 9.
4:35pm. Almost done with module 8. I am dying of boredom here. I should be done soon.
4:50pm. I'll skip the last 3 modules.
I get the gist of it by now.
5pm. Time for lunch.
5:20pm. Done with lunch. I am thinking what I should do for the rest of the day. There is nothing to it. I need to rest. What I did today was not too bad - some book learning. Tomorrow I need to sit down and refine it through practice. The kind of knowledge I learned today was merely a good overview. It gave me an idea what AWS is about.
It is actually pretty impressive. Imagine I was a regular web guy - managing databases, hardware, security and all of that would be hard. AWS really does offer to get all that tedious stuff out of the way allowing me to focus on what I am good at. With AWS, I could make a whole web app all by myself and scale it to any number of users. I wouldn't have thought it possible without these cloud services.
5:25pm. Tomorrow, what I will aim to do is very basic. Start an AWS instance and get PyTorch working on it. It is not even multi-GPU or distributed training - just get the thing to run. It won't be on any particular dataset. Maybe MNIST.
After that I'll look into multi GPU instances. Then I will look into distributed instances. After I have that, I'll have mastered the basics. It will not be hard. This exercise will introduce me to all the essential aspects of cloud computing on AWS as I need them for ML. In fact, it will also cover some other holes like databases for me. I remember installing SQL stuff last year on my PC, and it was a real pain to remove. I never ended up using or studying it properly. If I need those things I'll have them straight on the cloud where they won't get in the way on my own rig. I won't have to waste time setting it up.
After that challenge, it will be Azure's turn. Then I'll take look at the ML dedicated cloud service. Was it called something like Lambda Cube? I saw it in that Two Minute Paper video as an ad towards the end. I'll track it down when I need it.
5:45pm. I'll believe in my own skill and go forward, just like I always have. I thought I might have gotten rusty, but the ref counted C backend was a breeze. To the me of a couple of years ago, it would have been a very hard challenge. I know I am good.
My concurrency skills are only 4/5 rather than the apex, so I should be looking forward to exercising them more.
I admit, when I first started studying them seriously in 2020 I thought it might just be a boring web framework crap, but Rx and Hopac have both greatly expanded my horizons. Now I am glad about that experience. ZeroMQ is not bad either. Putting some more effort into the cloud aspect of it will give rise to a qualitative change in my capability. The main thing that is missing in me is proper attitude. I want to posses everything personally. Instead the right perspective is the cloud-centric perspective. Who cares who own what as long as it gives me improved capabilities.
It turns out, one of the unrealistic parts of Lain's Dream story that I wrote in 2014 was that I'd be able to get a neurochip out of the box. Whoever causes the Singularity will definitely be doing it out of the cloud."
[NTOS:SE] Properly handle dynamic counters in token
On current master, ReactOS faces these problems:
-
ObCreateObject charges both paged and non paged pool a size of TOKEN structure, not the actual dynamic contents of WHAT IS inside a token. For paged pool charge the size is that of the dynamic area (primary group + default DACL if any). This is basically what DynamicCharged is for. For the non paged pool charge, the actual charge is that of TOKEN structure upon creation. On duplication and filtering however, the paged pool charge size is that of the inherited dynamic charged space from an existing token whereas the non paged pool size is that of the calculated token body length for the new duplicated/filtered token. On current master, we're literally cheating the kernel by charging the wrong amount of quota not taking into account the dynamic contents which they come from UM.
-
Both DynamicCharged and DynamicAvailable are not fully handled (DynamicAvailable is pretty much poorly handled with some cases still to be taking into account). DynamicCharged is barely handled, like at all.
-
As a result of these two points above, NtSetInformationToken doesn't check when the caller wants to set up a new default token DACL or primary group if the newly DACL or the said group exceeds the dynamic charged boundary. So what happens is that I'm going to act like a smug bastard fat politician and whack the primary group and DACL of an token however I want to, because why in the hell not? In reality no, the kernel has to punish whoever attempts to do that, although we currently don't.
-
The dynamic area (aka DynamicPart) only picks up the default DACL but not the primary group as well. Generally the dynamic part is composed of primary group and default DACL, if provided.
In addition to that, we aren't returning the dynamic charged and available area in token statistics. SepComputeAvailableDynamicSpace helper is here to accommodate that. Apparently Windows is calculating the dynamic available area rather than just querying the DynamicAvailable field directly from the token. My theory regarding this is like the following: on Windows both TokenDefaultDacl and TokenPrimaryGroup classes are barely used by the system components during startup (LSASS provides both a DACL and primary group when calling NtCreateToken anyway). In fact DynamicAvailable is 0 during token creation, duplication and filtering when inspecting a token with WinDBG. So if an application wants to query token statistics that application will face a dynamic available space of 0.
