Go to advanced search

by PhilE
Sat Oct 12, 2019 4:06 pm
Forum: Device Tree
Topic: DPI displays - dpi24 overlay fails on RPi4 (SOLVED)
Replies: 32
Views: 3578

Re: DPI displays - dpi24 overlay fails on RPi4 (SOLVED)

The launch Pi 4 firmware didn't take into account the fact that the drive strength registers use a different scale on BCM2711, result in drive strengths that were half what they should have been. This was fixed in firmware released on September 17th.
by PhilE
Fri Oct 11, 2019 4:21 pm
Forum: Compute Module
Topic: Is CM3+ compatible with everything CM3?
Replies: 2
Views: 113

Re: Is CM3+ compatible with everything CM3?

From a software perspective it could have been a silent upgrade.
by PhilE
Fri Oct 11, 2019 1:39 pm
Forum: Advanced users
Topic: Pi4 64-bit raspbian kernel for testing - Focus on Pi4
Replies: 183
Views: 22067

Re: Pi4 64-bit raspbian kernel for testing - Focus on Pi4

Problem is that my kernels have non-standard names, not kernel7.img and kernel8.img And I want to prevent that an end-user could accidently brick their system by running for example rpi-update, which would replace the kernels by stock versions if they had standard names). The upcoming "os_prefix" w...
by PhilE
Wed Oct 09, 2019 8:26 pm
Forum: Compute Module
Topic: DSI Panel JDI LT070ME05000 bringup issues
Replies: 5
Views: 188

Re: DSI Panel JDI LT070ME05000 bringup issues

I think your diagnosis is correct, but I believe the fix is simpler than you think. The panel driver code that calls gpiod_set_value also calls functions that sleep, so I can't think of a reason not to replace them all with gpiod_set_value_cansleep instead.
by PhilE
Tue Oct 08, 2019 3:46 pm
Forum: Advanced users
Topic: Pi4 64-bit raspbian kernel for testing - Focus on Pi4
Replies: 183
Views: 22067

Re: Pi4 64-bit raspbian kernel for testing - Focus on Pi4

There's an "os_prefix" setting in the works that will allow multiple kernels to be installed on a single card (space permitting).
by PhilE
Tue Oct 08, 2019 3:13 pm
Forum: Advanced users
Topic: Pi4 64-bit raspbian kernel for testing - Focus on Pi4
Replies: 183
Views: 22067

Re: Pi4 64-bit raspbian kernel for testing - Focus on Pi4

Current rpi-update firmware will always prefer 32-bit kernels over 64-bit kernels. The only auto-selection occurs when only one is present, but it can be overruled (and made to fail to load) by setting arm_64bit explicitly to 1 or 0. As a result, you don't need a filter for 64-bit.
by PhilE
Tue Oct 08, 2019 2:26 pm
Forum: Advanced users
Topic: New RT (Real Time) kernel branch
Replies: 63
Views: 15694

Re: New RT (Real Time) kernel branch

Remove your blinkers and look again. You're starting to feel like a time waster, and you're polluting what is meant to be an informative topic. If you still want help start a new thread and try doing some of the leg work yourself.
by PhilE
Mon Oct 07, 2019 9:03 pm
Forum: Advanced users
Topic: New RT (Real Time) kernel branch
Replies: 63
Views: 15694

Re: New RT (Real Time) kernel branch

The instructions worked just fine - for a 3B+: sudo apt-get install git bc bison flex libssl-dev git clone --depth=1 --branch rpi-4.19.y-rt https://github.com/raspberrypi/linux cd linux KERNEL=kernel7 make bcm2709_defconfig make -j4 zImage modules dtbs sudo make modules_install sudo cp arch/arm/boot...
by PhilE
Mon Oct 07, 2019 5:45 pm
Forum: Advanced users
Topic: New RT (Real Time) kernel branch
Replies: 63
Views: 15694

Re: New RT (Real Time) kernel branch

What does "git branch -vv" report?
by PhilE
Mon Oct 07, 2019 3:24 pm
Forum: Advanced users
Topic: New RT (Real Time) kernel branch
Replies: 63
Views: 15694

Re: New RT (Real Time) kernel branch

1. You should be using bcm2709_defconfig for 3B+ (which you indicated is what you are building for) - bcm2711_defconfig is for 4B. 2. Both bcm2709_defconfig and bcm2711_defconfig in the rpi-4.19.y-rt branch (you did checkout the correct branch, didn't you?) define CONFIG_PREEMPT_RT_FULL, which resul...
by PhilE
Mon Oct 07, 2019 12:49 pm
Forum: Device Tree
Topic: enhance shutdown overlay
Replies: 3
Views: 176

Re: enhance shutdown overlay

That post explains how the gpio-shutdown overlay works (it generates a POWER key press in response to a GPIO level change), but it doesn't help to add the delay. As Dougie said, overlays don't include code - they just allow Device Tree properties to be changed in a way that a suitably configurable d...
by PhilE
Fri Oct 04, 2019 1:08 pm
Forum: Device Tree
Topic: inv-mpu6050 trigger probe fail -22
Replies: 5
Views: 246

Re: inv-mpu6050 trigger probe fail -22

It seems as though the only way this would work is if the interrupt for mpu9250 was not nested. Or if it were threaded? I found this page explaining the difference between chained and threaded interrupts: https://stackoverflow.com/questions/34377846/what-is-chained-irq-in-linux-when-are-they-need-t...
by PhilE
Fri Oct 04, 2019 8:26 am
Forum: Device Tree
Topic: 1-bit SDIO on 3B+ with internal WiFi
Replies: 1
Views: 135

