2024-02-20 11:36:15

by Georgi Djakov

Subject: Re: [PATCH] interconnect:Replace mutex with rt_mutex for icc_bw_lock

Hello,

On 20.02.24 9:43, Rumeng Wang wrote:
> From: wangrumeng <[email protected]>
>
> Replace existing mutex with rt_mutex to prevent priority inversion
> between clients, which can cause unacceptable delays in some cases.
>
> Signed-off-by: wangrumeng <[email protected]>

A similar patch [1] has been posted some time ago. Please check the review
comments.

Thanks,
Georgi

[1] https://lore.kernel.org/all/[email protected]/

> ---
> drivers/interconnect/core.c | 15 ++++++++-------
> 1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/interconnect/core.c b/drivers/interconnect/core.c
> index 50bac2d79d9b..467d42cc7e49 100644
> --- a/drivers/interconnect/core.c
> +++ b/drivers/interconnect/core.c
> @@ -14,6 +14,7 @@
> #include <linux/interconnect-provider.h>
> #include <linux/list.h>
> #include <linux/mutex.h>
> +#include <linux/rtmutex.h>
> #include <linux/slab.h>
> #include <linux/of.h>
> #include <linux/overflow.h>
> @@ -28,7 +29,7 @@ static LIST_HEAD(icc_providers);
> static int providers_count;
> static bool synced_state;
> static DEFINE_MUTEX(icc_lock);
> -static DEFINE_MUTEX(icc_bw_lock);
> +static DEFINE_RT_MUTEX(icc_bw_lock);
> static struct dentry *icc_debugfs_dir;
>
> static void icc_summary_show_one(struct seq_file *s, struct icc_node *n)
> @@ -698,7 +699,7 @@ int icc_set_bw(struct icc_path *path, u32 avg_bw, u32 peak_bw)
> if (WARN_ON(IS_ERR(path) || !path->num_nodes))
> return -EINVAL;
>
> - mutex_lock(&icc_bw_lock);
> + rt_mutex_lock(&icc_bw_lock);
>
> old_avg = path->reqs[0].avg_bw;
> old_peak = path->reqs[0].peak_bw;
> @@ -730,7 +731,7 @@ int icc_set_bw(struct icc_path *path, u32 avg_bw, u32 peak_bw)
> apply_constraints(path);
> }
>
> - mutex_unlock(&icc_bw_lock);
> + rt_mutex_unlock(&icc_bw_lock);
>
> trace_icc_set_bw_end(path, ret);
>
> @@ -939,7 +940,7 @@ void icc_node_add(struct icc_node *node, struct icc_provider *provider)
> return;
>
> mutex_lock(&icc_lock);
> - mutex_lock(&icc_bw_lock);
> + rt_mutex_lock(&icc_bw_lock);
>
> node->provider = provider;
> list_add_tail(&node->node_list, &provider->nodes);
> @@ -968,7 +969,7 @@ void icc_node_add(struct icc_node *node, struct icc_provider *provider)
> node->avg_bw = 0;
> node->peak_bw = 0;
>
> - mutex_unlock(&icc_bw_lock);
> + rt_mutex_unlock(&icc_bw_lock);
> mutex_unlock(&icc_lock);
> }
> EXPORT_SYMBOL_GPL(icc_node_add);
> @@ -1094,7 +1095,7 @@ void icc_sync_state(struct device *dev)
> return;
>
> mutex_lock(&icc_lock);
> - mutex_lock(&icc_bw_lock);
> + rt_mutex_lock(&icc_bw_lock);
> synced_state = true;
> list_for_each_entry(p, &icc_providers, provider_list) {
> dev_dbg(p->dev, "interconnect provider is in synced state\n");
> @@ -1107,7 +1108,7 @@ void icc_sync_state(struct device *dev)
> }
> }
> }
> - mutex_unlock(&icc_bw_lock);
> + rt_mutex_unlock(&icc_bw_lock);
> mutex_unlock(&icc_lock);
> }
> EXPORT_SYMBOL_GPL(icc_sync_state);

2024-02-27 03:41:16

by Rumeng Wang

Subject: [PATCH] interconnect:Replace mutex with rt_mutex for icc_bw_lock

Hello, it has been a week since the last reply. Is there any progress on this issue?
Thanks.