[HWMemSimImpl] Add support for two flavors of read latencies
The current version of HWMemSimImpl implements read latency by
inserting pipeline registers into the address and enable lines, which
delays the time until the underlying data array is probed for a value.
I think we have inherited this behaviour from SFC.
In real-world SRAMs there is a second component to the read latency: the time it takes for the value sampled in the data array to propagate to the output ports. The choices for this pre- and post-array latency has subtle implications on the read-under-write behaviour of the memory, mainly concerning the order of writes that a given read sees. From experience I would argue that most SRAMs have a post-array latency, and very little pre-array latency.
This commit splits the readLatency into two components: a pre- and
post-latency. By default, the pass continues to treat the read latency
annotated on FIR memories as a pure pre-array latency, and keeps the
post-array latency at zero. However, it also introduces the
read-latency-is-propagation-delay option which switches this behaviour
into modeling that read latency entirely with a post-array delay.
The intent here is to have a non-functional change, but with an opt-in switch such that we can start testing the alternate (and in my opinion correct behaviour) on real-world designs. I suspect that none of SiFive's designs relies on this behaviour to ensure portability between different technology nodes, and we can just switch to this new behaviour for completeness' sake.
Brainfuck commands (#45)
- vroummmmmmmm fix and remove unused import of turtle.up which was causing all sorts of problems with starting the bot, you know, not everyone configures their python for tkinter just for the fun of it, some people just want to run a dumb bot without having to get into computer graphics, after all, we live in a free country, don't we. In any case, this commit message is already quite a bit longer than the commit itself so I shall end it here, wish you a lovely day, and thank you for reading until the end.
- vrum
- vruum plugged security hole by tracking original sender through brainfuck command
Co-authored-by: gorilskij 43991400@users.noreply.github.com
Overheat warnings to both gatling guns (#742)
-
The notorious
-
Epic
-
FUCK YOU
-
I am going to beat you with a club
[PATCH] Fix ext2 readdir f_pos re-validation logic
This fixes not one, but two, silly (but admittedly hard to hit) bugs in the ext2 filesystem "readdir()" function. It also cleans up the code to avoid the unnecessary goto mess.
The bugs were related to re-valiating the f_pos value after somebody had either done an "lseek()" on the directory to an invalid offset, or when the offset had become invalid due to a file being unlinked in the directory. The code would not only set the f_version too eagerly, it would also not update f_pos appropriately for when the offset fixup took place.
When that happened, we'd occasionally subsequently fail the readdir() even when we shouldn't (no real harm done, but an ugly printk, and obviously you would end up not necessarily seeing all entries).
Thanks to Masoud Sharbiani masouds@google.com who noticed the problem and had a test-case for it, and also fixed up a thinko in the first version of this patch.
Signed-off-by: Al Viro viro@zeniv.linux.org.uk Acked-by: Masoud Sharbiani masouds@google.com Signed-off-by: Linus Torvalds torvalds@osdl.org
perf: Fix loss of notification with multi-event
When you do: $ perf record -e cycles,cycles,cycles noploop 10
You expect about 10,000 samples for each event, i.e., 10s at 1000samples/sec. However, this is not what's happening. You get much fewer samples, maybe 3700 samples/event:
$ perf report -D | tail -15 Aggregated stats: TOTAL events: 10998 MMAP events: 66 COMM events: 2 SAMPLE events: 10930 cycles stats: TOTAL events: 3644 SAMPLE events: 3644 cycles stats: TOTAL events: 3642 SAMPLE events: 3642 cycles stats: TOTAL events: 3644 SAMPLE events: 3644
On a Intel Nehalem or even AMD64, there are 4 counters capable of measuring cycles, so there is plenty of space to measure those events without multiplexing (even with the NMI watchdog active). And even with multiplexing, we'd expect roughly the same number of samples per event.
The root of the problem was that when the event that caused the buffer to become full was not the first event passed on the cmdline, the user notification would get lost. The notification was sent to the file descriptor of the overflowed event but the perf tool was not polling on it. The perf tool aggregates all samples into a single buffer, i.e., the buffer of the first event. Consequently, it assumes notifications for any event will come via that descriptor.
The seemingly straight forward solution of moving the waitq into the ringbuffer object doesn't work because of life-time issues. One could perf_event_set_output() on a fd that you're also blocking on and cause the old rb object to be freed while its waitq would still be referenced by the blocked thread -> FAIL.
