Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
On 05/16/2016 03:42 PM, Aaron Z wrote: On Mon, May 16, 2016 at 5:21 PM, Ben Greearwrote: Maybe get an IRC bot to dump irc logs to the mailing list every few hours? Might be a bit spammy though IMO, trying to follow IRC when its not realtime (ie: reading a log) makes it somewhat difficult to follow threads of conversations (which email does quite well) and it can be confusing unless there is some way to put a "this post goes in this thread" for each post (and at that point, you might as well use email). Yes, I like email too. But IRC has its uses, and maybe better to have it spammed to the mailing list that just plain lost/ignored. Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
Maybe get an IRC bot to dump irc logs to the mailing list every few hours? Might be a bit spammy though Ben On 05/16/2016 12:08 PM, Dave Taht wrote: znc and bitlbee are godsends when using irc asynchronously. In particular I feed all irc convos into bitlbee and then into erc on emacs (so I am not hurt when my connection goes away). On Mon, May 16, 2016 at 5:00 AM, David Woodhousewrote: On Mon, 2016-05-16 at 12:46 +0100, Bruno Randolf wrote: On 15/05/16 05:53, Daniel Dickinson wrote: I'd really appreciate if we could actually use the mailing list for the main communications venue rather than shutting out people not in the European timezones, which is what happens if IRC is the main way to participate in the community. I have been told that to really be part of the OpenWrt action I should have been on IRC; but I'm not in European timezone so that is not actually a useful suggestion, since most of the most people most relevant to decision-making are in Europe, and I would hope LEDE has a more *inter-continental focus than OpenWrt had. Agreed! I don't have time to hang out in IRC channels, usually. Or maybe a different working style... (it distracts me too much). Anyhow I think all important things and all things which don't need real-time interaction should go thru the mailing list, please. FWIW, IRC doesn't have to have real-time behaviour, and doesn't have to be constantly distracting. It's perfectly possible to have IRC conversations with people who are never awake at the same time as you. You say something, they respond when they wake up... and you respond when *you* wake up. Just like email, in fact. IRC gives you the *option* of real-time conversation, if you happen to be awake at the same time. And you don't have to pay attention to the channel all the time; an IRC client should highlight if your nick is mentioned because someone is talking you to specifically. IRC is great for conversations which are more ephemeral, and don't need to be archived for posterity. -- dwmw2 ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev -- Ben Greear Candela Technologies Inc http://www.candelatech.com ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
On 16-05-16 05:18 PM, Daniel Curran-Dickinson wrote: > The objective is so that you don't have isolated pools of TZ's where one > TZ has little insight into what the other TZ is doing and has no *good* > mechanisms for communicating across timezones. > > Reading IRC chat logs is an exercise is pain for many, so that's not > really the answer for cross-timezone communication. > > IRC, for many of us, is something that is only useful when it's realtime. > It is also difficult to follow *threads* of conversation that are severely time-delayed. In addition, in order to not miss messages you really need to set up a bouncer that is always on and keeps a unread messages even across an unexpected reboot or logging out of your console/gui session. Email is far superior for time-delayed threads of communication. If you're having problems with managing your email, it's your email strategy that needs to change, not the use of email. Regards, Dan iel ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
On 2016-05-16 21:14, Bruno Randolf wrote: > On 16/05/16 20:08, Dave Taht wrote: >> znc and bitlbee are godsends when using irc asynchronously. >> >> In particular I feed all irc convos into bitlbee and then into erc on >> emacs (so I am not hurt when my connection goes away). > > Well, thanks for the hints, but please just accept the fact that IRC is > not part of everyones working mode. One of the ideas that was tossed around was using LiquidFeedback to replace the IRC meeting voting stuff. We're still looking into it... - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] Concern over kernel: backport patches improving fq_codel drop behavior
