Do you know ROHM?

It was 3 years ago when I decided I had to see something new. That was when I switched from Nokia to ROHM Semiconductors – probably the biggest change in my work career so far. During these 3 years I’ve made a few observations. This should not be a surprize as I have a set of normally functioning eyes and ears. Nose and tongue too but I avoid using them at work. I don’t think I should poke my nose to all things – and tongue... Well, in order to stay appropriate, I’ll just say that I prefer a red tone over brown.

But back to the observation; I selected ROHM as a place to work by a coincidence. A former colleague and a friend of mine Mikko Mutanen had just started working at ROHM – and it was from him that I heard about the company and their open position. I had never heard of ROHM before. When I was leaving Nokia and I told my former colleagues about this job-change, I understood that none of them knew ROHM. A typical conversation with my former colleagues was something like this:

“Hi Matti, I heard you’re leaving Nokia!”

“Yes I am. I really love the work here but I think it’s time to learn something new.”

“Oh, what’s the plan, where are you going to work?”

“I’ll start at ROHM Semiconductors.”

“What? Where?”

“ROHM.”

“Ro… What was that again?”

“ ROHM, R - O - H - M. It’s a Japanese semiconductor company...”

I didn’t think much of it back then… But since then, this has started puzzling me. ROHM is not a new company, nor is it a terribly small one. Founded in 1958 and currently employing over 21000 persons ROHM has produced and sold a huge amount of different components varying from really simple resistors to amazingly complex TCONs and PMICs. Chances are you are happily using ROHM ICs in your handheld devices, gaming consoles or your car. The products produced by LAPIS Semiconductors and Kionix – which are part of the ROHM group – make the range of the IC catalogue even wider. It would probably be easier to list the type of ICs not produced by ROHM than the ICs it does produce. I think ROHM produces so many components one could sink a battleship with them.

So how come it’s so unknown? It’s for sure not that people at Nokia are more ignorant than others. On the contrary, my former colleagues were all working on in the tech field and most of them had various hobby projects related to electronics; an electric scooter, MIDI mixer/keyboard, device measuring water flowing in a pipe, lawnmower AI … just to name a few.

I think it’s because ROHM does not seem to make much noise about what it has done. Maybe that’s a characteristic of Japanese culture?

I decided to allow you to take a little peek on something we do here at ROHM Finland. As you may know, the Linux v5.12 was just released few days ago. And as always, after the release, maintainers start sending pull-requests to Linus for changes that should eventually make the next release.

It seems the v5.13-rc1 will again be a release introducing changes made by ROHM. As we can see from the pull-request from Lee Jones, the MFD maintainer, Linux v5.13-rc1 will probably bring support for ROHM BD71815, BD9576 and BD9573 PMICs and improvements to the BD71837 and BD71847. The BD71815 is widely used for powering older NXP i.MX-7 platforms, BD71837 and BD71847 are used with NXP's i.MX-8 SoCs and the BD9573 and BD9576 are intended to be used to power the Renesas R-Car SoCs. Also, the GPIO, RTC, CLK and Watchdog subsystems gained small improvements to ROHM IC drivers.

I think it is unsurprising that companies pay someone for writing software supporting their hardware although only a few go the extra mile and include support in mainline kernel. So, I think it is worth pointing out that ROHM does also encourage helping out the community with changes not directly related to ROHM ICs. As a few examples the driver-core maintained by Greg KH, received support for new devm-based delayed-work helper from a ROHM employer. This did also provide a dedicated place to add more such helpers. At the same time this change set cleaned a few non-ROHM IC drivers and killed a couple of bugs from them.

The regulator framework gained a generic helper for setting the regulator ramp-delay.

Speaking of which; the regulator subsystem is still being worked on by us to introduce a bigger regulator change. That is going to add a new IRQ helper to ease sending out the regulator notifications, extend the notifications to support warning level reports and finally add device-tree properties and driver call-backs for setting the protection/error/warning levels. Additionally, this should add protection API for gracefully shutting down the system on severe hardware malfunction in order to prevent further damage. I am cautiously hopeful that would land on the Linux v5.14-rc1 :)

So, ROHM may not be the biggest player in the field of software. Nor is it the biggest in the scale of Linux kernel development – but it still is a player. And we should be proud of that.

I saw this cool mathematical 'thought provoker' somewhere:

(1.00)^365 = 1.00

(1.01)^365 = 37.7

making constant small progress vs. doing nothing at all:

Let’s keep making the difference together!