VisionFive 2 eMMC boot issues.

VisionFive 2 eMMC boot issues.

[Milad]

Hello everyone,

Little reminder on this thread regarding my VisionFive 2 MMC boot issue.

Archive link:
https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html

Just following up to see if anyone has had a chance to look at the details
in my original message on the link.

Thanks!

Re: VisionFive 2 eMMC boot issues.

[Yao Zi]

On Tue, Jul 28, 2026 at 04:42:03PM +0200, Milad wrote:
> Hello everyone,
> 
> Little reminder on this thread regarding my VisionFive 2 MMC boot issue.
> 
> Archive link:
> https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html

The last time I tried upstream U-Boot with SD cards is about v2025.10 -
v2026.01 and it works fine on my side. Could you try these two versions,
too, so we know whether it's a new regression or just a bug yet to be
discovered?

Thanks,
Yao Zi

> Just following up to see if anyone has had a chance to look at the details
> in my original message on the link.
> 
> Thanks!

Re: VisionFive 2 eMMC boot issues.

[Milad]

Quick update, i also tested v2026.04 as well, and it also boots the SD card
correctly with the same clean pattern as v2025.10/v2026.01, straight
through to GRUB and Debian, no cluster read errors. Log below.

So the regression is most likely somewhere between v2026.04 and v2026.07.

Also worth mentioning, someone else in a community also reported having the
exact same errors on v2026.07 (SD/EFI boot failure), and mentioned v2026.04
worked for them too which is why i tried it so this isn't isolated to my
board specifically.

--- v2026.04 log ---
U-Boot SPL 2026.04 (Jul 31 2026 - 18:19:25 +0200)
DDR version: dc2e84f0.
Trying to boot from SPI


U-Boot 2026.04 (Jul 31 2026 - 18:19:25 +0200)

CPU: sifive,u74-mc
Model: StarFive VisionFive 2 v1.3B
DRAM: 8 GiB
Core: 159 devices, 30 uclasses, devicetree: board
WDT: Not starting watchdog@13070000
MMC: mmc@16010000: 0, mmc@16020000: 1
Loading Environment from SPIFlash... SF: Detected gd25lq128 with page size
256 Bytes, erase size 4 KiB, total 16 MiB
OK
StarFive EEPROM format v2

--------EEPROM INFO--------
Vendor : StarFive Technology Co., Ltd.
Product full SN: VF7110B1-2318-D008E000-18005137
data version: 0x2
PCB revision: 0xb2
BOM revision: A
Ethernet MAC0 address: 6c:cf:39:00:ba:b0
Ethernet MAC1 address: 6c:cf:39:00:ba:b1
--------EEPROM INFO--------

In: serial@10000000
Out: serial@10000000
Err: serial@10000000
Net: eth0: ethernet@16030000, eth1: ethernet@16040000
starting USB...
USB XHCI 1.00
Bus xhci_pci: 2 USB Device(s) found
       scanning usb for storage devices... 0 Storage Device(s) found
Working FDT set to ff6fab80
Hit any key to stop autoboot: 0
Card did not respond to voltage select! : -110
** Booting bootflow '<NULL>' with efi_mgr
Booting: Label: mmc 1 Device path:
/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,0000000000000000)/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,7200000000000000)/SD(1)/SD(1)
...
Loading Linux 6.12.96+deb13-riscv64 ...
Loading initial ramdisk ...
/dev/mmcblk1p2: clean, 61400/7364608 files, 4698361/29427968 blocks
Debian GNU/Linux 13 visionfive-2 ttyS0
visionfive-2 login:

Happy to keep looking into it this if useful, please let me know if there's
a specific commit or intermediate version i might need to test.