On 16/05/16 11:46, Felix Fietkau wrote: On 2016-05-13 11:57, Kevin Darbyshire-Bryant wrote: Hi All, I'm *very* much *for* the fq_codel batch drop functionality and that ideally should be in a lede 4.4 patch, without the API change. I attach a patch that I think more suitable...or would form the basis of something more suitable at least. Apologies for the somewhat rush job on this I'm 10 minutes from rushing out of the door for a week holiday :-) What do you think about moving cake to the LEDE source tree instead? - Felix That's certainly a plan! I'll do a PR next week :-) Kevin ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] ar71xx: performance decrease with kernel 4.4 (might be due to the qdisc/codel changes)
As a side note, you can do direct comparisons between data sets in flent, really easily. Load up all the data files in it, select something less noisy as a plot type (cdf, totals, or box totals), then select Data->Add other open data files. I am *ALWAYS* a fan of looking at the default, detailed graph (all or all_scaled)first, to see if there was any weirdnesses that would be missed by the other plot types. As an example of weirdnesses you can only detect by looking at a string of test's long term behavior, see the current thread and pictures on: https://plus.google.com/u/0/107942175615993706558/posts/WA915Pt4SRN One of the things I was concerned about going from 4.1 to 4.4 is that there was a lot of x86_64 based optimization of the routing cache... and I was certainly concerned about adding more variables to the fq_codel algorithm, also. I saw pretty bad register usage in cake on arm when I last looked, I've kind of always wanted people to try -O2 instead of -Os. On Mon, May 16, 2016 at 1:48 AM, Hannu Nymanwrote: > I already said to Felix yesterday that I felt that with kernel 4.4 my ar71xx > WNDR3800 seemed somehow more sluggish. Now I tested the matter with "flent". > > And sadly, with kernel 4.4 my router's performance decreases significantly. > With kernel 4.1 the router achieves about 20% higher download throughput > than with the 4.4 build :-( > > I used "flent" to measure wired connection throughput with > - two LEDE builds: r241 with kernel 4.1 and r253 with kernel 4.4 > - two separate speed limits for SQM simple fq_codel QoS: 85000/1 kb/s > that should leave some CPU power free in the router, and 11/15000 that > should fully utilise the router's CPU. > - otherwise identical settings, all measurements made inside 30 minutes so > no changes in traffic conditions > > The achieved speeds were: > > r241 kernel 4.1 - 85/10: 79 Mb/s down, 8.1 Mb/s up > r253 kernel 4.4 - 85/10: 67 Mb/s down, 8.5 Mb/s up > > r241 kernel 4.1 - 110/15: 85 down, 10.5 up > r253 kernel 4.4 - 110/15: 70 down, 10.8 up > > (ping latency stays at 16-17 ms with all four tries) > > With both SQM speed limits, the kernel 4.1 build performs significantly > better. > > This performance decrease might be due to the kernel version bump to 4.4 or > the qdisc/codel changes made to the 4.4 patches a few days earlier. > > This chart sums it up: > https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0=Kernel41_44_comparison.png > > Full flent data (summary pics & rrul data files) for all four tries is > available in: > https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0 > > > > ___ > Lede-dev mailing list > Lede-dev@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/lede-dev -- Dave Täht Let's go make home routers and wifi faster! With better software! http://blog.cerowrt.org ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH] include/kernel: Do not strip kernel's Elf
On Mon, May 16, 2016 at 05:42:43PM +0300, Alexey Brodkin wrote: > If an image gets built as an Elf there's a chance > it will be used by developers for debugging purposes. > In that case it's very helpful to keep debugging info > in that image. > > I would think that most OWRT-powered devices in production > will use some form of binary image for booting so Elf > flavours could be left a bit bulkier with more debug info > inside. Well, some of the tiniest AR7 boxes out there come with a bootloader expecting ELF binaries... I didn't have a close look, just assuming that they might be affected. May Florian knows more...? Cheers Daniel > > Signed-off-by: Alexey Brodkin> --- > include/kernel-defaults.mk | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/include/kernel-defaults.mk b/include/kernel-defaults.mk > index 406fd46..0166dc5 100644 > --- a/include/kernel-defaults.mk > +++ b/include/kernel-defaults.mk > @@ -142,7 +142,7 @@ endif > define Kernel/CopyImage > cmp -s $(LINUX_DIR)/vmlinux $(KERNEL_BUILD_DIR)/vmlinux$(1).debug || { \ > $(KERNEL_CROSS)objcopy -O binary $(OBJCOPY_STRIP) -S > $(LINUX_DIR)/vmlinux $(LINUX_KERNEL)$(1); \ > - $(KERNEL_CROSS)objcopy $(OBJCOPY_STRIP) -S $(LINUX_DIR)/vmlinux > $(KERNEL_BUILD_DIR)/vmlinux$(1).elf; \ > + $(KERNEL_CROSS)objcopy $(OBJCOPY_STRIP) $(LINUX_DIR)/vmlinux > $(KERNEL_BUILD_DIR)/vmlinux$(1).elf; \ > $(CP) $(LINUX_DIR)/vmlinux > $(KERNEL_BUILD_DIR)/vmlinux$(1).debug; \ > $(foreach k, \ > $(if $(KERNEL_IMAGES),$(KERNEL_IMAGES),$(filter-out > dtbs,$(KERNELNAME))), \ > -- > 2.5.5 > > > ___ > Lede-dev mailing list > Lede-dev@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCH] include/kernel: Do not strip kernel's Elf
If an image gets built as an Elf there's a chance it will be used by developers for debugging purposes. In that case it's very helpful to keep debugging info in that image. I would think that most OWRT-powered devices in production will use some form of binary image for booting so Elf flavours could be left a bit bulkier with more debug info inside. Signed-off-by: Alexey Brodkin--- include/kernel-defaults.mk | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/kernel-defaults.mk b/include/kernel-defaults.mk index 406fd46..0166dc5 100644 --- a/include/kernel-defaults.mk +++ b/include/kernel-defaults.mk @@ -142,7 +142,7 @@ endif define Kernel/CopyImage cmp -s $(LINUX_DIR)/vmlinux $(KERNEL_BUILD_DIR)/vmlinux$(1).debug || { \ $(KERNEL_CROSS)objcopy -O binary $(OBJCOPY_STRIP) -S $(LINUX_DIR)/vmlinux $(LINUX_KERNEL)$(1); \ - $(KERNEL_CROSS)objcopy $(OBJCOPY_STRIP) -S $(LINUX_DIR)/vmlinux $(KERNEL_BUILD_DIR)/vmlinux$(1).elf; \ + $(KERNEL_CROSS)objcopy $(OBJCOPY_STRIP) $(LINUX_DIR)/vmlinux $(KERNEL_BUILD_DIR)/vmlinux$(1).elf; \ $(CP) $(LINUX_DIR)/vmlinux $(KERNEL_BUILD_DIR)/vmlinux$(1).debug; \ $(foreach k, \ $(if $(KERNEL_IMAGES),$(KERNEL_IMAGES),$(filter-out dtbs,$(KERNELNAME))), \ -- 2.5.5 ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] [PATCHv2 1/1] [brcm63xx] Display the correct detected CPU ID
The following changes since commit f8abb68e3a1b78ff5a8d616994daa9721fe8fc83: tools/cmake: bump to 3.5.2 (2016-05-13 17:03:54 +0200) are available in the git repository at: https://github.com/Xotic750/mirror-lede.git fix_cpu_id for you to fetch changes up to 3dc3f588e333466374a688607414eb16c0bb6520: Fix the logged CPU variant (2016-05-14 00:35:38 +0200) Xotic750 (1): Fix the logged CPU variant target/linux/brcm63xx/patches-4.1/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch | 9 + target/linux/brcm63xx/patches-4.4/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch | 9 + 2 files changed, 18 insertions(+) diff --git a/target/linux/brcm63xx/patches-4.1/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch b/target/linux/brcm63xx/patches-4.1/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch index 661abf6..23491aa 100644 --- a/target/linux/brcm63xx/patches-4.1/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch +++ b/target/linux/brcm63xx/patches-4.1/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch @@ -41,6 +41,15 @@ Subject: [PATCH 40/53] MIPS: BCM63XX: add a new cpu variant helper bcm63xx_cpu_rev = (tmp & REV_REVID_MASK) >> REV_REVID_SHIFT; switch (bcm63xx_cpu_id) { +@@ -377,7 +387,7 @@ void __init bcm63xx_cpu_init(void) + bcm63xx_memory_size = detect_memory_size(); + + printk(KERN_INFO "Detected Broadcom 0x%04x CPU revision %02x\n", +- bcm63xx_cpu_id, bcm63xx_cpu_rev); ++ bcm63xx_cpu_variant, bcm63xx_cpu_rev); + printk(KERN_INFO "CPU frequency is %u MHz\n", + bcm63xx_cpu_freq / 100); + printk(KERN_INFO "%uMB of RAM installed\n", --- a/arch/mips/include/asm/mach-bcm63xx/bcm63xx_cpu.h +++ b/arch/mips/include/asm/mach-bcm63xx/bcm63xx_cpu.h @@ -19,6 +19,7 @@ diff --git a/target/linux/brcm63xx/patches-4.4/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch b/target/linux/brcm63xx/patches-4.4/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch index 661abf6..330a9e6 100644 --- a/target/linux/brcm63xx/patches-4.4/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch +++ b/target/linux/brcm63xx/patches-4.4/330-MIPS-BCM63XX-add-a-new-cpu-variant-helper.patch @@ -41,6 +41,15 @@ Subject: [PATCH 40/53] MIPS: BCM63XX: add a new cpu variant helper bcm63xx_cpu_rev = (tmp & REV_REVID_MASK) >> REV_REVID_SHIFT; switch (bcm63xx_cpu_id) { +@@ -377,7 +387,7 @@ void __init bcm63xx_cpu_init(void) + bcm63xx_memory_size = detect_memory_size(); + + pr_info("Detected Broadcom 0x%04x CPU revision %02x\n", +- bcm63xx_cpu_id, bcm63xx_cpu_rev); ++ bcm63xx_cpu_variant, bcm63xx_cpu_rev); + pr_info("CPU frequency is %u MHz\n", + bcm63xx_cpu_freq / 100); + pr_info("%uMB of RAM installed\n", --- a/arch/mips/include/asm/mach-bcm63xx/bcm63xx_cpu.h +++ b/arch/mips/include/asm/mach-bcm63xx/bcm63xx_cpu.h @@ -19,6 +19,7 @@ ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
