Thread (20 messages) 20 messages, 3 authors, 2025-09-28

Re: [PATCH 1/9] alpha: Convert mapping routine to rely on physical address

From: Magnus Lindholm <linmag7@gmail.com>
Date: 2025-09-27 19:20:28
Also in: linux-alpha, linux-iommu, linux-mips, lkml, sparclinux, virtualization, xen-devel

quoted
Suggest testing the same branch with the alpha patch reverted just to
rule out any issue in the core code. If it reproduces suggest to
bisect Leon's branch.
Hi again, I've booted up the ES40 again with the kernel build from Leons
branch, it boots up but message log is full off messages like
"EXT4-fs error (device sda4): ext4_find_extent:939: inode
#16257327: comm init: pblk 65114257 bad header/extent:
invalid magic"

The filesystem is broken after just booting with the kernel.
This time fsck did not fix it, I needed to re-install gentoo stage3.
So it's for sure reproducible as well as destructive.  It's not possible to
revert all the commits individually, since this will leave the source tree
in a state where the kernel doesn't build. I've started off by reverting
the following commits:

e78a9d72517a88faa6f16dab4d1c6f966ed378ae
(dma-mapping: remove unused map_page callback)

d459e3b80ad1c81bf596d63d2e3347cf8c7bb0d9
(alpha: Convert mapping routine to rely on physical address)

3cd47242d513050d7a81ac6e7020fd3ef5462ad4
(block-dma: properly take MMIO path)

7950995bef32aa7e5f74699c7d0fdac41d2dad14
 (block-dma: migrate to dma_map_phys instead of map_page)


After reverting the above commits, I'm able to build a working kernel,
that is, no filesystem corruption occurs. I'll take a closer look at this
after the weekend.

Regards

Magnus
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help