From: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com> To: Lee Jones <lee.jones@linaro.org> Cc: linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, will.deacon@arm.com, jonas.aberg@stericsson.com, linux@arm.linux.org.uk, linus.walleij@stericsson.com, arnd@arndb.de Subject: Re: [PATCH 0/2] Boottime: measure bootloader and kernel bootup time Date: Thu, 15 Nov 2012 11:57:52 +0100 [thread overview] Message-ID: <20121115105752.GC4398@game.jcrosoft.org> (raw) In-Reply-To: <20121115103528.GA21682@gmail.com> On 10:35 Thu 15 Nov , Lee Jones wrote: > On Thu, 15 Nov 2012, Jean-Christophe PLAGNIOL-VILLARD wrote: > > > On 10:04 Thu 15 Nov , Lee Jones wrote: > > > This patchset supplies a new tool which measures boottime inclusive > > > of bootloader involvement. It seems to be accurate and adds no > > > latency to the booting of the system. > > > > > > In the first round of testing we booted the kernel 3 times each with > > > boottime enabled and disabled. Actually, when it was disabled, the > > > code was completely removed from the kernel source and the kernel was > > > rebuilt. The difference between present & enabled and completely > > > removed was very little indeed. In fact, averaging out the 3 runs of > > > each, when boottime was enabled the system booted 4ms _faster_, go > > > figure! > > > > > > Enabled Disabled > > > 1st run 2.983093 2.985168 > > > 2nd run 2.973266 2.971801 > > > 3rd run 2.975402 2.987304 > > > Average 2.977254 2.981424 -0.004171 (-4ms) > > > > > > > > > For the next round of testing, we delayed the bootloader for varying > > > amounts of time, then started to remove key components from kernel > > > functionality to see if the boottime system would pick-up on them. > > > > > > /* Hang around in the bootloader for a while. */ > > > $ cat /sys/kernel/debug/boottime/summary > > > kernel: 42864 msecs > > > total: 42864 msecs > > > kernel: cpu0 system: 56% idle: 43% iowait: 0% irq: 0% > > > cpu1 system: 4% idle: 95% iowait: 0% irq: 0% > > > > > > /* Hang around in the bootloader for a smaller amount of time. */ > > > kernel: 12351 msecs > > > total: 12351 msecs > > > kernel: cpu0 system: 56% idle: 43% iowait: 0% irq: 0% > > > cpu1 system: 2% idle: 97% iowait: 0% irq: 0% > > > > > > /* Instant boot. (1st run) */ > > > kernel: 7833 msecs > > > total: 7833 msecs > > > kernel: cpu0 system: 55% idle: 44% iowait: 0% irq: 0% > > > cpu1 system: 3% idle: 96% iowait: 0% irq: 0% > > > > > > /* Instant boot. (2nd run) */ > > > kernel: 7817 msecs > > > total: 7817 msecs > > > kernel: cpu0 system: 53% idle: 46% iowait: 0% irq: 0% > > > cpu1 system: 8% idle: 91% iowait: 0% irq: 0% > > > > > > /* Instant boot. (3rd run) */ > > > kernel: 7747 msecs > > > total: 7747 msecs > > > kernel: cpu0 system: 56% idle: 42% iowait: 0% irq: 0% > > > cpu1 system: 2% idle: 97% iowait: 0% irq: 0% > > 7/8s? on a u9500 wow it's really slow > > No, this is a u8500. > > > I do 2/3s here > > Really? You installed this quick. > > Are you sure you did it correctly? yes we spend less than 2s (~ 1.5s) in the bootloader I measure it via scope > > Don't forget, this times from power-on to userspace. yes to the shell prompt and gexbox 13s IIRC application Best Regards, J. > > -- > Lee Jones > Linaro ST-Ericsson Landing Team Lead > Linaro.org │ Open source software for ARM SoCs > Follow Linaro: Facebook | Twitter | Blog
WARNING: multiple messages have this Message-ID (diff)
From: plagnioj@jcrosoft.com (Jean-Christophe PLAGNIOL-VILLARD) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH 0/2] Boottime: measure bootloader and kernel bootup time Date: Thu, 15 Nov 2012 11:57:52 +0100 [thread overview] Message-ID: <20121115105752.GC4398@game.jcrosoft.org> (raw) In-Reply-To: <20121115103528.GA21682@gmail.com> On 10:35 Thu 15 Nov , Lee Jones wrote: > On Thu, 15 Nov 2012, Jean-Christophe PLAGNIOL-VILLARD wrote: > > > On 10:04 Thu 15 Nov , Lee Jones wrote: > > > This patchset supplies a new tool which measures boottime inclusive > > > of bootloader involvement. It seems to be accurate and adds no > > > latency to the booting of the system. > > > > > > In the first round of testing we booted the kernel 3 times each with > > > boottime enabled and