On Mon, 2016-05-16 at 12:46 +0100, Bruno Randolf wrote: > On 15/05/16 05:53, Daniel Dickinson wrote: > > > > I'd really appreciate if we could actually use the mailing list for the > > main communications venue rather than shutting out people not in the > > European timezones, which is what happens if IRC is the main way to > > participate in the community. > > > > I have been told that to really be part of the OpenWrt action I should > > have been on IRC; but I'm not in European timezone so that is not > > actually a useful suggestion, since most of the most people most > > relevant to decision-making are in Europe, and I would hope LEDE has a > > more *inter-continental focus than OpenWrt had. > Agreed! I don't have time to hang out in IRC channels, usually. Or maybe > a different working style... (it distracts me too much). Anyhow I think > all important things and all things which don't need real-time > interaction should go thru the mailing list, please. FWIW, IRC doesn't have to have real-time behaviour, and doesn't have to be constantly distracting. It's perfectly possible to have IRC conversations with people who are never awake at the same time as you. You say something, they respond when they wake up... and you respond when *you* wake up. Just like email, in fact. IRC gives you the *option* of real-time conversation, if you happen to be awake at the same time. And you don't have to pay attention to the channel all the time; an IRC client should highlight if your nick is mentioned because someone is talking you to specifically. IRC is great for conversations which are more ephemeral, and don't need to be archived for posterity. -- dwmw2 smime.p7s Description: S/MIME cryptographic signature ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] A request not making IRC necessary to be part of the action
On 15/05/16 05:53, Daniel Dickinson wrote: > I'd really appreciate if we could actually use the mailing list for the > main communications venue rather than shutting out people not in the > European timezones, which is what happens if IRC is the main way to > participate in the community. > > I have been told that to really be part of the OpenWrt action I should > have been on IRC; but I'm not in European timezone so that is not > actually a useful suggestion, since most of the most people most > relevant to decision-making are in Europe, and I would hope LEDE has a > more *inter-continental focus than OpenWrt had. Agreed! I don't have time to hang out in IRC channels, usually. Or maybe a different working style... (it distracts me too much). Anyhow I think all important things and all things which don't need real-time interaction should go thru the mailing list, please. bruno ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] possible PPPoE fix
On 2016-05-16 12:31, Syrone Wong wrote: > Hi everyone, > > I notice that if I send SIGHUP to pppd, it works the first time, but > if you send the second time, pppd crashes and kernel may crash as > well. I find the musl-libc > fixes from openembedded works well for me, but I don't know the root > cause. I tested the fix on ar71xx/Qihoo-c301 and mvebu/wrt1900ac, the > issue above disappears. Did you reproduce the issue only on mvebu or also on ar71xx? - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
