linux-kernel - Re: [PATCH v3 04/11] kernel/fork: Initialize mm's PASID

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 4 Feb 2022 16:25:34 -0800
From:   Fenghua Yu <fenghua.yu@...el.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Dave Hansen <dave.hansen@...ux.intel.com>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Peter Zijlstra <peterz@...radead.org>,
        Andy Lutomirski <luto@...nel.org>,
        Tony Luck <tony.luck@...el.com>,
        Lu Baolu <baolu.lu@...ux.intel.com>,
        Joerg Roedel <joro@...tes.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Jacob Pan <jacob.jun.pan@...ux.intel.com>,
        Ashok Raj <ashok.raj@...el.com>,
        Ravi V Shankar <ravi.v.shankar@...el.com>,
        iommu@...ts.linux-foundation.org, x86 <x86@...nel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3 04/11] kernel/fork: Initialize mm's PASID

Hi, Thomas,

On Sat, Feb 05, 2022 at 12:22:12AM +0100, Thomas Gleixner wrote:
> On Fri, Jan 28 2022 at 12:28, Fenghua Yu wrote:
> > A new mm doesn't have a PASID yet when it's created. Initialize
> > the mm's PASID on fork() or for init_mm to INVALID_IOASID (-1).
> 
> I must be missing something here.
> 
> > diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h
> > index aa5f09ca5bcf..c74d1edbac2f 100644
> > --- a/include/linux/sched/mm.h
> > +++ b/include/linux/sched/mm.h
> > @@ -8,6 +8,7 @@
> >  #include <linux/mm_types.h>
> >  #include <linux/gfp.h>
> >  #include <linux/sync_core.h>
> > +#include <linux/ioasid.h>
> >  
> >  /*
> >   * Routines for handling mm_structs
> > @@ -433,4 +434,13 @@ static inline void membarrier_update_current_mm(struct mm_struct *next_mm)
> >  }
> >  #endif
> >  
> > +#ifdef CONFIG_IOMMU_SVA
> > +static inline void mm_pasid_init(struct mm_struct *mm)
> > +{
> > +	mm->pasid = INVALID_IOASID;
> > +}
> > +#else
> > +static inline void mm_pasid_init(struct mm_struct *mm) {}
> > +#endif
> > +
> >  #endif /* _LINUX_SCHED_MM_H */
> 
> So this adds mm_pasid_init() to linux/sched/mm.h which replaces:
> 
> > -static void mm_init_pasid(struct mm_struct *mm)
> > -{
> > -#ifdef CONFIG_IOMMU_SVA
> > -	mm->pasid = INIT_PASID;
> > -#endif
> > -}
> > -
> 
> I.e. already existing code which is initializing mm->pasid with
> INIT_PASID (0) while the replacement initializes it with INVALID_IOASID
> (-1).
> 
> The change log does not have any information about why INIT_PASID is the
> wrong initialization value and why this change is not having any side
> effects.

I should add the following info in the commit message to explain why
change INIT_PASID (0) to INVALID_IOASID (-1):

INIT_PASID (0) is reserved for kernel legacy DMA PASID. It cannot be
allocated to a user process. Initialize the process's PASID to 0 may
cause confusion that why the process uses reserved kernel legacy DMA
PASID. Initializing the PASID to INVALID_IOASID (-1) explicitly
tells the process doesn't have a valid PASID yet initially.

Is it OK for you?

> 
> It neither mentions why having this in a global available header makes
> sense when the only call site is in the C file from which the already
> existing function is removed.

This series defines three helpers mm_pasid_init(), mm_pasid_set(), and
mm_pasid_drop() in mm because they handle the pasid member in mm_struct
and should be part of mm operations. I explained why mm_pasid_set() and
mm_pasid_drop() are defined in mm, but I didn't explain why mm_pasid_init()
is define in mm.

Is it OK to add the following explanation on why mm_pasid_init() is defined?

mm_pasid_init() is defined in mm and replaces mm_init_pasid() because
the PASID init operation initializes the pasid member in mm_struct and
should be part of mm operations.

Thanks,

-Fenghua