IHIFI770C gets stuck during boot process

[fuyi]

The IHIFI770C using daily builds fails to start and gets stuck at the Rockbox boot logo. I tried building from the Master branch source code, but the same issue occurs. However, after removing the INIT_ATTR from the line extern void system_init(void) INIT_ATTR; in the firmware/export/system.h file, it starts normally. My C language skills are poor, and I’m unsure if removing this might cause any side effects. Could someone help check how to properly fix this?

[speachy]

The IHIFI770C using daily builds fails to start and gets stuck at the Rockbox boot logo. I tried building from the Master branch source code, but the same issue occurs. However, after removing the INIT_ATTR from the line extern void system_init(void) INIT_ATTR; in the firmware/export/system.h file, it starts normally. My C language skills are poor, and I’m unsure if removing this might cause any side effects. Could someone help check how to properly fix this?

Wow, I've been trying to find someone with one of these devices (770/770C/800) for several years now; as far as I am aware you are the first person who has tried this stuff out since it was merged.

The change you made is an interesting one; can I ask why you tried that in particular?  Also, where did the bootloader come from?

As for the failure, system_init() is only called as part of the early boot process by other early boot code, so nothing immediately stands out.  I'm looking into it though

[fuyi]

The IHIFI770C using daily builds fails to start and gets stuck at the Rockbox boot logo. I tried building from the Master branch source code, but the same issue occurs. However, after removing the INIT_ATTR from the line extern void system_init(void) INIT_ATTR; in the firmware/export/system.h file, it starts normally. My C language skills are poor, and I’m unsure if removing this might cause any side effects. Could someone help check how to properly fix this?

Wow, I've been trying to find someone with one of these devices (770/770C/800) for several years now; as far as I am aware you are the first person who has tried this stuff out since it was merged.

The change you made is an interesting one; can I ask why you tried that in particular?  Also, where did the bootloader come from?

As for the failure, system_init() is only called as part of the early boot process by other early boot code, so nothing immediately stands out.  I'm looking into it though

Thank you for addressing the issue. I own the iHiFi 770C, HM-601 Slim, and HM-603 Slim—all RK27XX-based devices. Rockbox runs smoothly on the iHiFi 770C, even more stably than on my HM-60x series.

When I tried updating Rockbox and encountered the boot hang on the iHiFi 770C, I traced back through Git commit history to find the last working version. I discovered that starting from commit 39439f69, the device failed to boot. After reviewing the changes, I noticed modifications to system_init(), so I attempted to revert those changes in the latest code—and it successfully booted.

Interestingly, the HM-60x series (also RK27XX-based) doesn’t exhibit this issue, which suggests the problem might lie in the iHiFi 770C’s device-specific code.

The official Rockbox installation instructions don’t include the iHiFi 770C. I sourced the bootloader from Rockbox-Vortex. Comparing the code merged into Rockbox, I confirmed it’s identical. The installation process for iHiFi devices resembles the HM-60x method: renaming bootloader.rkw to BASE.rkw and overwriting it in the original firmware. The key difference is that the iHiFi requires flashing via RK27Upgrade_image.exe, a tool provided on the Rockbox-Vortex website.