[PATCH] arm_mpam: return allocation errors as error pointers

[PATCH] arm_mpam: return allocation errors as error pointers

[Sang-Hoon Choi]

resctrl_arch_mon_ctx_alloc() stores the result of monitor allocation in
an allocated int. If monitor allocation fails, it still returns that pointer.
Callers check IS_ERR(), so they treat the failed allocation as a valid
monitor context and pass the error value on as a monitor index.

For example, if the wait for a free monitor ends because a signal is
pending, the stored result is still -ENOSPC. Other allocation errors
also reach the same return path.

Free the temporary allocation and return ERR_PTR(err) for negative
results. Leave successful indices and USE_PRE_ALLOCATED unchanged.

Fixes: 2a3c79c61539 ("arm_mpam: resctrl: Allow resctrl to allocate monitors")
Reported-by: Changyul Lee <lcy8047@gmail.com>
Link: https://github.com/lcy8047/radix-tree-bug-temp/blob/main/bug_list/idr-ida-xarray/wrong-id-value/IDR-BUG-33-mpam-monitor-id.md
Assisted-by: Sang-Hoon Choi <csh0052@gmail.com>
Signed-off-by: Sang-Hoon Choi <csh0052@gmail.com>
---
Compile-tested on arm64 with allmodconfig and W=1 against mainline
6f8319e3e9a44dd537d17f41565a8453c560a581. No hardware runtime test was
performed.

 drivers/resctrl/mpam_resctrl.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/resctrl/mpam_resctrl.c b/drivers/resctrl/mpam_resctrl.c
index 9d223057953a..945be7465ad3 100644
--- a/drivers/resctrl/mpam_resctrl.c
+++ b/drivers/resctrl/mpam_resctrl.c
@@ -413,6 +413,13 @@ void *resctrl_arch_mon_ctx_alloc(struct rdt_resource *r,
 	} while (*ret == -ENOSPC && !signal_pending(current));
 	finish_wait(&resctrl_mon_ctx_waiters, &wait);
 
+	if (*ret < 0) {
+		int err = *ret;
+
+		kfree(ret);
+		return ERR_PTR(err);
+	}
+
 	return ret;
 }
 
-- 
2.43.0

Re: [PATCH] arm_mpam: return allocation errors as error pointers

[Ben Horgan]

Hi Sang-Hoon,

On 9/30/26 08:08, Sang-Hoon Choi wrote:
> resctrl_arch_mon_ctx_alloc() stores the result of monitor allocation in
> an allocated int. If monitor allocation fails, it still returns that pointer.
> Callers check IS_ERR(), so they treat the failed allocation as a valid
> monitor context and pass the error value on as a monitor index.
> 
> For example, if the wait for a free monitor ends because a signal is
> pending, the stored result is still -ENOSPC. Other allocation errors
> also reach the same return path.

Thanks for report and fix. Is this addressing a hypothetical problem or
have you seen resctrl_arch_mon_ctx_alloc() fail?

> 
> Free the temporary allocation and return ERR_PTR(err) for negative
> results. Leave successful indices and USE_PRE_ALLOCATED unchanged.
> 
> Fixes: 2a3c79c61539 ("arm_mpam: resctrl: Allow resctrl to allocate monitors")
> Reported-by: Changyul Lee <lcy8047@gmail.com>
> Link: https://github.com/lcy8047/radix-tree-bug-temp/blob/main/bug_list/idr-ida-xarray/wrong-id-value/IDR-BUG-33-mpam-monitor-id.md

I don't see anything at this link. This line can just go.

> Assisted-by: Sang-Hoon Choi <csh0052@gmail.com>

Should this be?

Assisted-by: LLM

Thanks,

Ben

> Signed-off-by: Sang-Hoon Choi <csh0052@gmail.com>
> ---
> Compile-tested on arm64 with allmodconfig and W=1 against mainline
> 6f8319e3e9a44dd537d17f41565a8453c560a581. No hardware runtime test was
> performed.
> 
>  drivers/resctrl/mpam_resctrl.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/resctrl/mpam_resctrl.c b/drivers/resctrl/mpam_resctrl.c
> index 9d223057953a..945be7465ad3 100644
> --- a/drivers/resctrl/mpam_resctrl.c
> +++ b/drivers/resctrl/mpam_resctrl.c
> @@ -413,6 +413,13 @@ void *resctrl_arch_mon_ctx_alloc(struct rdt_resource *r,
>  	} while (*ret == -ENOSPC && !signal_pending(current));
>  	finish_wait(&resctrl_mon_ctx_waiters, &wait);
>  
> +	if (*ret < 0) {
> +		int err = *ret;
> +
> +		kfree(ret);
> +		return ERR_PTR(err);
> +	}
> +
>  	return ret;
>  }
>  

Re: [PATCH] arm_mpam: return allocation errors as error pointers

[Ben Horgan]

Hi Sang-Hoon,

On 30/09/2026 08:08, Sang-Hoon Choi wrote:
> resctrl_arch_mon_ctx_alloc() stores the result of monitor allocation in
> an allocated int. If monitor allocation fails, it still returns that pointer.
> Callers check IS_ERR(), so they treat the failed allocation as a valid
> monitor context and pass the error value on as a monitor index.
> 
> For example, if the wait for a free monitor ends because a signal is
> pending, the stored result is still -ENOSPC. Other allocation errors
> also reach the same return path.
> 
> Free the temporary allocation and return ERR_PTR(err) for negative
> results. Leave successful indices and USE_PRE_ALLOCATED unchanged.
> 
> Fixes: 2a3c79c61539 ("arm_mpam: resctrl: Allow resctrl to allocate monitors")
> Reported-by: Changyul Lee <lcy8047@gmail.com>
> Link: https://github.com/lcy8047/radix-tree-bug-temp/blob/main/bug_list/idr-ida-xarray/wrong-id-value/IDR-BUG-33-mpam-monitor-id.md
> Assisted-by: Sang-Hoon Choi <csh0052@gmail.com>
> Signed-off-by: Sang-Hoon Choi <csh0052@gmail.com>
> ---
> Compile-tested on arm64 with allmodconfig and W=1 against mainline
> 6f8319e3e9a44dd537d17f41565a8453c560a581. No hardware runtime test was
> performed.
> 
>  drivers/resctrl/mpam_resctrl.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/resctrl/mpam_resctrl.c b/drivers/resctrl/mpam_resctrl.c
> index 9d223057953a..945be7465ad3 100644
> --- a/drivers/resctrl/mpam_resctrl.c
> +++ b/drivers/resctrl/mpam_resctrl.c
> @@ -413,6 +413,13 @@ void *resctrl_arch_mon_ctx_alloc(struct rdt_resource *r,
>  	} while (*ret == -ENOSPC && !signal_pending(current));
>  	finish_wait(&resctrl_mon_ctx_waiters, &wait);
>  
> +	if (*ret < 0) {
> +		int err = *ret;
> +
> +		kfree(ret);
> +		return ERR_PTR(err);
> +	}
> +
>  	return ret;

This seems like a good place to use the cleanup.h helpers, __free(kfree) and then pass on the
pointer with return_ptr(ret).

Thanks,

Ben

>  }
>