From: Cyrill Gorcunov <gorcunov@gmail.com>
To: "Maciej W. Rozycki" <macro@linux-mips.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>,
Yinghai Lu <yhlu.kernel@gmail.com>, Ingo Molnar <mingo@elte.hu>,
Thomas Gleixner <tglx@linutronix.de>,
"H. Peter Anvin" <hpa@zytor.com>,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] x86: let 32bit use apic_ops too
Date: Mon, 14 Jul 2008 22:09:08 +0400 [thread overview]
Message-ID: <20080714180908.GB6986@asus> (raw)
In-Reply-To: <Pine.LNX.4.55.0807141805500.16334@cliff.in.clinika.pl>
[Maciej W. Rozycki - Mon, Jul 14, 2008 at 06:20:26PM +0100]
| On Mon, 14 Jul 2008, Cyrill Gorcunov wrote:
|
| > Maciej, but if we eliminate LOCK# by using simple MOV there will not
| > be guarantee for atomicity. Am I wrong?
|
| You are right, but we do not care about atomicity. We only care about
| interrupts. This is because the local APIC is private to its associated
| CPU and inaccessible from the outside, at least for writes (mind the
| Remote Read command), so as long as the local CPU does not issue
| consecutive write cycles, there is no problem with another CPU getting in
| the way. Which means any RMW instruction would suffice here, with the R
| part of the cycle separating any possible preceding write from one
| immediately following, but unfortunately the only one we can use is the
| XCHG and it has always implied the LOCK#, since the 8086, which at that
| point was considered a microoptimization (the LOCK# was cheap and an extra
| memory byte, otherwise needed for the LOCK prefix, expensive back then).
| So atomicity is an unfortunate side effect rather than a part of the
| design here.
|
| Now if we know the APIC does not suffer from the double-write erratum,
| then we can use a straight MOV as consecutive writes are not a concern
| anymore.
|
| Maciej
|
Maciej, check me please (it's a bit shame but I don't understand the problem
that deep) - we have only two errata here 3AP and 11AP. 3AP says - "Writes to
error register clears register" so we don't care about locking there since
our mostly task is to read error number or clear it (well we're recommened
to write before read - but that is different and not related to the hw
error).
The second problem - 11AP says the following: "Back to back assertions of
HOLD or BOFF# may cause lost APIC write cycle". For this case we use LOCK#
since - HOLD is not recognized during LOCK cycles (as Intel docs says).
Did I miss something? Or maybe it's completely out-of-topic? :)
- Cyrill -
next prev parent reply other threads:[~2008-07-14 18:09 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-07-08 8:41 [PATCH] x86: introduce page_size_mask for 64bit Yinghai Lu
2008-07-08 8:43 ` [PATCH] x86: not overmap than end in init_memory_mapping - 64bit Yinghai Lu
2008-07-09 7:38 ` Ingo Molnar
2008-07-09 8:34 ` Ingo Molnar
2008-07-09 8:37 ` Yinghai Lu
2008-07-09 8:46 ` Ingo Molnar
2008-07-09 8:58 ` Yinghai Lu
2008-07-09 10:01 ` Yinghai Lu
2008-07-09 10:30 ` Ingo Molnar
2008-07-09 8:45 ` Ingo Molnar
2008-07-10 3:15 ` [PATCh] x86: overmapped fix when 4K pages on tail " Yinghai Lu
2008-07-10 3:16 ` [PATCH] x86: merge __acpi_map_table Yinghai Lu
2008-07-10 3:17 ` [PATCH] x86: make e820_end return end_of_ram again for 64bit Yinghai Lu
2008-07-10 7:00 ` Ingo Molnar
2008-07-10 11:17 ` [PATCH] x86: e820 remove the range instead of update it to reserved Yinghai Lu