Re: 1-bit SDIO on 3B+ with internal WiFi

The BCM2835/6/7 SoCs have two SD interfaces: 1. SD0 (also called SDHost), a homebrew Broadcom interface that we use to drive the SD card slot, and 2. SD1 (also called MMC), a more standard Arasan SDHCI interface (but without any ADMA support) that is normally used to drive the WiFi device over SDIO....
by PhilE
Thu Oct 03, 2019 12:35 pm
Forum: Device Tree
Topic: inv-mpu6050 trigger probe fail -22
Replies: 5
Views: 246

Re: inv-mpu6050 trigger probe fail -22

I'm wondering if the sx1509 driver is rejecting the attempt to use an interrupt, or whether it's happening in one of the Linux frameworks. Since you're already building your own kernel, try adding some tracing to drivers/pinctrl/pinctrl-sx150x.c, concentrating initially on the places where it return...
by PhilE
Wed Oct 02, 2019 8:50 pm
Forum: General discussion
Topic: Raspbian Jessie linux 4.4.9 Severe Performance Degradation
Replies: 103
Views: 44370

Re: Raspbian Jessie linux 4.4.9 Severe Performance Degradation

"0x" is a prefix used by the C programming language (and others) to indicate that the number is in hexadecimal (you can Google that). When we speak of bit 0 we mean the binary digit with the place value of 1 (2**0). Bit 1 is has the place value of 2 (2**1). Bit 16 is 65536 (2**16). Each hexadecimal ...
by PhilE
Wed Oct 02, 2019 9:13 am
Forum: Device Tree
Topic: LED gpio in device tree, __overrides__
Replies: 1
Views: 187

Re: LED gpio in device tree, __overrides__

Read section 2.2 of the Device Tree guide, then come back with any questions you have.
by PhilE
Tue Oct 01, 2019 4:39 pm
Forum: Device Tree
Topic: Adding touchscreen driver in device tree for i2c?
Replies: 4
Views: 250

Re: Adding touchscreen driver in device tree for i2c?

The official Linux kernel repository doesn't really recognise overlays. All of the Raspberry Pi overlays are hosted in our "linux" GitHub repo: https://github.com/raspberrypi/linux The current release branch is rpi-4.19.y, and once accepted we will take care of replicating it to other active branches.
by PhilE
Tue Oct 01, 2019 8:43 am
Forum: Compute Module
Topic: RPi 3 model B+ electrical damage
Replies: 1
Views: 141

Re: RPi 3 model B+ electrical damage

1. This is a Compute Module forum.
2. If there is a published reduced schematic for a board we are very unlikely to subsequently release a less-reduced schematic for the same board.
by PhilE
Mon Sep 30, 2019 9:42 am
Forum: Device Tree
Topic: Adding touchscreen driver in device tree for i2c?
Replies: 4
Views: 250

Re: Adding touchscreen driver in device tree for i2c?

The good news is that you've done the hard bit already. To make this your own overlay you'll need to rename it and add the name of the `.dtbo` to the Makefile in `arch/arm/boot/dts/`. You should also add a description to `arch/arm/boot/dts/README`. At this point, using the `dtbs` kernel make target ...
by PhilE
Sun Sep 29, 2019 7:29 am
Forum: Troubleshooting
Topic: VLI firmware v2.0 - powersaving features enabled
Replies: 65
Views: 7667

Re: VLI firmware v2.0 - powersaving features enabled

The hang is something that only happens occasionally immediately after the EEPROM is reprogrammed.
by PhilE
Sat Sep 28, 2019 12:03 pm
Forum: Compute Module
Topic: CM3+ UART1 on Pin14/15 does not work together with UART0
Replies: 4
Views: 268

Re: CM3+ UART1 on Pin14/15 does not work together with UART0

Nice job with the overlay. If you give me a line of the form:

Code: Select all

 Signed-off-by: Phil Elwell <phil@raspberrypi.org>
I'll use it when I submit the patch to the kernel tree. It's no problem if you'd rather not.
by PhilE
Fri Sep 27, 2019 4:43 pm
Forum: Compute Module
Topic: CM3+ UART1 on Pin14/15 does not work together with UART0
Replies: 4
Views: 268

Re: CM3+ UART1 on Pin14/15 does not work together with UART0

The behaviour you've observed is caused by a hack in the uart0 overlay. The issue is caused by the fact that the normal Pi dt-blob enables a UART on GPIOs 14 & 15. The firmware knows how to switch this to UART1 when needed, but it doesn't understand that when an overlay enables the same UART on othe...
by PhilE
Fri Sep 27, 2019 9:46 am
Forum: Device Tree
Topic: Trouble adding a gpio device to the device tree
Replies: 3
Views: 238

Re: Trouble adding a gpio device to the device tree

I'm happy you got it working - good catch with the "target = ". There are a few more things I meant to say: 1. I'm trying to standardise on "brcm,bcm2835" as the compatible string for overlays, but at the moment neither the firmware nor the kernel are checking. 2. The firmware should have logged a f...
by PhilE
Thu Sep 26, 2019 3:28 pm
Forum: Device Tree
Topic: Trouble adding a gpio device to the device tree
Replies: 3
Views: 238

Re: Trouble adding a gpio device to the device tree

Your overlay has conflated GPIO controllers with pin control groups. They're related concepts, and because they usually appear together in overlays it's easy to see where the confusion could creep in. Fragment 0 declares a pin control group in order that the pinctrl subsystem can reserve some pins o...

Go to advanced search