[PATCH] lib: test_hmm: use device devt for coherent device range selection

[PATCH] lib: test_hmm: use device devt for coherent device range selection

[Stanislav Kinsburskii]

Commit af69016dab96 ("lib: test_hmm: implement a device release method")
moved the initial dmirror_allocate_chunk() call before cdev_device_add().
That means the struct cdev has not been added yet, so cdev_add() has not
initialized mdevice->cdevice.dev.

The coherent-device range selection uses the device minor to choose between
spm_addr_dev0 and spm_addr_dev1. Reading MINOR(mdevice->cdevice.dev) before
cdev_add() therefore always sees an uninitialized dev_t. As a result, both
coherent devices select the same physical range, and adding the second
device fails due to the overlapping dev_pagemap range.

Use mdevice->device.devt instead. It is initialized in
dmirror_device_init() before dmirror_allocate_chunk() is called and is the
same dev_t later passed to cdev_device_add().

Fixes: af69016dab96 ("lib: test_hmm: implement a device release method")
Signed-off-by: Stanislav Kinsburskii <skinsburskii@gmail.com>
---
 lib/test_hmm.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/test_hmm.c b/lib/test_hmm.c
index 9c59d1ceb5b5..c4adbf98fac7 100644
--- a/lib/test_hmm.c
+++ b/lib/test_hmm.c
@@ -581,7 +581,7 @@ static int dmirror_allocate_chunk(struct dmirror_device *mdevice,
 		devmem->pagemap.type = MEMORY_DEVICE_PRIVATE;
 		break;
 	case HMM_DMIRROR_MEMORY_DEVICE_COHERENT:
-		devmem->pagemap.range.start = (MINOR(mdevice->cdevice.dev) - 2) ?
+		devmem->pagemap.range.start = (MINOR(mdevice->device.devt) - 2) ?
 							spm_addr_dev0 :
 							spm_addr_dev1;
 		devmem->pagemap.range.end = devmem->pagemap.range.start +

Re: [PATCH] lib: test_hmm: use device devt for coherent device range selection

[Andrew Morton]

On Mon, 29 Jun 2026 16:30:14 -0700 Stanislav Kinsburskii <skinsburskii@gmail.com> wrote:

> Commit af69016dab96 ("lib: test_hmm: implement a device release method")
> moved the initial dmirror_allocate_chunk() call before cdev_device_add().
> That means the struct cdev has not been added yet, so cdev_add() has not
> initialized mdevice->cdevice.dev.
> 
> The coherent-device range selection uses the device minor to choose between
> spm_addr_dev0 and spm_addr_dev1. Reading MINOR(mdevice->cdevice.dev) before
> cdev_add() therefore always sees an uninitialized dev_t. As a result, both
> coherent devices select the same physical range, and adding the second
> device fails due to the overlapping dev_pagemap range.
> 
> Use mdevice->device.devt instead. It is initialized in
> dmirror_device_init() before dmirror_allocate_chunk() is called and is the
> same dev_t later passed to cdev_device_add().

Thanks.

> Fixes: af69016dab96 ("lib: test_hmm: implement a device release method")

We should cc the people involved with that commit.  Added.

> --- a/lib/test_hmm.c
> +++ b/lib/test_hmm.c
> @@ -581,7 +581,7 @@ static int dmirror_allocate_chunk(struct dmirror_device *mdevice,
>  		devmem->pagemap.type = MEMORY_DEVICE_PRIVATE;
>  		break;
>  	case HMM_DMIRROR_MEMORY_DEVICE_COHERENT:
> -		devmem->pagemap.range.start = (MINOR(mdevice->cdevice.dev) - 2) ?
> +		devmem->pagemap.range.start = (MINOR(mdevice->device.devt) - 2) ?
>  							spm_addr_dev0 :
>  							spm_addr_dev1;
>  		devmem->pagemap.range.end = devmem->pagemap.range.start +
> 

af69016dab96 had cc:stable, so I'll add cc:stable to this patch, to
help ensure that it lands in the appropriate places.

Re: [PATCH] lib: test_hmm: use device devt for coherent device range selection

[Alistair Popple]

On 2026-06-30 at 09:30 +1000, Stanislav Kinsburskii <skinsburskii@gmail.com> wrote...
> Commit af69016dab96 ("lib: test_hmm: implement a device release method")
> moved the initial dmirror_allocate_chunk() call before cdev_device_add().
> That means the struct cdev has not been added yet, so cdev_add() has not
> initialized mdevice->cdevice.dev.
> 
> The coherent-device range selection uses the device minor to choose between
> spm_addr_dev0 and spm_addr_dev1. Reading MINOR(mdevice->cdevice.dev) before
> cdev_add() therefore always sees an uninitialized dev_t. As a result, both
> coherent devices select the same physical range, and adding the second
> device fails due to the overlapping dev_pagemap range.
> 
> Use mdevice->device.devt instead. It is initialized in
> dmirror_device_init() before dmirror_allocate_chunk() is called and is the
> same dev_t later passed to cdev_device_add().
> 
> Fixes: af69016dab96 ("lib: test_hmm: implement a device release method")
> Signed-off-by: Stanislav Kinsburskii <skinsburskii@gmail.com>
> ---
>  lib/test_hmm.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/test_hmm.c b/lib/test_hmm.c
> index 9c59d1ceb5b5..c4adbf98fac7 100644
> --- a/lib/test_hmm.c
> +++ b/lib/test_hmm.c
> @@ -581,7 +581,7 @@ static int dmirror_allocate_chunk(struct dmirror_device *mdevice,
>  		devmem->pagemap.type = MEMORY_DEVICE_PRIVATE;
>  		break;
>  	case HMM_DMIRROR_MEMORY_DEVICE_COHERENT:
> -		devmem->pagemap.range.start = (MINOR(mdevice->cdevice.dev) - 2) ?
> +		devmem->pagemap.range.start = (MINOR(mdevice->device.devt) - 2) ?

Makes sense to me. As Andrew mentioned we should cc: stable, but I assume he
will add that so:

Reviewed-by: Alistair Popple <apopple@nvidia.com>

>  							spm_addr_dev0 :
>  							spm_addr_dev1;
>  		devmem->pagemap.range.end = devmem->pagemap.range.start +
> 
> 
>