Closed AnalogBot closed 3 months ago
@AnalogBot thanks for the detailed report, and doing some sleuthing! I've made a patch for vsl4 that I can test later and then backport to vsl . The single card in my dev box is vsl4 atm, hence doing that one first: https://github.com/RemixVSL/iomemory-vsl4/compare/main...vsl%23131. The MQ code is defunct atm afaik, but I've updated it irrespective.
@AnalogBot please check https://github.com/RemixVSL/iomemory-vsl/tree/bug%23131, which contains the fix. I'll merge it with master if you can confirm it works as desired? (otherwise we have to wait for me to swap cards at some point)
https://github.com/RemixVSL/iomemory-vsl/tree/bug%23131 builds and the module loads without issue. The device appears to be operating normally with this version loaded.
Looks like that fixes it. Thank you for your work on this!
@AnalogBot thanks for the bug report and the feedback!
I can confirm it works with kernel 6.9.4 on Fedora 40. Great work guys!
Bug description
Build is failing on 6.9.2
How to reproduce
Possible solution
Note: I'm not really familiar with kernel dev, I don't even really know enough C to be anything other then dangerous. But I tried to dig in to see if it was something simple. This is what I found, but it's outside of my capabilities to fix.
Looks like in 6.9 these functions now require an additional 'queue_limits' argument. blk_alloc_disk blk_mq_alloc_disk
https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.9 - Looks like most in tree block device drivers were updated to support this change. https://github.com/openzfs/zfs/pull/16033/commits/c5ad47920fdbf20dfe56f77368a15568c17b57eb - ZFS also recently updated to support.
Environment information
Information about the system the module is used on
Linux thebria 6.9.2-arch1-1 #1 SMP PREEMPT_DYNAMIC Sun, 26 May 2024 01:30:29 +0000 x86_64 GNU/Linux
fio-status: driver isn't loaded
lspci -b -nn 0c:00.0 Mass storage controller [0180]: SanDisk ioDrive2 [1aed:2001] (rev 04)