From: Paul Burton <paul.burton@imgtec.com>
To: Thomas Gleixner <tglx@linutronix.de>, Ralf Baechle <ralf@linux-mips.org>
Cc: dianders@chromium.org, James Hogan <james.hogan@imgtec.com>,
Brian Norris <briannorris@chromium.org>,
Jason Cooper <jason@lakedaemon.net>,
jeffy.chen@rock-chips.com, Marc Zyngier <marc.zyngier@arm.com>,
linux-kernel@vger.kernel.org, linux-mips@linux-mips.org,
tfiga@chromium.org, Paul Burton <paul.burton@imgtec.com>
Subject: [RFC PATCH v1 0/9] Support shared percpu interrupts; clean up MIPS hacks
Date: Thu, 7 Sep 2017 16:25:33 -0700 [thread overview]
Message-ID: <20170907232542.20589-1-paul.burton@imgtec.com> (raw)
Message-ID: <20170907232533.-JAhz_8hpBD0SIp2xVUWVgtaCaqeoKieTcxfP_GvJtM@z> (raw)
In-Reply-To: <1682867.tATABVWsV9@np-p-burton>
This series introduces support for percpu shared interrupts and makes
use of this support to clean up some hacks that have been used to
support such interrupts on MIPS.
- Patch 1 allows users of shared interrupts to opt into IRQ_NOAUTOEN
behaviour & avoid warnings from doing so.
- Patch 2 introduces support for shared percpu_devid interrupts.
- Patch 3 introduces a helper allowing users to detect whether an
interrupt is a percpu_devid interrupt or not, which is useful
during the transition phase where interrupts may be either.
- Patches 4 & 5 removes an ugly custom implementation of shared
interrupts between the MIPS cevt-r4k timer driver & users of
performance counters, in favor of using standard IRQF_SHARED &
multiple handlers.
- Patches 6 & 7 add percpu interrupt support to the MIPS perf &
cevt-r4k timer drivers respectively.
- Patch 8 configures the MIPS cop 0 count/compare, fast debug channel &
performance counter overflow interrupts as percpu_devid when they are
mapped by the irqchip-mips-cpu driver.
- Patch 9 removes a hack from the irqchip-mips-gic driver that was
used to enable & disable an interrupt across all CPUs, which is no
longer necessary with users of those interrupts using the percpu
interrupt APIs correctly. This mirrors patch 8 for systems where we
map the CPU local interrupts via the GIC.
There's a little more work necessary before this could go in - the
MIPS oprofile code needs adjusting to use the percpu interrupt APIs, as
does the fast debug channel driver.
Applies atop next-20170905.
Paul Burton (9):
genirq: Allow shared interrupt users to opt into IRQ_NOAUTOEN
genirq: Support shared per_cpu_devid interrupts
genirq: Introduce irq_is_percpu_devid()
MIPS: Remove perf_irq interrupt sharing fallback
MIPS: Remove perf_irq
MIPS: perf: percpu_devid interrupt support
MIPS: cevt-r4k: percpu_devid interrupt support
irqchip: mips-cpu: Set timer, FDC & perf interrupts percpu_devid
irqchip: mips-gic: Remove gic_all_vpes_local_irq_controller
arch/mips/include/asm/time.h | 1 -
arch/mips/kernel/cevt-r4k.c | 77 ++++++++++++-----------------
arch/mips/kernel/perf_event_mipsxx.c | 71 +++++++++++----------------
arch/mips/kernel/time.c | 9 ----
arch/mips/kernel/traps.c | 2 +-
arch/mips/oprofile/op_impl.h | 2 -
arch/mips/oprofile/op_model_loongson3.c | 39 +++++++--------
arch/mips/oprofile/op_model_mipsxx.c | 10 +---
drivers/irqchip/irq-mips-cpu.c | 9 +++-
drivers/irqchip/irq-mips-gic.c | 69 +++-----------------------
include/linux/interrupt.h | 2 +
include/linux/irqdesc.h | 8 +++
kernel/irq/chip.c | 8 +--
kernel/irq/handle.c | 8 ++-
kernel/irq/manage.c | 86 +++++++++++++++++++++++++--------
kernel/irq/settings.h | 5 ++
16 files changed, 189 insertions(+), 217 deletions(-)
--
2.14.1
next parent reply other threads:[~2017-09-07 23:26 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <1682867.tATABVWsV9@np-p-burton>
2017-09-07 23:25 ` Paul Burton [this message]
2017-09-07 23:25 ` [RFC PATCH v1 0/9] Support shared percpu interrupts; clean up MIPS hacks Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 1/9] genirq: Allow shared interrupt users to opt into IRQ_NOAUTOEN Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 2/9] genirq: Support shared per_cpu_devid interrupts Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-25 21:06 ` Thomas Gleixner
2017-09-26 12:00 ` Thomas Gleixner
2017-10-19 14:08 ` Thomas Gleixner
2017-09-07 23:25 ` [RFC PATCH v1 3/9] genirq: Introduce irq_is_percpu_devid() Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 4/9] MIPS: Remove perf_irq interrupt sharing fallback Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 5/9] MIPS: Remove perf_irq Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 6/9] MIPS: perf: percpu_devid interrupt support Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-10-19 14:12 ` Thomas Gleixner
2017-09-07 23:25 ` [RFC PATCH v1 7/9] MIPS: cevt-r4k: " Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 8/9] irqchip: mips-cpu: Set timer, FDC & perf interrupts percpu_devid Paul Burton
2017-09-07 23:25 ` Paul Burton
2017-09-07 23:25 ` [RFC PATCH v1 9/9] irqchip: mips-gic: Remove gic_all_vpes_local_irq_controller Paul Burton
2017-09-07 23:25 ` Paul Burton
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=20170907232542.20589-1-paul.burton@imgtec.com \
--to=paul.burton@imgtec.com \
--cc=briannorris@chromium.org \
--cc=dianders@chromium.org \
--cc=james.hogan@imgtec.com \
--cc=jason@lakedaemon.net \
--cc=jeffy.chen@rock-chips.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mips@linux-mips.org \
--cc=marc.zyngier@arm.com \
--cc=ralf@linux-mips.org \
--cc=tfiga@chromium.org \
--cc=tglx@linutronix.de \
/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).