Build of 6.18.20.1 Linux kernel for WSL fails (dxgkrnl/__dma_fence_is_later signature mismatch)

[faratech]

Windows Version

Microsoft Windows [Version 10.0.26200.8117]

WSL Version

2.7.1.0

Are you using WSL 1 or WSL 2?

• WSL 2
• WSL 1

Kernel Version

6.6.123.2-microsoft-standard-WSL2+ (running; 6.18.20.1 is what fails to build)

Distro Version

Ubuntu 25.10

Other Software

• GCC 15.2.0 (Ubuntu 15.2.0-4ubuntu4)
• Host arch: aarch64 (Snapdragon X Elite)

Repro Steps

git clone --depth 1 -b linux-msft-wsl-6.18.20.1 \
    https://github.com/microsoft/WSL2-Linux-Kernel.git
cd WSL2-Linux-Kernel
cp Microsoft/config-wsl-arm64 .config      # or config-wsl
make ARCH=arm64 olddefconfig               # or ARCH=x86
make ARCH=arm64 -j"$(nproc)" Image         # or bzImage

Expected Behavior

Kernel builds cleanly from the linux-msft-wsl-6.18.20.1 release tag using the shipped default config.

Actual Behavior

drivers/hv/dxgkrnl/dxgsyncfile.c: In function 'dxgdmafence_signaled':
drivers/hv/dxgkrnl/dxgsyncfile.c:449:46: error: passing argument 1 of
  '__dma_fence_is_later' makes pointer from integer without a cast
  [-Wint-conversion]
  449 |     return __dma_fence_is_later(syncpoint->fence_value, fence->seqno,
./include/linux/dma-fence.h:471:59: note: expected 'struct dma_fence *'
  but argument is of type 'u64'
drivers/hv/dxgkrnl/dxgsyncfile.c:450:42: error: passing argument 3 of
  '__dma_fence_is_later' makes integer from pointer without a cast
  [-Wint-conversion]

Root cause

__dma_fence_is_later changed signature between 6.6.y and 6.18.y:

Tag	include/linux/dma-fence.h
linux-msft-wsl-6.6.123.2	(u64 f1, u64 f2, const struct dma_fence_ops *ops)
linux-msft-wsl-6.18.20.1	(struct dma_fence *fence, u64 f1, u64 f2)

Every other in-tree caller was updated to the new form (dma-fence-chain.c, sw_sync.c, xe_sched_job.c, xe_hw_fence.c), but drivers/hv/dxgkrnl/dxgsyncfile.c:449 still uses the old calling convention — so dxgkrnl is inconsistent with the very header it includes. Verified by diffing the file between the two tags: the dxgkrnl line is byte-identical in 6.6.123.2 and 6.18.20.1 despite the header change, i.e., the driver was simply missed during the 6.6 → 6.18 rebase.

The file is arch-independent and CONFIG_DXGKRNL=y in both config-wsl and config-wsl-arm64, so this reproduces on any host building the release tag from source. Confirmed firsthand on aarch64.

Fix

One-line, pure API-compat, no semantic change:

--- a/drivers/hv/dxgkrnl/dxgsyncfile.c
+++ b/drivers/hv/dxgkrnl/dxgsyncfile.c
@@ -446,8 +446,7 @@ static bool dxgdmafence_signaled(struct dma_fence *fence)
 	syncpoint = to_syncpoint(fence);
 	if (syncpoint == 0)
 		return true;
-	return __dma_fence_is_later(syncpoint->fence_value, fence->seqno,
-				    fence->ops);
+	return __dma_fence_is_later(fence, syncpoint->fence_value, fence->seqno);
 }

Builds cleanly on aarch64 with config-wsl-arm64 + this patch.

Why this tag probably shipped (compiler note)

Both Microsoft/config-wsl and Microsoft/config-wsl-arm64 on this tag begin with:

CONFIG_CC_VERSION_TEXT="gcc (GCC) 13.2.0"
CONFIG_GCC_VERSION=130200

The -Wint-conversion / -Wincompatible-pointer-types diagnostics this bug triggers were warnings on GCC ≤ 13 but were promoted to errors by default in GCC 14. So on GCC 13 this code warns but still produces a working image; on any current mainstream distro toolchain (Debian 13, Ubuntu 24.04+, Fedora 40+, Arch, openSUSE Tumbleweed — all GCC 14 or 15), it's a hard failure. That strongly suggests whatever verifies the public tag before publishing is still on GCC 13, which means this class of regression slips through silently. Bumping release-verification to GCC ≥14 would catch all four of the recent "tag doesn't build" reports (#12431, #13086, #13511, this one) going forward.