Therefore link all events to the ringbuffer and broadcast the wakeup from the ringbuffer object to all possible events that could be waited upon. This is rather ugly, and we're open to better solutions but it works for now.
Reported-by: Stephane Eranian eranian@google.com Finished-by: Stephane Eranian eranian@google.com Reviewed-by: Stephane Eranian eranian@google.com Signed-off-by: Peter Zijlstra a.p.zijlstra@chello.nl Link: http://lkml.kernel.org/r/20111126014731.GA7030@quad Signed-off-by: Ingo Molnar mingo@elte.hu
Revisiting The Goliath: Or, that time I dripped out the SBC Starfury just because (#68126)
Drips the SHIT out of the SBC Starfury while not completely overhauling it. Touches everything NOT in engineering or southward (because I love how scuffed that part is and refuse to touch it on principle) - Also converts one map varedit into a real boy subtype, and moves tiny fans to their own file.
Mandatory disclosure on the gameplay changes: Fighters 1 and 3 are now NOT in the hangar, and are now attached to the formerly unused gunnery rooms. Cryo now works. Yeah. I know. You can actually open the anesthetic closet now. Everyone now shares three spawners. This doesn't reduce the amount of people who can play when this rolls, as I've adjusted var/uses in accordance: it just reduces clutter. A few of the horizontal double airlocks have been compacted into glass_large airlocks. The bar windows now actually have grilles like they were meant to. Four turbines have shown up. They aren't functional*, they just look like gunnery and conveniently fit in the spots. I'm sure this is space OSHA compliant. The map is ever so slightly smaller, vertically. This should distance us from an edge case where somehow all space levels are too cluttered for this to spawn properly, for the time being.
*Technically there's nothing stopping you from using them besides the amount of time it'd take for the operatives to kick your ass
This map was originally designed wayyy back before we even had the computer sprites we have now, (#27760 if you want to see SOUL) and it shows. While it will never have it's SM again, we can at least make the thing much nicer to look at.
Tool-Brute_force-Zip_File-Encryption-Broken
◀ مرحباً, أنا احمد | مهندس سوفت وير
ما أقدمه هو " أداة فك تشفير ملف مضغوط - E n c r y p t i o n "
↩بعد الشرح ، ستتمكن من فهم مستقبل الأداة سواء كنت طالب في مرحلة النمو أو مطور برمجيات خبير , أمنحك التجربة التي تحتاجها لبرمجياتك الخاصة ، ومواءمة
فريقك، وبناء على تجربتي الخاصة تقنية القوة الغاشمة أو ما تسمى بالانجليزية ' Brute-force ' فأن كسر تشفير ملف مضغوط باستخدام القوة الغاشمة في بايثون بستخدام ' Brute-force ' يعد من اقوى التقنيات في كسر كلمات المرور و تجاوزها ↩حيث أن طريقتها في كسر كلمات المرور و تجاوزها تكمن في شيئين وهما إما:
نعم صديقي: هنا تبني قصتك, يجب أن تتحلى بالشفافية بشأن هوايتك "وما تؤمن به" وأن تنقله [باستمرار] عبر كل نقطة اتصال مثل مواقع التوصل الاجتماعي.
✅ "معرفة هوايتك هو سر القيام بذلك بشكل فعال"
بعد عرض هذه الأداة ، إذا كان لديك خبرة في لغة Python ، يمكنك تطوير الأداة لتنشأ ادوات فك تشفير تحتوي على مختلف الصيغ حسب تطويرك وطلب جمهورك ثم إتاحة تنزيلها لجميع المستخدمين
What I offer is "Zip file decryption tool - E n c r y p t i o n "
↪️After the explanation, you will be able to understand the future of the tool
Whether you are a growing student or an experienced software developer, I give you the experience you need for your software, and match
For your team, based on my own experience, the 'brute-force' technique Breaking the encryption of a compressed file using brute force in Python using 'Brute-force' is one of the most powerful techniques for cracking and bypassing passwords. ↪️ As its way of cracking and bypassing passwords lies in two things:
✅ "Knowing your hobby is the secret to doing it effectively"
After viewing this tool, if you have experience in the Python language, you can develop the tool to create decoders containing various formats as per your development and audience demand and then make it available for download to all users
Fix rowcount estimate for SubqueryScan that's under a Gather.