2008-07-11 8:20 ` Ingo Molnar
2008-07-11 3:36 ` [PATCH] x86: save slit Yinghai Lu
2008-07-11 8:22 ` Ingo Molnar
2008-07-11 3:38 ` [PATCH] x86: introduce max_low_pfn_mapped for 64bit Yinghai Lu
2008-07-11 8:26 ` Ingo Molnar
2008-07-11 8:39 ` Yinghai Lu
2008-07-11 8:51 ` Ingo Molnar
2008-07-12 1:41 ` [PATCH] x86: let 32bit use apic_ops too Yinghai Lu
2008-07-12 1:43 ` [PATCH] x86: mach_apicdef.h need to include before smp.h Yinghai Lu
2008-07-12 1:44 ` [PATCH] x86: make read_apic_id return final apicid Yinghai Lu
2008-07-12 8:01 ` [PATCH] x86: make 64bit have get_apic_id Yinghai Lu
2008-07-13 6:28 ` Ingo Molnar
2008-07-13 6:59 ` Ingo Molnar
2008-07-13 7:05 ` Yinghai Lu
2008-07-13 9:23 ` Ingo Molnar
2008-07-13 9:28 ` Ingo Molnar
2008-07-13 16:15 ` Suresh Siddha
2008-07-13 1:19 ` [PATCH] x86: make read_apic_id return final apicid Suresh Siddha
2008-07-13 1:08 ` [PATCH] x86: let 32bit use apic_ops too Suresh Siddha
2008-07-13 2:04 ` Yinghai Lu
2008-07-13 16:28 ` Suresh Siddha
2008-07-13 16:51 ` Maciej W. Rozycki
2008-07-13 17:16 ` Cyrill Gorcunov
2008-07-13 23:46 ` Maciej W. Rozycki
2008-07-14 16:48 ` Cyrill Gorcunov
2008-07-14 17:20 ` Maciej W. Rozycki
2008-07-14 18:09 ` Cyrill Gorcunov [this message]
2008-07-14 18:24 ` Maciej W. Rozycki
2008-07-14 18:32 ` Cyrill Gorcunov
2008-07-13 1:43 ` Maciej W. Rozycki
2008-07-13 1:45 ` Yinghai Lu
2008-07-13 1:54 ` Maciej W. Rozycki
2008-07-13 16:43 ` Suresh Siddha
2008-07-13 17:05 ` Maciej W. Rozycki
2008-07-14 5:19 ` [PATCH] x86: let 32bit use apic_ops too - fix Yinghai Lu
2008-07-14 7:12 ` Ingo Molnar
2008-07-14 16:49 ` Suresh Siddha
2008-07-14 17:00 ` Yinghai Lu
2008-07-14 18:03 ` Suresh Siddha
2008-07-18 17:06 ` Ingo Molnar
2008-07-15 17:33 ` Suresh Siddha
2008-07-15 18:10 ` Yinghai Lu
2008-07-15 18:27 ` Suresh Siddha
2008-07-18 17:07 ` Ingo Molnar
2008-07-12 21:30 ` [PATCH] x86: max_low_pfn_mapped fix #1 Yinghai Lu
2008-07-13 9:45 ` Ingo Molnar
2008-07-12 21:31 ` [PATCH] x86: max_low_pfn_mapped fix #2 Yinghai Lu
2008-07-12 21:32 ` [PATCH] x86: max_low_pfn_mapped fix #3 Yinghai Lu
2008-07-13 21:29 ` [PATCH] x86: max_low_pfn_mapped fix #4 Yinghai Lu
2008-07-13 21:30 ` [PATCH] x86: get x86_phys_bits early Yinghai Lu
2008-07-13 21:32 ` [PATCH] x86: make 64bit hpet_set_mapping to use ioremap too Yinghai Lu
2008-07-13 21:50 ` [PATCH] x86: make 64bit hpet_set_mapping to use ioremap too v2 Yinghai Lu
2008-07-10 6:54 ` [PATCH] x86: merge __acpi_map_table Ingo Molnar
2008-07-10 6:53 ` [PATCh] x86: overmapped fix when 4K pages on tail - 64bit Ingo Molnar
2008-07-10 6:57 ` Yinghai Lu
2008-07-10 7:20 ` Ingo Molnar
2008-07-10 7:32 ` Yinghai Lu
2008-07-10 14:16 ` Arjan van de Ven
2008-07-13 14:57 ` Andi Kleen
2008-07-13 15:33 ` Arjan van de Ven
2008-07-13 18:25 ` Andi Kleen
2008-07-13 18:17 ` Yinghai Lu
2008-07-13 18:48 ` Andi Kleen
2008-07-13 19:00 ` Yinghai Lu
2008-07-13 20:32 ` Ingo Molnar
2008-07-13 20:51 ` Andi Kleen
2008-07-14 0:04 ` H. Peter Anvin
2008-07-14 6:39 ` Andi Kleen
2008-07-09 7:38 ` [PATCH] x86: introduce page_size_mask for 64bit Ingo Molnar
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=20080714180908.GB6986@asus \
--to=gorcunov@gmail.com \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=macro@linux-mips.org \
--cc=mingo@elte.hu \
--cc=suresh.b.siddha@intel.com \
--cc=tglx@linutronix.de \
--cc=yhlu.kernel@gmail.com \
/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