From: Peter Zijlstra <peterz@infradead.org>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
x86@kernel.org, Bin Yang <bin.yang@intel.com>,
Dave Hansen <dave.hansen@intel.com>,
Mark Gross <mark.gross@intel.com>
Subject: Re: [patch V3 02/11] x86/mm/cpa: Split, rename and clean up try_preserve_large_page()
Date: Tue, 18 Sep 2018 10:19:09 +0200 [thread overview]
Message-ID: <20180918081909.GI24106@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20180917143545.830507216@linutronix.de>
On Mon, Sep 17, 2018 at 04:29:08PM +0200, Thomas Gleixner wrote:
> @@ -1288,23 +1287,23 @@ static int __change_page_attr(struct cpa
> err = split_large_page(cpa, kpte, address);
> if (!err) {
> /*
> + * Do a global flush tlb after splitting the large page
> + * and before we do the actual change page attribute in the PTE.
> + *
> + * With out this, we violate the TLB application note, that says
> + * "The TLBs may contain both ordinary and large-page
> * translations for a 4-KByte range of linear addresses. This
> * may occur if software modifies the paging structures so that
> * the page size used for the address range changes. If the two
> * translations differ with respect to page frame or attributes
> * (e.g., permissions), processor behavior is undefined and may
> * be implementation-specific."
> + *
> + * We do this global tlb flush inside the cpa_lock, so that we
> * don't allow any other cpu, with stale tlb entries change the
> * page attribute in parallel, that also falls into the
> * just split large page entry.
> + */
> flush_tlb_all();
> goto repeat;
> }
this made me look at the tlb invalidation of that thing again; do we
want something like the below?
---
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -285,16 +285,6 @@ static void cpa_flush_all(unsigned long
on_each_cpu(__cpa_flush_all, (void *) cache, 1);
}
-static void __cpa_flush_range(void *arg)
-{
- /*
- * We could optimize that further and do individual per page
- * tlb invalidates for a low number of pages. Caveat: we must
- * flush the high aliases on 64bit as well.
- */
- __flush_tlb_all();
-}
-
static void cpa_flush_range(unsigned long start, int numpages, int cache)
{
unsigned int i, level;
@@ -303,7 +293,7 @@ static void cpa_flush_range(unsigned lon
BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
WARN_ON(PAGE_ALIGN(start) != start);
- on_each_cpu(__cpa_flush_range, NULL, 1);
+ flush_tlb_all();
if (!cache)
return;
@@ -1006,14 +996,24 @@ __split_large_page(struct cpa_data *cpa,
__set_pmd_pte(kpte, address, mk_pte(base, __pgprot(_KERNPG_TABLE)));
/*
- * Intel Atom errata AAH41 workaround.
+ * Do a global flush tlb after splitting the large page
+ * and before we do the actual change page attribute in the PTE.
*
- * The real fix should be in hw or in a microcode update, but
- * we also probabilistically try to reduce the window of having
- * a large TLB mixed with 4K TLBs while instruction fetches are
- * going on.
+ * Without this, we violate the TLB application note, that says
+ * "The TLBs may contain both ordinary and large-page
+ * translations for a 4-KByte range of linear addresses. This
+ * may occur if software modifies the paging structures so that
+ * the page size used for the address range changes. If the two
+ * translations differ with respect to page frame or attributes
+ * (e.g., permissions), processor behavior is undefined and may
+ * be implementation-specific."
+ *
+ * We do this global tlb flush inside the cpa_lock, so that we
+ * don't allow any other cpu, with stale tlb entries change the
+ * page attribute in parallel, that also falls into the
+ * just split large page entry.
*/
- __flush_tlb_all();
+ flush_tlb_all();
spin_unlock(&pgd_lock);
return 0;
@@ -1538,28 +1538,8 @@ static int __change_page_attr(struct cpa
* We have to split the large page:
*/
err = split_large_page(cpa, kpte, address);
- if (!err) {
- /*
- * Do a global flush tlb after splitting the large page
- * and before we do the actual change page attribute in the PTE.
- *
- * With out this, we violate the TLB application note, that says
- * "The TLBs may contain both ordinary and large-page
- * translations for a 4-KByte range of linear addresses. This
- * may occur if software modifies the paging structures so that
- * the page size used for the address range changes. If the two
- * translations differ with respect to page frame or attributes
- * (e.g., permissions), processor behavior is undefined and may
- * be implementation-specific."
- *
- * We do this global tlb flush inside the cpa_lock, so that we
- * don't allow any other cpu, with stale tlb entries change the
- * page attribute in parallel, that also falls into the
- * just split large page entry.
- */
- flush_tlb_all();
+ if (!err)
goto repeat;
- }
return err;
}
next prev parent reply other threads:[~2018-09-18 8:19 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-17 14:29 [patch V3 00/11] x86/mm/cpa: Improve large page preservation handling Thomas Gleixner
2018-09-17 14:29 ` [patch V3 01/11] x86/mm/init32: Mark text and rodata RO in one go Thomas Gleixner
2018-09-21 16:15 ` Dave Hansen
2018-09-27 18:45 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 02/11] x86/mm/cpa: Split, rename and clean up try_preserve_large_page() Thomas Gleixner
2018-09-18 7:03 ` Peter Zijlstra
2018-09-18 8:19 ` Peter Zijlstra [this message]
2018-09-18 12:14 ` Peter Zijlstra
2018-09-18 22:34 ` Thomas Gleixner
2018-09-21 16:22 ` Dave Hansen
2018-09-27 18:46 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 03/11] x86/mm/cpa: Rework static_protections() Thomas Gleixner
2018-09-21 16:33 ` Dave Hansen
2018-09-27 18:46 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 04/11] x86/mm/cpa: Allow range check for static protections Thomas Gleixner
2018-09-21 16:36 ` Dave Hansen
2018-09-27 18:47 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 05/11] x86/mm/cpa: Add debug mechanism Thomas Gleixner
2018-09-21 16:40 ` Dave Hansen
2018-09-22 10:33 ` Peter Zijlstra
2018-09-27 18:48 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 06/11] x86/mm/cpa: Add large page preservation statistics Thomas Gleixner
2018-09-21 19:59 ` Dave Hansen
2018-09-27 18:48 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 07/11] x86/mm/cpa: Avoid static protection checks on unmap Thomas Gleixner
2018-09-21 20:01 ` Dave Hansen
2018-09-27 18:49 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 08/11] x86/mm/cpa: Add sanity check for existing mappings Thomas Gleixner
2018-09-18 7:14 ` Peter Zijlstra
2018-09-21 20:07 ` Dave Hansen
2018-09-27 18:49 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 09/11] x86/mm/cpa: Optimize same protection check Thomas Gleixner
2018-09-21 20:12 ` Dave Hansen
2018-09-27 18:07 ` Thomas Gleixner
2018-09-27 18:50 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 10/11] x86/mm/cpa: Do the range check early Thomas Gleixner
2018-09-21 20:26 ` Dave Hansen
2018-09-27 18:50 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
2018-09-17 14:29 ` [patch V3 11/11] x86/mm/cpa: Avoid the 4k pages check completely Thomas Gleixner
2018-09-21 20:32 ` Dave Hansen
2018-09-27 18:51 ` [tip:x86/mm] " tip-bot for Thomas Gleixner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180918081909.GI24106@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=bin.yang@intel.com \
--cc=dave.hansen@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.gross@intel.com \
--cc=tglx@linutronix.de \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).