SubqueryScan was always getting labeled with a rowcount estimate appropriate for non-parallel cases. However, nodes that are underneath a Gather should be treated as processing only one worker's share of the rows, whether the particular node is explicitly parallel-aware or not. Most non-scan-level node types get this right automatically because they base their rowcount estimate on that of their input sub-Path(s). But SubqueryScan didn't do that, instead using the whole-relation rowcount estimate as if it were a non-parallel-aware scan node. If there is a parallel-aware node below the SubqueryScan, this is wrong, and it results in inflating the cost estimates for nodes above the SubqueryScan, which can cause us to not choose a parallel plan, or choose a silly one --- as indeed is visible in the one regression test whose results change with this patch. (Although that plan tree appears to contain no SubqueryScans, there were some in it before setrefs.c deleted them.)
To fix, use path->subpath->rows not baserel->tuples as the number of input tuples we'll process. This requires estimating the quals' selectivity afresh, which is slightly annoying; but it shouldn't really add much cost thanks to the caching done in RestrictInfo.
This is pretty clearly a bug fix, but I'll refrain from back-patching as people might not appreciate plan choices changing in stable branches. The fact that it took us this long to identify the bug suggests that it's not a major problem.
Per report from bucoo, though this is not his proposed patch.
Discussion: https://postgr.es/m/202204121457159307248@sohu.com
Fix: foo, poisoned by a cursed amulet of life saving.
Wearing a cursed amulet of life saving will not work, and the player will die via whatever caused them to die in the first place. The inclusion of KILLED_BY here was not entirely correct. Yes, the cursed amulet of life saving didn't help you, but that's not why you died.
This commit incorporates NetHack 3.7 commit 47cef18, which provides a proper 'while' statement if an amulet of life saving doesn't work. Then we call savelife() in the routine that causes a cursed amulet to fail. End product is proper death reason output.
Update README benchmark to latest numbers (#724)
The README benchmark numbers were last updated in January 2021, so this PR updates it to see what's changed since then.
To help understand the underlying reasons for changes in numbers, I updated each piece incrementally and re-ran the benchmark at each step. For the numbers shown here, I ran the full benchmark 3 times and took the best time for each tool.
Here's the original benchmark:
Time Speed
Sucrase 1.64 seconds 220221 lines per second
swc 2.13 seconds 169502 lines per second
esbuild 3.02 seconds 119738 lines per second
TypeScript 24.18 seconds 14937 lines per second
Babel 27.22 seconds 13270 lines per second
Since running the previous benchmark, I switched from a 2016 Intel MacBook Pro to a 2020 M1 MacBook Air, so I simulated the original benchmark on my new laptop by using Sucrase 3.17.0, the pinned versions of all other tools, and Node 14. Results:
Sucrase 0.76 seconds 477950 lines per second
swc 1.05 seconds 344215 lines per second
esbuild 1.43 seconds 253143 lines per second
TypeScript 9.28 seconds 38943 lines per second
Babel 11.03 seconds 32758 lines per second
Over a 2x improvement for all tools just from the new laptop! Pretty incredible in my opinion.
Then I upgraded to Node 18, still running the old version of all tools:
Sucrase 0.64 seconds 563963 lines per second
swc 1.04 seconds 348526 lines per second
esbuild 1.42 seconds 254111 lines per second
TypeScript 7.99 seconds 45210 lines per second
Babel 10.3 seconds 35076 lines per second
About a 10-20% improvement for tools written in JS just from upgrading Node. As expected, swc and esbuild aren't affected by the node upgrade.
Then I upgraded all tools to latest (and addressed a breaking change in esbuild usage):
Sucrase 0.65 seconds 558226 lines per second
swc 1.24 seconds 290311 lines per second
esbuild 1.47 seconds 245861 lines per second
TypeScript 8.97 seconds 40258 lines per second
Babel 9.02 seconds 40032 lines per second
Interestingly, most tools got a little slower, which isn't too surprising given that bug fixes and scope creep can gradually hurt performance.
Then I updated the usage of Sucrase and swc to account for better configuration
since last time. Most notably, Sucrase now has a disableESTransforms option
that should be used in most cases these days (and I'm hoping to make it the
default in a future semver-major release), and that better matches how all of
the other tools are configured here. Results:
Sucrase 0.57 seconds 636975 lines per second
swc 1.19 seconds 304526 lines per second
esbuild 1.45 seconds 248692 lines per second
TypeScript 8.98 seconds 40240 lines per second
Babel 9.18 seconds 39366 lines per second
It looks like disableESTransforms resulted in about a 15% speedup for Sucrase.
As always, there are plenty of caveats with these measurements, which are mentioned in the docs in the benchmark file, but I think they still are fair.
One part that I didn't update yet is the Jest codebase being transpiled. I don't anticipate that having a meaningful effect on the benchmark numbers.