On Tue, 28 Jul 2026, 18:28 Yao Zi, <me@ziyao.cc> wrote:

> On Tue, Jul 28, 2026 at 04:42:03PM +0200, Milad wrote:
> > Hello everyone,
> >
> > Little reminder on this thread regarding my VisionFive 2 MMC boot issue.
> >
> > Archive link:
> > https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html
>
> The last time I tried upstream U-Boot with SD cards is about v2025.10 -
> v2026.01 and it works fine on my side. Could you try these two versions,
> too, so we know whether it's a new regression or just a bug yet to be
> discovered?
>
> Thanks,
> Yao Zi
>
> > Just following up to see if anyone has had a chance to look at the
> details
> > in my original message on the link.
> >
> > Thanks!
>

Re: VisionFive 2 eMMC boot issues.

[Yao Zi]

On Fri, Jul 31, 2026 at 06:38:38PM +0200, Milad wrote:
> Quick update, i also tested v2026.04 as well, and it also boots the SD card
> correctly with the same clean pattern as v2025.10/v2026.01, straight
> through to GRUB and Debian, no cluster read errors. Log below.
> 
> So the regression is most likely somewhere between v2026.04 and v2026.07.

Thanks for the information. I'll take a look later this day.

Best regards,
Yao Zi

> Also worth mentioning, someone else in a community also reported having the
> exact same errors on v2026.07 (SD/EFI boot failure), and mentioned v2026.04
> worked for them too which is why i tried it so this isn't isolated to my
> board specifically.
> 
> --- v2026.04 log ---
> U-Boot SPL 2026.04 (Jul 31 2026 - 18:19:25 +0200)
> DDR version: dc2e84f0.
> Trying to boot from SPI
> 
> 
> U-Boot 2026.04 (Jul 31 2026 - 18:19:25 +0200)
> 
> CPU: sifive,u74-mc
> Model: StarFive VisionFive 2 v1.3B
> DRAM: 8 GiB
> Core: 159 devices, 30 uclasses, devicetree: board
> WDT: Not starting watchdog@13070000
> MMC: mmc@16010000: 0, mmc@16020000: 1
> Loading Environment from SPIFlash... SF: Detected gd25lq128 with page size
> 256 Bytes, erase size 4 KiB, total 16 MiB
> OK
> StarFive EEPROM format v2
> 
> --------EEPROM INFO--------
> Vendor : StarFive Technology Co., Ltd.
> Product full SN: VF7110B1-2318-D008E000-18005137
> data version: 0x2
> PCB revision: 0xb2
> BOM revision: A
> Ethernet MAC0 address: 6c:cf:39:00:ba:b0
> Ethernet MAC1 address: 6c:cf:39:00:ba:b1
> --------EEPROM INFO--------
> 
> In: serial@10000000
> Out: serial@10000000
> Err: serial@10000000
> Net: eth0: ethernet@16030000, eth1: ethernet@16040000
> starting USB...
> USB XHCI 1.00
> Bus xhci_pci: 2 USB Device(s) found
>        scanning usb for storage devices... 0 Storage Device(s) found
> Working FDT set to ff6fab80
> Hit any key to stop autoboot: 0
> Card did not respond to voltage select! : -110
> ** Booting bootflow '<NULL>' with efi_mgr
> Booting: Label: mmc 1 Device path:
> /VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,0000000000000000)/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,7200000000000000)/SD(1)/SD(1)
> ...
> Loading Linux 6.12.96+deb13-riscv64 ...
> Loading initial ramdisk ...
> /dev/mmcblk1p2: clean, 61400/7364608 files, 4698361/29427968 blocks
> Debian GNU/Linux 13 visionfive-2 ttyS0
> visionfive-2 login:
> 
> Happy to keep looking into it this if useful, please let me know if there's
> a specific commit or intermediate version i might need to test.
> 
> 
> On Tue, 28 Jul 2026, 18:28 Yao Zi, <me@ziyao.cc> wrote:
> 
> > On Tue, Jul 28, 2026 at 04:42:03PM +0200, Milad wrote:
> > > Hello everyone,
> > >
> > > Little reminder on this thread regarding my VisionFive 2 MMC boot issue.
> > >
> > > Archive link:
> > > https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html
> >
> > The last time I tried upstream U-Boot with SD cards is about v2025.10 -
> > v2026.01 and it works fine on my side. Could you try these two versions,
> > too, so we know whether it's a new regression or just a bug yet to be
> > discovered?
> >
> > Thanks,
> > Yao Zi
> >
> > > Just following up to see if anyone has had a chance to look at the
> > details
> > > in my original message on the link.
> > >
> > > Thanks!
> >