Ok, I tried "git request-pull", I assume that it has been submitted ok, how do I know/track it? [graham@tyr mirror-lede]$ git request-pull f8abb68e3a1b78ff5a8d616994daa9721fe8fc83 https://github.com/Xotic750/mirror-lede.git evg2000 The following changes since commit f8abb68e3a1b78ff5a8d616994daa9721fe8fc83: tools/cmake: bump to 3.5.2 (2016-05-13 17:03:54 +0200) are available in the git repository at: https://github.com/Xotic750/mirror-lede.git evg2000 for you to fetch changes up to c2f97ed220ad8eeb182e80c0fc006ece966679aa: Ran 'make target/linux/refresh V=s' after update to kernel 4.4.10 from 4.4.8 where the initial patch was added. (2016-05-14 00:34:17 +0200) Xotic750 (2): Add initial support for EVG2000 Ran 'make target/linux/refresh V=s' after update to kernel 4.4.10 from 4.4.8 where the initial patch was added. target/linux/brcm63xx/base-files/etc/board.d/01_leds | 7 + target/linux/brcm63xx/base-files/etc/board.d/02_network | 1 + target/linux/brcm63xx/base-files/etc/diag.sh | 3 +++ target/linux/brcm63xx/base-files/etc/uci-defaults/09_fix_crc | 2 +- target/linux/brcm63xx/base-files/lib/brcm63xx.sh | 3 +++ target/linux/brcm63xx/base-files/lib/preinit/05_init_interfaces_brcm63xx | 1 + target/linux/brcm63xx/dts/evg2000.dts | 103 +++ target/linux/brcm63xx/image/Makefile | 2 ++ target/linux/brcm63xx/patches-4.1/805-board_EVG2000.patch | 62 +++ target/linux/brcm63xx/patches-4.4/202-MTD-DEVICES-m25p80-use-parsers-if-provided-in-flash-.patch | 2 +- target/linux/brcm63xx/patches-4.4/203-MTD-DEVICES-m25p80-add-support-for-limiting-reads.patch | 4 +-- target/linux/brcm63xx/patches-4.4/414-MTD-m25p80-allow-passing-pp_data.patch | 2 +- target/linux/brcm63xx/patches-4.4/805-board_EVG2000.patch | 62 +++ target/linux/brcm63xx/profiles/netgear.mk | 10 +++ target/linux/generic/patches-4.4/032-fq_codel-add-batch-ability-to-fq_codel_drop.patch | 4 +-- target/linux/generic/patches-4.4/330-MIPS-kexec-Accept-command-line-parameters-from-users.patch | 10 +++ target/linux/generic/patches-4.4/531-debloat_lzma.patch | 33 --- target/linux/generic/patches-4.4/645-bridge_multicast_to_unicast.patch | 5 ++-- target/linux/generic/patches-4.4/834-ledtrig-libata.patch | 4 +-- 19 files changed, 289 insertions(+), 31 deletions(-) create mode 100644 target/linux/brcm63xx/dts/evg2000.dts create mode 100644 target/linux/brcm63xx/patches-4.1/805-board_EVG2000.patch create mode 100644 target/linux/brcm63xx/patches-4.4/805-board_EVG2000.patch On 16 May 2016 at 12:24, Conor O'Gormanwrote: > On 16/05/16 11:14, Graham Fairweather wrote: >> >> Everything that I read talks about creating PRs when you have forked >> from another github repo, and I have successfully create PRs in this >> situation in the past. I have not seen any information about how to >> create a PR when my repo was forked from a git repo that is hosted >> outside of github. When I go to github and choose create PR, I can >> only create it to branches within my fork, the upstream >> (http://git.lede-project.org/openwrt/source.git) is not listed. Is >> there a github repo that I should have forked from instead? >> > If you have a publicly accessible git repository you can use "git > request-pull" > https://git-scm.com/docs/git-request-pull > > If you are using github you can fork from > https://github.com/lede-project/source > > ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] Concern over kernel: backport patches improving fq_codel drop behavior
On 2016-05-13 11:57, Kevin Darbyshire-Bryant wrote: > Hi All, > > I've just seen > https://github.com/lede-project/source/commit/fad8bdfa40df8636a52d770bbab010086c1976ec > go through and wish to raise a concern: > > The patch introduces a kernel API change 'qdisc_tree_decrease_qlen' > replaces 'qdisc_tree_decrease_qlen' without a corresponding kernel > version number change. This will cause kmod_sched_cake to fail to build > on lede 4.4 kernels. A patch for kmod-sched-cake could be done, however > that would make the package incompatible with openwrt until it also > carries the above kernel patch (it would fail the 'other way around') > > Personally, I think the patch is too invasive until an official backport > is done (with corresponding kernel revision which can be tested for by > kmod-sched-cake) > > I'm *very* much *for* the fq_codel batch drop functionality and that > ideally should be in a lede 4.4 patch, without the API change. I attach > a patch that I think more suitable...or would form the basis of > something more suitable at least. > > Apologies for the somewhat rush job on this I'm 10 minutes from > rushing out of the door for a week holiday :-) What do you think about moving cake to the LEDE source tree instead? - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] possible PPPoE fix
