I think what everyone needs to just think and remember for a minute: * OpenWRT is built by the community and not by highly paid engineers * OpenWRT is often far more feature rich than stock firmwares written for the devices supported * OpenWRT is even used sometimes by hardware vendors as a baseline for their own firmwares
I think that suggesting OpenWRT is done by guesswork and not done by people who know their code is just counter productive and if we're to criticize the hard work of the OpenWRT team then we should at least be upbuilding in what we say rather than just saying OpenWRT performance sucks balls. -----Original Message----- From: openwrt-devel [mailto:openwrt-devel-boun...@lists.openwrt.org] On Behalf Of Zefir Kurtisi Sent: Wednesday, 11 March 2015 11:13 PM To: Felix Fietkau; David Lang; José Vázquez Cc: OpenWrt Development List Subject: Re: [OpenWrt-Devel] Why OpenWrt sucks? On 03/10/2015 10:39 PM, Felix Fietkau wrote: > On 2015-03-10 20:22, Zefir Kurtisi wrote: > >> This gives at least two sources for optimization for the reference >> driver: 1) save inter-layer processing overhead (passing commands >> from hostapd directly to HW is cheaper than passing it through 4 >> layers), and 2) vastly reduce locking when you are in full control of concurrency. > The overhead of commands issued by hostapd is pretty much irrelevant. > All that matters for driver performance is the actual data path > (network stack, mac80211 tx/rx, driver tx/rx). > Also, I've not seen any indication in my tests that locking is a > measurable cause of performance issues. > When it comes to performance issues, guesswork is meaningless, > measurements (e.g. profiling results) matter! > > - Felix > Hm, maybe my memory is deceiving me, but I'm pretty sure my statement is not based on guesswork but in fact on measurements. Like 3 years ago, for a PPC based platform we had to decide whether to go with the reference driver or with aht9k. For that, thorough performance measurements were performed and ath9k was somewhere between 10 and 30% behind throughput wise. Guessing that synchronization might be one factor, we removed locking in ath9k and results improved noticeably by up to 5%. At that time and on a PPC platform it was a factor, it might be irrelevant today. I remember it well since it indicated that ath9k has some optimization potential left and we finally decided for it. Anyway, that was not my point here. What I claim is that a proprietary driver always opens more optimization potential (cross-layer wise). Cheers, Zefir _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
