mainline boot fail nvme/block? [BISECTED]

[Genes Lists]

On Mon, 2025-10-13 at 16:46 +0800, Inochi Amaoto wrote:
> On Fri, Oct 10, 2025 at 07:49:34PM -0400, Genes Lists wrote:
> > On Fri, 2025-10-10 at 08:54 -0600, Jens Axboe wrote:
> > > On 10/10/25 8:29 AM, Genes Lists wrote:
> > > > Mainline fails to boot - 6.17.1 works fine.
> > > > Same kernel on an older laptop without any nvme works just
> > > > fine.
> > > > 
> > > > It seems to get stuck enumerating disks within the initramfs
> > > > created by
> > > > dracut.
> > > > 
> > > > ,
...

> > Bisect landed here. (cc linux-pci at vger.kernel.org)
> > Hopefully it is helpful, even though I don't see MSI in lspci
> > output
> > (which is provided below).
> > 
> > gene
> > 
> > 
> > 54f45a30c0d0153d2be091ba2d683ab6db6d1d5b is the first bad commit
> > commit 54f45a30c0d0153d2be091ba2d683ab6db6d1d5b (HEAD)
> > Author: Inochi Amaoto <inochiama at gmail.com>
> > Date:   Thu Aug 14 07:28:32 2025 +0800
> > 
> >     PCI/MSI: Add startup/shutdown for per device domains
> > 
> >     As the RISC-V PLIC cannot apply affinity settings without
> > invoking
> >     irq_enable(), it will make the interrupt unavailble when used
> > as an
> >     underlying interrupt chip for the MSI controller.
> > 
> >     Implement the irq_startup() and irq_shutdown() callbacks for
> > the
> > PCI MSI
> >     and MSI-X templates.
> > 
> >     For chips that specify MSI_FLAG_PCI_MSI_STARTUP_PARENT, the
> > parent
> > startup
> >     and shutdown functions are invoked. That allows the interrupt
> > on
> > the parent
> >     chip to be enabled if the interrupt has not been enabled during
> >     allocation. This is necessary for MSI controllers which use
> > PLIC as
> >     underlying parent interrupt chip.
> > 
> >     Suggested-by: Thomas Gleixner <tglx at linutronix.de>
> >     Signed-off-by: Inochi Amaoto <inochiama at gmail.com>
> >     Signed-off-by: Thomas Gleixner <tglx at linutronix.de>
> >     Tested-by: Chen Wang <unicorn_wang at outlook.com> # Pioneerbox
> >     Reviewed-by: Chen Wang <unicorn_wang at outlook.com>
> >     Acked-by: Bjorn Helgaas <bhelgaas at google.com>
> >     Link: https://lore.kernel.org/all/20250813232835.43458-3-
> > inochiama at gmail.com
> > 
> >  drivers/pci/msi/irqdomain.c | 52
> > ++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  include/linux/msi.h         |  2 ++
> >  2 files changed, 54 insertions(+)
> > 
> > 
...

> 
> 
> I think this is caused by VMD device, which I have a temporary
> solution
> here [1]. Since I have no idea about how VMD works, I hope if anyone
> can help to convert this as an formal fix.
> 
> [1]
> https://lore.kernel.org/all/qs2vydzm6xngul77xuwjli7h757gzfhmb4siiklzo
> gihz5oplw at gsvgn75lib6t/
> 
> Regards,
> Inochi

Thank you Inochi

I tried this patch over 6.18-rc1.

 It get's further than without the patch but around the time I get
prompted for passphrase for the luks partition
(root is not encrypted) it crashes. 

I have uploaded 2 images I took of the screen when this happens and
uploaded them to here:

    https://0x0.st/KSNz.jpg
    https://0x0.st/KSNi.jpg



-- 
Gene
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: This is a digitally signed message part
URL: <http://lists.infradead.org/pipermail/linux-nvme/attachments/20251013/9aee88ea/attachment.sig>