Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1190157 > unrolled thread

[PATCH] x86, pat: Add comments to cachemode translation tables

Back to article view | Back to linux.kernel

Contents

  [PATCH] x86, pat: Add comments to cachemode translation tables Toshi Kani <toshi.kani@hp.com> - 2015-07-22 20:10 +0200
    Re: [PATCH] x86, pat: Add comments to cachemode translation  tables "Jan Beulich" <JBeulich@suse.com> - 2015-07-23 08:50 +0200
      Re: [PATCH] x86, pat: Add comments to cachemode translation tables Thomas Gleixner <tglx@linutronix.de> - 2015-08-02 12:10 +0200
        Re: [PATCH] x86, pat: Add comments to cachemode translation tables Toshi Kani <toshi.kani@hp.com> - 2015-08-03 17:10 +0200

#1190157 — [PATCH] x86, pat: Add comments to cachemode translation tables

Add comments to the cachemode translation tables to clarify that
the default values are set as minimal supported mode, which are
necessary to handle WC and WT fallback to UC- when they are not
enabled.

Signed-off-by: Toshi Kani <toshi.kani@hp.com>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
---
Ingo, please replace the patch below with this patch.
https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg937806.html
---
 arch/x86/mm/init.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index 8533b46..1d8a83d 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -30,8 +30,11 @@
 /*
  * Tables translating between page_cache_type_t and pte encoding.
  *
- * Minimal supported modes are defined statically, they are modified
- * during bootup if more supported cache modes are available.
+ * The default values are defined statically as minimal supported mode;
+ * WC and WT fall back to UC-.  pat_init() updates these values to support
+ * more cache modes, WC and WT, when it is safe to do so.  See pat_init()
+ * for the details.  Note, __early_ioremap() used during early boot-time
+ * takes pgprot_t (pte encoding) and does not use these tables.
  *
  *   Index into __cachemode2pte_tbl[] is the cachemode.
  *
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1190661 — Re: [PATCH] x86, pat: Add comments to cachemode translation tables

>>> On 22.07.15 at 20:06, <toshi.kani@hp.com> wrote:
> Add comments to the cachemode translation tables to clarify that
> the default values are set as minimal supported mode, which are
> necessary to handle WC and WT fallback to UC- when they are not
> enabled.

Wait - shouldn't WT fall back to UC (so to not be affected by WC
MTRR settings)?

Jan

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1198323

On Thu, 23 Jul 2015, Toshi Kani wrote:
> On Thu, 2015-07-23 at 09:36 -0600, Jan Beulich wrote:
> > > 
> > > > > On 23.07.15 at 17:25, <toshi.kani@hp.com> wrote:
> > > Yes, I agree with you.  But such risk is very low -- 1) the regular 
> > > case
> > > (no fallback) is used most of the cases, 2) the code using WT knows 
> > > what
> > > type of memory it is dealing with.  For example, pmem may map NVDIMM 
> > > with
> > > WT, and any sane BIOS sets MTRR to WB for NVDIMM. 
> > 
> > Do the words "sane" and "BIOS" really fit together in your opinion?
> 
> :-)
> 
> Anyway, I am not disagreeing with you... When UC is ready for both regular
> memory and IO memory, it should be changed to fall back to UC.

What's the resolution of this discussion? Is that patch correct as is
or do we get an updated version?

Thanks,

	tglx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1198931

On Sun, 2015-08-02 at 12:07 +0200, Thomas Gleixner wrote:
> On Thu, 23 Jul 2015, Toshi Kani wrote:
> > On Thu, 2015-07-23 at 09:36 -0600, Jan Beulich wrote:
> > > > 
> > > > > > On 23.07.15 at 17:25, <toshi.kani@hp.com> wrote:
> > > > Yes, I agree with you.  But such risk is very low -- 1) the regular 
> > > > case
> > > > (no fallback) is used most of the cases, 2) the code using WT knows 
> > > > what
> > > > type of memory it is dealing with.  For example, pmem may map NVDIMM 
> > > > 
> > > > with
> > > > WT, and any sane BIOS sets MTRR to WB for NVDIMM. 
> > > 
> > > Do the words "sane" and "BIOS" really fit together in your opinion?
> > 
> > :-)
> > 
> > Anyway, I am not disagreeing with you... When UC is ready for both 
> > regular
> > memory and IO memory, it should be changed to fall back to UC.
> 
> What's the resolution of this discussion? Is that patch correct as is
> or do we get an updated version?

Yes, this patch is correct and we are in agreement.

What Jan mentioned about falling back to UC, instead of UC-, is a separate
item, and the code does not support it yet.

Thanks,
-Toshi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web