Diagnostic Logs

N/A — pre-boot compile failure, not a runtime issue.

---

cc @chessturo — flagging since you cut the linux-msft-wsl-6.18.20.1 tag.

[github-actions]

Logs are required for review from WSL team

If this a feature request, please reply with '/feature'. If this is a question, reply with '/question'.
Otherwise, please attach logs by following the instructions below, your issue will not be reviewed unless they are added. These logs will help us understand what is going on in your machine.

How to collect WSL logs

Download and execute collect-wsl-logs.ps1 in an administrative powershell prompt:

Invoke-WebRequest -UseBasicParsing "https://raw.githubusercontent.com/microsoft/WSL/master/diagnostics/collect-wsl-logs.ps1" -OutFile collect-wsl-logs.ps1
Set-ExecutionPolicy Bypass -Scope Process -Force
.\collect-wsl-logs.ps1

The script will output the path of the log file once done.

If this is a networking issue, please use .\collect-wsl-logs.ps1 -LogProfile networking instead, following the instructions in Collect WSL logs for networking issues

Once completed please upload the output files to this GitHub issue.

See Collect WSL logs (recommended method).

If you choose to email these logs instead of attaching them to the bug, please send them to wsl-gh-logs@microsoft.com with the GitHub issue number in the subject, and include a link to your GitHub issue comment in the message body, and reply with '/emailed-logs'.

[faratech]

This is a compile-time failure while building the public release tag from source — no kernel image, WSL VM, or runtime exists at the point of failure, so collect-wsl-logs.ps1 has nothing relevant to collect. The full compiler output, root cause, and one-line fix are already in the issue body.

[asaf400]

I can confirm a successful build on amd64 as well using the fix.

Running it now 👍
Linux Lenovo-AsafL 6.18.20.1-microsoft-standard-WSL2+ #1 SMP PREEMPT_DYNAMIC Mon Apr 13 13:14:20 IDT 2026 x86_64 GNU/Linux

[dschulman-repay]

Similar build failure seen on WSL 2.6.3.0 under Windows 10.0.26200.8037, running Ubuntu 24.04.4.

drivers/hv/dxgkrnl/dxgsyncfile.c: In function ‘dxgdmafence_signaled’:
drivers/hv/dxgkrnl/dxgsyncfile.c:449:46: error: passing argument 1 of ‘__dma_fence_is_later’ makes pointer from integer without a cast [-Wint-conversion]
  449 |         return __dma_fence_is_later(syncpoint->fence_value, fence->seqno,
      |                                     ~~~~~~~~~^~~~~~~~~~~~~
      |                                              |
      |                                              u64 {aka long long unsigned int}
In file included from ./include/linux/sync_file.h:20,
                 from drivers/hv/dxgkrnl/dxgsyncfile.h:17,
                 from drivers/hv/dxgkrnl/dxgsyncfile.c:36:
./include/linux/dma-fence.h:471:59: note: expected ‘struct dma_fence *’ but argument is of type ‘u64’ {aka ‘long long unsigned int’}
  471 | static inline bool __dma_fence_is_later(struct dma_fence *fence, u64 f1, u64 f2)
      |                                         ~~~~~~~~~~~~~~~~~~^~~~~
drivers/hv/dxgkrnl/dxgsyncfile.c:450:42: error: passing argument 3 of ‘__dma_fence_is_later’ makes integer from pointer without a cast [-Wint-conversion]
  450 |                                     fence->ops);
      |                                     ~~~~~^~~~~
      |                                          |
      |                                          const struct dma_fence_ops *
./include/linux/dma-fence.h:471:78: note: expected ‘u64’ {aka ‘long long unsigned int’} but argument is of type ‘const struct dma_fence_ops *’
  471 | static inline bool __dma_fence_is_later(struct dma_fence *fence, u64 f1, u64 f2)
      |                                                                          ~~~~^~
make[5]: *** [scripts/Makefile.build:287: drivers/hv/dxgkrnl/dxgsyncfile.o] Error 1
make[4]: *** [scripts/Makefile.build:544: drivers/hv/dxgkrnl] Error 2
make[3]: *** [scripts/Makefile.build:544: drivers/hv] Error 2
make[2]: *** [scripts/Makefile.build:544: drivers] Error 2
make[1]: *** [/home/<userid>/Downloads/WSL2-Linux-Kernel-linux-msft-wsl-6.18.20.1/Makefile:2021: .] Error 2
make: *** [Makefile:248: __sub-make] Error 2

This appears to be fixed by @faratech's patch.