Hi everyone, I notice that if I send SIGHUP to pppd, it works the first time, but if you send the second time, pppd crashes and kernel may crash as well. I find the musl-libc fixes from openembedded works well for me, but I don't know the root cause. I tested the fix on ar71xx/Qihoo-c301 and mvebu/wrt1900ac, the issue above disappears. What I have done: 1. https://github.com/wongsyrone/openwrt-1/commit/25a690bb809012e5194bc87b14b8ade30bd35e57 2.https://github.com/wongsyrone/openwrt-1/commit/4c1e96e9ad2b58d50d7b6f32f7f817a357698838 3.https://github.com/wongsyrone/openwrt-1/commit/84e28c8f303d9b74b1debb9d3d0b2293c8c4a220 Some data I collected: https://github.com/wongsyrone/openwrt-1/issues/87 Related ticket: https://dev.openwrt.org/ticket/22045 It's very appreciated if someone has better solutions against this issue. Best Regards, Syrone Wong ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
Everything that I read talks about creating PRs when you have forked from another github repo, and I have successfully create PRs in this situation in the past. I have not seen any information about how to create a PR when my repo was forked from a git repo that is hosted outside of github. When I go to github and choose create PR, I can only create it to branches within my fork, the upstream (http://git.lede-project.org/openwrt/source.git) is not listed. Is there a github repo that I should have forked from instead? On 16 May 2016 at 11:29, Conor O'Gormanwrote: > On 16/05/16 10:26, Graham Fairweather wrote: >> >> How do I do that? I forked the LEDE repo >> (http://git.lede-project.org/openwrt/source.git) to github, created a >> branch for the work >> (https://github.com/Xotic750/mirror-lede/tree/evg2000), the LEDE >> document (https://www.lede-project.org/development.html) says that PRs >> are accepted, but no detail of how to do it? >> > https://help.github.com/articles/using-pull-requests/ ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
On 16/05/16 11:14, Graham Fairweather wrote: Everything that I read talks about creating PRs when you have forked from another github repo, and I have successfully create PRs in this situation in the past. I have not seen any information about how to create a PR when my repo was forked from a git repo that is hosted outside of github. When I go to github and choose create PR, I can only create it to branches within my fork, the upstream (http://git.lede-project.org/openwrt/source.git) is not listed. Is there a github repo that I should have forked from instead? If you have a publicly accessible git repository you can use "git request-pull" https://git-scm.com/docs/git-request-pull If you are using github you can fork from https://github.com/lede-project/source ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
On 16/05/16 10:26, Graham Fairweather wrote: How do I do that? I forked the LEDE repo (http://git.lede-project.org/openwrt/source.git) to github, created a branch for the work (https://github.com/Xotic750/mirror-lede/tree/evg2000), the LEDE document (https://www.lede-project.org/development.html) says that PRs are accepted, but no detail of how to do it? https://help.github.com/articles/using-pull-requests/ ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
How do I do that? I forked the LEDE repo (http://git.lede-project.org/openwrt/source.git) to github, created a branch for the work (https://github.com/Xotic750/mirror-lede/tree/evg2000), the LEDE document (https://www.lede-project.org/development.html) says that PRs are accepted, but no detail of how to do it? On 16 May 2016 at 11:08, John Crispinwrote: > > > On 16/05/2016 11:06, Graham Fairweather wrote: >> Ok, thanks for the feedback. I tried so hard to make sure that it was >> correct. :( >> > > >> On 16 May 2016 at 10:17, Conor O'Gorman wrote: >>> You are still having line wrap issues with your patches, as Felix previously >>> mentioned. > > > you can use github to send a PR if you like. that way you dont have to > use a mail client. > > ___ > Lede-dev mailing list > Lede-dev@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/lede-dev ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] ar71xx: performance decrease with kernel 4.4 (might be due to the qdisc/codel changes)