disabled. Actually, when it was disabled, the > > > code was completely removed from the kernel source and the kernel was > > > rebuilt. The difference between present & enabled and completely > > > removed was very little indeed. In fact, averaging out the 3 runs of > > > each, when boottime was enabled the system booted 4ms _faster_, go > > > figure! > > > > > > Enabled Disabled > > > 1st run 2.983093 2.985168 > > > 2nd run 2.973266 2.971801 > > > 3rd run 2.975402 2.987304 > > > Average 2.977254 2.981424 -0.004171 (-4ms) > > > > > > > > > For the next round of testing, we delayed the bootloader for varying > > > amounts of time, then started to remove key components from kernel > > > functionality to see if the boottime system would pick-up on them. > > > > > > /* Hang around in the bootloader for a while. */ > > > $ cat /sys/kernel/debug/boottime/summary > > > kernel: 42864 msecs > > > total: 42864 msecs > > > kernel: cpu0 system: 56% idle: 43% iowait: 0% irq: 0% > > > cpu1 system: 4% idle: 95% iowait: 0% irq: 0% > > > > > > /* Hang around in the bootloader for a smaller amount of time. */ > > > kernel: 12351 msecs > > > total: 12351 msecs > > > kernel: cpu0 system: 56% idle: 43% iowait: 0% irq: 0% > > > cpu1 system: 2% idle: 97% iowait: 0% irq: 0% > > > > > > /* Instant boot. (1st run) */ > > > kernel: 7833 msecs > > > total: 7833 msecs > > > kernel: cpu0 system: 55% idle: 44% iowait: 0% irq: 0% > > > cpu1 system: 3% idle: 96% iowait: 0% irq: 0% > > > > > > /* Instant boot. (2nd run) */ > > > kernel: 7817 msecs > > > total: 7817 msecs > > > kernel: cpu0 system: 53% idle: 46% iowait: 0% irq: 0% > > > cpu1 system: 8% idle: 91% iowait: 0% irq: 0% > > > > > > /* Instant boot. (3rd run) */ > > > kernel: 7747 msecs > > > total: 7747 msecs > > > kernel: cpu0 system: 56% idle: 42% iowait: 0% irq: 0% > > > cpu1 system: 2% idle: 97% iowait: 0% irq: 0% > > 7/8s? on a u9500 wow it's really slow > > No, this is a u8500. > > > I do 2/3s here > > Really? You installed this quick. > > Are you sure you did it correctly? yes we spend less than 2s (~ 1.5s) in the bootloader I measure it via scope > > Don't forget, this times from power-on to userspace. yes to the shell prompt and gexbox 13s IIRC application Best Regards, J. > > -- > Lee Jones > Linaro ST-Ericsson Landing Team Lead > Linaro.org ? Open source software for ARM SoCs > Follow Linaro: Facebook | Twitter | Blog
next prev parent reply other threads:[~2012-11-15 10:59 UTC|newest] Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top 2012-11-15 10:04 [PATCH 0/2] Boottime: measure bootloader and kernel bootup time Lee Jones 2012-11-15 10:04 ` Lee Jones 2012-11-15 10:04 ` [PATCH 1/2] Boottime: A tool for automatic measurement of kernel/bootloader boot time Lee Jones 2012-11-15 10:04 ` Lee Jones 2012-11-15 10:16 ` Kyungmin Park 2012-11-15 10:16 ` Kyungmin Park 2012-11-15 11:25 ` Lee Jones 2012-11-15 11:25 ` Lee Jones 2012-11-15 12:25 ` Arnd Bergmann 2012-11-15 12:25 ` Arnd Bergmann 2012-11-15 17:31 ` Linus Walleij 2012-11-15 17:31 ` Linus Walleij 2012-11-20 16:16 ` Lee Jones 2012-11-20 16:16 ` Lee Jones 2012-11-29 20:35 ` Simon Glass 2012-11-29 20:35 ` Simon Glass 2012-11-30 8:21 ` Lee Jones 2012-11-30 8:21 ` Lee Jones 2012-11-15 13:52 ` Andrew Murray 2012-11-15 13:52 ` Andrew Murray 2012-11-15 14:37 ` Lee Jones 2012-11-15 14:37 ` Lee Jones 2012-11-15 10:04 ` [PATCH 2/2] clocksource: clksrc-dbx500-prcmu: Add boottime support Lee Jones 2012-11-15 10:04 ` Lee Jones 2012-11-15 10:15 ` [PATCH 0/2] Boottime: measure bootloader and kernel bootup time Jean-Christophe PLAGNIOL-VILLARD 2012-11-15 10:15 ` Jean-Christophe PLAGNIOL-VILLARD 2012-11-15 10:35 ` Lee Jones 2012-11-15 10:35 ` Lee Jones 2012-11-15 10:57 ` Jean-Christophe PLAGNIOL-VILLARD [this message] 2012-11-15 10:57 ` Jean-Christophe PLAGNIOL-VILLARD
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=20121115105752.GC4398@game.jcrosoft.org \ --to=plagnioj@jcrosoft.com \ --cc=arnd@arndb.de \ --cc=jonas.aberg@stericsson.com \ --cc=lee.jones@linaro.org \ --cc=linus.walleij@stericsson.com \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux@arm.linux.org.uk \ --cc=will.deacon@arm.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.