Re: VisionFive 2 eMMC boot issues.

[Conor Dooley]

On 28/07/2026 15:42, Milad wrote:
> Hello everyone,
> 
> Little reminder on this thread regarding my VisionFive 2 MMC boot issue.
> 
> Archive link:
> https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html
> 
> Just following up to see if anyone has had a chance to look at the details
> in my original message on the link.
> 
> Thanks!

I don't think you CCed anyone that cares about the platform on that
patch so that's probably the problem there.

Is this the same issue that Ilya attempted to fix in the upstream version
of the devicetree for this platform?
https://lore.kernel.org/all/20260306171939.84568-1-k0tran@altlinux.org/

My subsequent PR for that lead to issues and it never got applied.
https://lore.kernel.org/soc/4dd4ffe6-307a-442b-ae99-50c88d4e5b84@freeshell.de/

This happened back in March, and I just have not had the time nor
inclination to actually sort it out, and nobody else has stepped up.

I think I will just pick that one patch back up again and resubmit
it for this cycle. After all, it doesn't change the configuration, just
the point at which the configuration should be done.

If you could let me know if that patch helps, then it would be another
argument for actually (finally) getting it merged.

Thanks,
Conor.

Re: VisionFive 2 eMMC boot issues.

[Yao Zi]

On Tue, Jul 28, 2026 at 07:02:02PM +0100, Conor Dooley wrote:
> On 28/07/2026 15:42, Milad wrote:
> > Hello everyone,
> > 
> > Little reminder on this thread regarding my VisionFive 2 MMC boot issue.
> > 
> > Archive link:
> > https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html
> > 
> > Just following up to see if anyone has had a chance to look at the details
> > in my original message on the link.
> > 
> > Thanks!
> 
> I don't think you CCed anyone that cares about the platform on that
> patch so that's probably the problem there.
> 
> Is this the same issue that Ilya attempted to fix in the upstream version
> of the devicetree for this platform?
> https://lore.kernel.org/all/20260306171939.84568-1-k0tran@altlinux.org/

Quoting logs in the link,

> SD card (mmc1) fails to initialize in U-Boot. Full serial output:
>
> Card did not respond to voltage select! : -110
> ** Booting bootflow '<NULL>' with efi_mgr
> Loading Boot0000 'usb 0' failed
> Error reading cluster
> Loading Boot0001 'mmc 1' failed
> EFI boot manager: Cannot load any image
> Boot failed (err=-14)
> Card did not respond to voltage select! : -110
> Card did not respond to voltage select! : -110
> Card did not respond to voltage select! : -110
> Card did not respond to voltage select! : -110

The problem happens in U-Boot shell instead of SPL stage, so it probably
has nothing to do with the bootph-pre-ram change. Even if the pinctrl
isn't available at SPL stage, it should still be picked by the proper
U-Boot.

> My subsequent PR for that lead to issues and it never got applied.
> https://lore.kernel.org/soc/4dd4ffe6-307a-442b-ae99-50c88d4e5b84@freeshell.de/
> 
> This happened back in March, and I just have not had the time nor
> inclination to actually sort it out, and nobody else has stepped up.
> 
> I think I will just pick that one patch back up again and resubmit
> it for this cycle. After all, it doesn't change the configuration, just
> the point at which the configuration should be done.
> 
> If you could let me know if that patch helps, then it would be another
> argument for actually (finally) getting it merged.
> 
> Thanks,
> Conor.