On 2016-05-16 10:48, Hannu Nyman wrote: > I already said to Felix yesterday that I felt that with kernel 4.4 my ar71xx > WNDR3800 seemed somehow more sluggish. Now I tested the matter with "flent". > > And sadly, with kernel 4.4 my router's performance decreases significantly. > With kernel 4.1 the router achieves about 20% higher download throughput than > with the 4.4 build :-( > > I used "flent" to measure wired connection throughput with > - two LEDE builds: r241 with kernel 4.1 and r253 with kernel 4.4 > - two separate speed limits for SQM simple fq_codel QoS: 85000/1 kb/s > that should leave some CPU power free in the router, and 11/15000 that > should fully utilise the router's CPU. > - otherwise identical settings, all measurements made inside 30 minutes so no > changes in traffic conditions > > The achieved speeds were: > > r241 kernel 4.1 - 85/10: 79 Mb/s down, 8.1 Mb/s up > r253 kernel 4.4 - 85/10: 67 Mb/s down, 8.5 Mb/s up > > r241 kernel 4.1 - 110/15: 85 down, 10.5 up > r253 kernel 4.4 - 110/15: 70 down, 10.8 up > > (ping latency stays at 16-17 ms with all four tries) > > With both SQM speed limits, the kernel 4.1 build performs significantly > better. > > This performance decrease might be due to the kernel version bump to 4.4 or > the qdisc/codel changes made to the 4.4 patches a few days earlier. > > This chart sums it up: > https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0=Kernel41_44_comparison.png > > Full flent data (summary pics & rrul data files) for all four tries is > available in: > https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0 I've also noticed a throughput decrease on other platforms, I don't have any hard data on that though. I've tried tracking down the source of this regression, but haven't gotten anywhere with that yet. Once we're done with more urgent stuff, I plan to do some more work on optimizing the network stack. - Felix ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
On 16/05/2016 11:06, Graham Fairweather wrote: > Ok, thanks for the feedback. I tried so hard to make sure that it was > correct. :( > > On 16 May 2016 at 10:17, Conor O'Gormanwrote: >> You are still having line wrap issues with your patches, as Felix previously >> mentioned. you can use github to send a PR if you like. that way you dont have to use a mail client. ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
Ok, thanks for the feedback. I tried so hard to make sure that it was correct. :( On 16 May 2016 at 10:17, Conor O'Gormanwrote: > You are still having line wrap issues with your patches, as Felix previously > mentioned. ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] ar71xx: performance decrease with kernel 4.4 (might be due to the qdisc/codel changes)
I already said to Felix yesterday that I felt that with kernel 4.4 my ar71xx WNDR3800 seemed somehow more sluggish. Now I tested the matter with "flent". And sadly, with kernel 4.4 my router's performance decreases significantly. With kernel 4.1 the router achieves about 20% higher download throughput than with the 4.4 build :-( I used "flent" to measure wired connection throughput with - two LEDE builds: r241 with kernel 4.1 and r253 with kernel 4.4 - two separate speed limits for SQM simple fq_codel QoS: 85000/1 kb/s that should leave some CPU power free in the router, and 11/15000 that should fully utilise the router's CPU. - otherwise identical settings, all measurements made inside 30 minutes so no changes in traffic conditions The achieved speeds were: r241 kernel 4.1 - 85/10: 79 Mb/s down, 8.1 Mb/s up r253 kernel 4.4 - 85/10: 67 Mb/s down, 8.5 Mb/s up r241 kernel 4.1 - 110/15: 85 down, 10.5 up r253 kernel 4.4 - 110/15: 70 down, 10.8 up (ping latency stays at 16-17 ms with all four tries) With both SQM speed limits, the kernel 4.1 build performs significantly better. This performance decrease might be due to the kernel version bump to 4.4 or the qdisc/codel changes made to the 4.4 patches a few days earlier. This chart sums it up: https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0=Kernel41_44_comparison.png Full flent data (summary pics & rrul data files) for all four tries is available in: https://www.dropbox.com/sh/89pntkzjxydnn4c/AAC4x9cScJERL9Wfxm4k43kma?dl=0 ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
[LEDE-DEV] mount_root issues
Hi, the recent changes i made to mount_root that caused lots of carnage for lost of people should now be fixed. if you had issues with the root not being mounted or being mounted twice then please update to latest HEAD and try again. John ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH 1/1] [brcm63xx] Add support for the NetGear EVG2000
You are still having line wrap issues with your patches, as Felix previously mentioned. ___ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
Re: [LEDE-DEV] [PATCH] ath25: update kernel from 3.18 to 4.4
