[PATCHv5 1/2] block: accumulate memory segment gaps per bio

[Keith Busch]

On Tue, Nov 11, 2025 at 04:26:21AM +0000, Matthew Wilcox wrote:
> On Tue, Oct 14, 2025 at 08:04:55AM -0700, Keith Busch wrote:
> > device's virtual boundary, but the code had been depending on that queue
> > limit to know ahead of time if the request is guaranteed to align to
> > that optimization.
> > 
> > Rather than rely on that queue limit, which many devices may not report,
> > save the lowest set bit of any boundary gap between each segment in the
> > bio while checking the segments. The request stores the value for
> > merging and quickly checking per io if the request can use iova
> > optimizations.
> 
> Hi Keith,
> 
> I just hit this bug:

...
 
> RIP: 0010:bio_get_last_bvec+0x20/0xe0
> Code: 90 90 90 90 90 90 90 90 90 90 55 49 89 f2 48 89 f9 48 89 e5 53 8b 77 2c 8b 47 30 44 8b 4f 28 49 89 f0 49 c1 e0 04 4c 03 47 50 <41> 8b 78 08 41 8b 58 0c 4d 8b 18 29 c7 44 39 cf 4d 89 1a 41 0f 47
> RSP: 0018:ffff88814d1ef8d8 EFLAGS: 00010246
> RAX: 0000000000000000 RBX: ffff8881044b5500 RCX: ffff88811fd23d78
> RDX: ffff88811fd235f8 RSI: 0000000000000000 RDI: ffff88811fd23d78
> RBP: ffff88814d1ef8e0 R08: 0000000000000000 R09: 0000000000020000
> R10: ffff88814d1ef8f0 R11: 0000000000000200 R12: 0000000000000000
> R13: ffff88811fd235f8 R14: 0000000000000000 R15: ffff8881044b5500
> FS:  0000000000000000(0000) GS:ffff8881f6ac9000(0000) knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 0000000000000008 CR3: 000000000263a000 CR4: 0000000000750eb0
> PKRU: 55555554
> Kernel panic - not syncing: Fatal exception
> Kernel Offset: disabled
> ---[ end Kernel panic - not syncing: Fatal exception ]---
> 
> I'm not saying it's definitely your patch; after all, there's 17 of
> my slab patches on top of next-20251110, but when I looked on lore for
> 'bio_get_last_bvec' this was the only patch since 2021 that mentioned it,
> so I thought I'd drop you a note in case you see the bug immediately.
> I'm heading to bed, and will be out tomorrow, so my opportunities to be
> helpful will be limited.

Thanks for the heads up. This is in the path I'd been modifying lately,
so sounds plausible that I introduced the bug. The information here
should be enough for me to make progress: it looks like req->bio is NULL
in your trace, which I did not expect would happen. But it's late here
too, so look with fresh eyes in the morning.