Best regards,
Yao Zi

Re: VisionFive 2 eMMC boot issues.

[Milad]

Thanks to both for the replies.

I built and tested both v2025.10 and v2026.01 on my board (VisionFive 2,
v1.3B, 8GiB), and can confirm both boot the SD card correctly, no cluster
read failure, straight into GRUB and Debian on both. Only v2026.07 fails
for me.

So the regression window looks to be somewhere between v2026.01 and
v2026.07. Full serial logs for both below.

Given the earlier note that this happens in the main U-Boot shell rather
than at SPL stage, I've held off testing the devicetree/pinctrl patch,
since it sounds like that may not be the relevant fix here let me know if
you still think it's worth trying. I'd be happy to look more into it or try
some patches.

--- v2025.10 log ---
U-Boot SPL 2025.10 (Jul 28 2026 - 19:21:15 +0200)
DDR version: dc2e84f0.
Trying to boot from SPI


U-Boot 2025.10 (Jul 28 2026 - 19:21:15 +0200)

CPU: sifive,u74-mc
Model: StarFive VisionFive 2 v1.3B
DRAM: 8 GiB
Core: 159 devices, 29 uclasses, devicetree: board
WDT: Not starting watchdog@13070000
MMC: mmc@16010000: 0, mmc@16020000: 1
Loading Environment from SPIFlash... SF: Detected gd25lq128 with page size
256 Bytes, erase size 4 KiB, total 16 MiB
OK
StarFive EEPROM format v2

--------EEPROM INFO--------
Vendor : StarFive Technology Co., Ltd.
Product full SN: VF7110B1-2318-D008E000-18005137
data version: 0x2
PCB revision: 0xb2
BOM revision: A
Ethernet MAC0 address: 6c:cf:39:00:ba:b0
Ethernet MAC1 address: 6c:cf:39:00:ba:b1
--------EEPROM INFO--------

starfive_7110_pcie pcie@940000000: Starfive PCIe bus probed.
starfive_7110_pcie pcie@9c0000000: Starfive PCIe bus probed.
In: serial@10000000
Out: serial@10000000
Err: serial@10000000
Net: eth0: ethernet@16030000, eth1: ethernet@16040000
starting USB...
Register 5000420 NbrPorts 5
Starting the controller
USB XHCI 1.00
Bus xhci_pci: 2 USB Device(s) found
       scanning usb for storage devices... 0 Storage Device(s) found
Working FDT set to ff71c720
Hit any key to stop autoboot: 0
Card did not respond to voltage select! : -110
** Booting bootflow '<NULL>' with efi_mgr
Booting: Label: mmc 1 Device path:
/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,0000000000000000)/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,6f00000000000000)/SD(1)/SD(1)
...
Loading Linux 6.12.96+deb13-riscv64 ...
Loading initial ramdisk ...
/dev/mmcblk1p2: recovering journal
/dev/mmcblk1p2: clean, 33993/861568 files, 371273/3443456 blocks
Debian GNU/Linux 13 visionfive-2 ttyS0
visionfive-2 login:

--- v2026.01 log ---
U-Boot SPL 2026.01 (Jul 29 2026 - 12:11:02 +0200)
DDR version: dc2e84f0.
Trying to boot from SPI


U-Boot 2026.01 (Jul 29 2026 - 12:11:02 +0200)

CPU: sifive,u74-mc
Model: StarFive VisionFive 2 v1.3B
DRAM: 8 GiB
Core: 158 devices, 29 uclasses, devicetree: board
WDT: Not starting watchdog@13070000
MMC: mmc@16010000: 0, mmc@16020000: 1
Loading Environment from SPIFlash... SF: Detected gd25lq128 with page size
256 Bytes, erase size 4 KiB, total 16 MiB
OK
StarFive EEPROM format v2