I've diffed the bootlogs of 3.18 and 4.4 for my particular image, included below: --- /tmp/ath25-3.18-boot.log2016-05-16 00:50:16.984951380 -0700 +++ /tmp/ath25-4.4-boot.log 2016-05-16 00:50:30.337017589 -0700 @@ -13,23 +13,23 @@ FLASH: 0xa800 - 0xa87f, 128 blocks of 0x0001 bytes each. == Executing boot script in 3.000 seconds - enter ^C to abort RedBoot> fis load -l vmlinux.bin.l7 -Image loaded from 0x80041000-0x803af7ec +Image loaded from 0x80041000-0x803cc014 RedBoot> exec Now booting linux kernel: Base address 0x8003 Entry 0x80041000 Cmdline : -Linux version 3.18.29 (openwrt@hawg) (gcc version 5.3.0 (LEDE GCC 5.3.0 r241) ) #1 Mon May 16 05:25:24 UTC 2016 +Linux version 4.4.10 (openwrt@hawg) (gcc version 5.3.0 (LEDE GCC 5.3.0 r241) ) #1 Mon May 16 07:13:54 UTC 2016 bootconsole [early0] enabled CPU0 revision is: 00019064 (MIPS 4KEc) Determined physical RAM map: memory: 0200 @ (usable) Initrd not found or empty - disabling initrd Zone ranges: - Normal [mem 0x-0x01ff] + Normal [mem 0x-0x01ff] Movable zone start for each node Early memory node ranges - node 0: [mem 0x-0x01ff] -Initmem setup node 0 [mem 0x-0x01ff] + node 0: [mem 0x-0x01ff] +Initmem setup node 0 [mem 0x-0x01ff] Primary instruction cache 16kB, VIPT, 4-way, linesize 16 bytes. Primary data cache 16kB, 4-way, VIPT, no aliases, linesize 16 bytes Built 1 zonelists in Zone order, mobility grouping on. Total pages: 8128 @@ -37,18 +37,23 @@ PID hash table entries: 128 (order: -3, 512 bytes) Dentry cache hash table entries: 4096 (order: 2, 16384 bytes) Inode-cache hash table entries: 2048 (order: 1, 8192 bytes) -Memory: 28452K/32768K available (2666K kernel code, 107K rwdata, 596K rodata, 136K init, 193K bss, 4316K reserved) +Memory: 28324K/32768K available (2749K kernel code, 108K rwdata, 632K rodata, 148K init, 192K bss, K reserved, 0K cma-reserved) +SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1 NR_IRQS:128 +clocksource: MIPS: mask: 0x max_cycles: 0x, max_idle_ns: 20774570075 ns +sched_clock: 32 bits at 92MHz, resolution 10ns, wraps every 23342213114ns Calibrating delay loop... 183.70 BogoMIPS (lpj=918528) pid_max: default: 32768 minimum: 301 Mount-cache hash table entries: 1024 (order: 0, 4096 bytes) Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes) +clocksource: jiffies: mask: 0x max_cycles: 0x, max_idle_ns: 1911260446275 ns NET: Registered protocol family 16 registering PCI controller with io_map_base unset ar2315-pci ar2315-pci: register PCI controller PCI host bridge to bus :00 pci_bus :00: root bus resource [mem 0x8000-0xbfff] pci_bus :00: root bus resource [io 0x] +pci_bus :00: root bus resource [??? 0x flags 0x0] pci_bus :00: No busn resource found for root bus, will use [bus 00-ff] pci :00:00.0: BAR 1: assigned [mem 0x8000-0x83ff] pci :00:01.0: BAR 1: assigned [mem 0x8400-0x87ff] @@ -68,12 +73,11 @@ pci :00:04.0: BAR 0: assigned [mem 0x9986-0x9987] pci :00:05.0: BAR 0: assigned [mem 0x9988-0x9989] pci :00:06.0: BAR 0: assigned [mem 0x998a-0x998b] -Switched to clocksource MIPS +clocksource: Switched to clocksource MIPS NET: Registered protocol family 2 TCP established hash table entries: 1024 (order: 0, 4096 bytes) TCP bind hash table entries: 1024 (order: 0, 4096 bytes) TCP: Hash tables configured (established 1024 bind 1024) -TCP: reno registered UDP hash table entries: 256 (order: 0, 4096 bytes) UDP-Lite hash table entries: 256 (order: 0, 4096 bytes) NET: Registered protocol family 1 @@ -82,7 +86,6 @@ futex hash table entries: 256 (order: -1, 3072 bytes) squashfs: version 4.0 (2009/01/31) Phillip Lougher jffs2: version 2.2 (NAND) (SUMMARY) (LZMA) (RTIME) (CMODE_PRIORITY) (c) 2001-2006 Red Hat, Inc. -msgmni has been set to 55 io scheduler noop registered io scheduler deadline registered (default) Serial: 8250/16550 driver, 1 ports, IRQ sharing disabled @@ -97,31 +100,30 @@ 6 RedBoot partitions found on MTD device spiflash Creating 6 MTD partitions on "spiflash": 0x-0x0003 : "RedBoot" -0x0003-0x0015 : "vmlinux.bin.l7" -0x0015-0x007e : "rootfs" +0x0003-0x0016 : "vmlinux.bin.l7" +0x0016-0x007e : "rootfs" mtd: device 2 (rootfs) set to be root filesystem 1 squashfs-split partitions found on MTD device rootfs -0x0052-0x007e : "rootfs_data" +0x0053-0x007e : "rootfs_data" 0x007e-0x007ef000 : "FIS directory" 0x007ef000-0x007f : "RedBoot config" 0x007f-0x0080 : "boardconfig" eth0: Atheros AR231x: 00:12:cf:83:7b:08, irq 4 libphy: ar231x_eth_mii: probed eth0: attached PHY driver [Generic