Re: [PATCH v8 17/21] x86,fs/resctrl: Move the resctrl filesystem code to live in /fs/resctrl

[Fenghua Yu]

Hi, James,

This comment was buried in the middle of this huge patch.

To see my previous comment easily, I cut all other irrelevant code.

On 4/11/25 17:18, Fenghua Yu wrote:
> Hi, James,
>
> On 4/11/25 09:42, James Morse wrote:
> > Resctrl is a filesystem interface to hardware that provides cache
> > allocation policy and bandwidth control for groups of tasks or CPUs.
> >
> > To support more than one architecture, resctrl needs to live in /fs/.
> >
> > Move the code that is concerned with the filesystem interface to
> > /fs/resctrl.
> >
> > Signed-off-by: James Morse <james.morse@xxxxxxx>
> > ---
> > Changes since *:
> >   * This patch is generated by a script.
> > ---
> >   arch/x86/kernel/cpu/resctrl/ctrlmondata.c     |  636 ---
> >   arch/x86/kernel/cpu/resctrl/internal.h        |  380 +-
> >   arch/x86/kernel/cpu/resctrl/monitor.c         |  902 +---
> >   arch/x86/kernel/cpu/resctrl/monitor_trace.h   |   18 +-
> >   arch/x86/kernel/cpu/resctrl/pseudo_lock.c     | 1080 +---
> >   .../kernel/cpu/resctrl/pseudo_lock_trace.h    |    2 +
> >   arch/x86/kernel/cpu/resctrl/rdtgroup.c        | 4556 +----------------
> >   fs/resctrl/ctrlmondata.c                      |  660 +++
> >   fs/resctrl/internal.h                         |  442 ++
> >   fs/resctrl/monitor.c                          |  932 ++++
> >   fs/resctrl/monitor_trace.h                    |   33 +
> >   fs/resctrl/pseudo_lock.c                      | 1115 ++++
> >   fs/resctrl/pseudo_lock_trace.h                |   17 +
> >   fs/resctrl/rdtgroup.c                         | 4313 ++++++++++++++++
> >   14 files changed, 7668 insertions(+), 7418 deletions(-)

[SNIP]

> > diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c 
> > b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> > index 326b3048d728..c7a7f0ae373a 100644
> > --- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
> > +++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c 

[SNIP]

> > -static int mkdir_rdt_prepare(struct kernfs_node *parent_kn,
> > -                 const char *name, umode_t mode,
> > -                 enum rdt_group_type rtype, struct rdtgroup **r)
> > -{
> > -    struct rdtgroup *prdtgrp, *rdtgrp;
> > -    unsigned long files = 0;
> > -    struct kernfs_node *kn;
> > -    int ret;
> > -
> > -    prdtgrp = rdtgroup_kn_lock_live(parent_kn);
> > -    if (!prdtgrp) {
> > -        ret = -ENODEV;
> > -        goto out_unlock;
> > -    }
> > -
> > -    /*
> > -     * Check that the parent directory for a monitor group is a
> > -     * "mon_groups" directory.

The follow is my previous original comment on the merge conflict and the 
fix:

> Due to the slight difference between here and the upstream commit 
> 45c2e30bbd64 ("x86/resctrl: Fix rdtgroup_mkdir()'s unlocked use of 
> kernfs_node::name"), merge conflicts.
>
> You may consider to add the following patch in patch #17 to fix the 
> conflict?
>
>         /*
> -        * Check that the parent directory for a monitor group is a 
> "mon_groups"
> -        * directory.
> +        * Check that the parent directory for a monitor group is a
> +        * "mon_groups" directory.
>          */
[SNIP]

Thanks.

-Fenghua