--------EEPROM INFO--------
Vendor : StarFive Technology Co., Ltd.
Product full SN: VF7110B1-2318-D008E000-18005137
data version: 0x2
PCB revision: 0xb2
BOM revision: A
Ethernet MAC0 address: 6c:cf:39:00:ba:b0
Ethernet MAC1 address: 6c:cf:39:00:ba:b1
--------EEPROM INFO--------

In: serial@10000000
Out: serial@10000000
Err: serial@10000000
Net: eth0: ethernet@16030000, eth1: ethernet@16040000
starting USB...
Starting the controller
USB XHCI 1.00
Bus xhci_pci: 2 USB Device(s) found
       scanning usb for storage devices... 0 Storage Device(s) found
Working FDT set to ff71cd40
Hit any key to stop autoboot: 0
Card did not respond to voltage select! : -110
** Booting bootflow '<NULL>' with efi_mgr
Booting: Label: mmc 1 Device path:
/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,0000000000000000)/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b,7000000000000000)/SD(1)/SD(1)
...
Loading Linux 6.12.96+deb13-riscv64 ...
Loading initial ramdisk ...
/dev/mmcblk1p2: clean, 33991/861568 files, 371273/3443456 blocks
Debian GNU/Linux 13 visionfive-2 ttyS0
visionfive-2 login:

Thanks for looking into this.
Milad H.


On Wed, 29 Jul 2026, 11:39 Yao Zi, <me@ziyao.cc> wrote:

> On Tue, Jul 28, 2026 at 07:02:02PM +0100, Conor Dooley wrote:
> > On 28/07/2026 15:42, Milad wrote:
> > > Hello everyone,
> > >
> > > Little reminder on this thread regarding my VisionFive 2 MMC boot
> issue.
> > >
> > > Archive link:
> > >
> https://lists.u-boot-project.org/pipermail/u-boot/2026-July/624724.html
> > >
> > > Just following up to see if anyone has had a chance to look at the
> details
> > > in my original message on the link.
> > >
> > > Thanks!
> >
> > I don't think you CCed anyone that cares about the platform on that
> > patch so that's probably the problem there.
> >
> > Is this the same issue that Ilya attempted to fix in the upstream version
> > of the devicetree for this platform?
> > https://lore.kernel.org/all/20260306171939.84568-1-k0tran@altlinux.org/
>
> Quoting logs in the link,
>
> > SD card (mmc1) fails to initialize in U-Boot. Full serial output:
> >
> > Card did not respond to voltage select! : -110
> > ** Booting bootflow '<NULL>' with efi_mgr
> > Loading Boot0000 'usb 0' failed
> > Error reading cluster
> > Loading Boot0001 'mmc 1' failed
> > EFI boot manager: Cannot load any image
> > Boot failed (err=-14)
> > Card did not respond to voltage select! : -110
> > Card did not respond to voltage select! : -110
> > Card did not respond to voltage select! : -110
> > Card did not respond to voltage select! : -110
>
> The problem happens in U-Boot shell instead of SPL stage, so it probably
> has nothing to do with the bootph-pre-ram change. Even if the pinctrl
> isn't available at SPL stage, it should still be picked by the proper
> U-Boot.
>
> > My subsequent PR for that lead to issues and it never got applied.
> >
> https://lore.kernel.org/soc/4dd4ffe6-307a-442b-ae99-50c88d4e5b84@freeshell.de/
> >
> > This happened back in March, and I just have not had the time nor
> > inclination to actually sort it out, and nobody else has stepped up.
> >
> > I think I will just pick that one patch back up again and resubmit
> > it for this cycle. After all, it doesn't change the configuration, just
> > the point at which the configuration should be done.
> >
> > If you could let me know if that patch helps, then it would be another
> > argument for actually (finally) getting it merged.
> >
> > Thanks,
> > Conor.
>
> Best regards,
> Yao Zi
>