| 00:16.49 | *** join/#edev Lars_G_ (~Lars@unaffiliated/lars-g/x-000001) |
| 03:12.54 | *** join/#edev landley (~landley@70.114.134.56) |
| 04:10.51 | *** join/#edev TheSeven (~quassel@rockbox/developer/TheSeven) |
| 04:27.39 | _av500_ | sh still exists? |
| 04:27.52 | dalias | :) |
| 04:28.10 | dalias | apparently the kernel has not been tested on sh since the gcc3 era or earlier |
| 04:28.33 | dalias | because it's making utterly ridiculous assumptions about the compiler not modifying the argument zone on the stack |
| 04:30.47 | _av500_ | and you are still using it? |
| 04:30.53 | dalias | no |
| 04:31.03 | _av500_ | ah |
| 04:31.07 | dalias | if you check the scrollback, i have someone trying to port musl to sh |
| 04:31.09 | dalias | "just because" |
| 04:31.14 | _av500_ | ic :) |
| 04:31.19 | dalias | and he ran into this while doing the port |
| 04:31.29 | dalias | threads weren't working because clone() is totally broken in the kernel |
| 04:31.59 | dalias | because it accesses parts of the pt_regs structure that the compiler clobbered because the function arguments aliased the part of the stack where pt_regs was stored |
| 04:32.13 | _av500_ | we were recently discussing dropping sh support in libav |
| 04:32.29 | dalias | what would be involved in "dropping support" ? |
| 04:32.35 | *** join/#edev rektide (~rektide@eldergods.com) |
| 04:32.39 | dalias | the C should work anywhere |
| 04:32.42 | _av500_ | yes |
| 04:32.49 | _av500_ | but there is asm in dsputil |
| 04:32.49 | dalias | is there sh asm in libav? :-p |
| 04:32.54 | dalias | haha wow |
| 04:32.56 | dalias | who wrote that? |
| 04:32.58 | _av500_ | and dsputil is being cleaned up |
| 04:33.09 | _av500_ | dalias: some guy in 2007 or so |
| 04:33.14 | _av500_ | but never heard of again |
| 04:33.31 | dalias | if there's no maintenance burden i'd just leave it |
| 04:33.37 | _av500_ | there is |
| 04:33.47 | _av500_ | as said, its getting cleaned up |
| 04:33.50 | _av500_ | so code is being moved |
| 04:33.56 | dalias | but if it needs to be kept in sync with changes in the C, or if it has a chance of breaking build on sh, i'd remove it |
| 04:33.59 | _av500_ | without being tested, we have no idea if its breaking |
| 04:34.17 | _av500_ | renesas told me last year that sh is dead to them |
| 04:34.20 | dalias | basically my criterion would be the stability of the C functions it's replacing |
| 04:34.21 | _av500_ | they use ARM now |
| 04:34.54 | dalias | if they have permanent contracts that won't change, the sh asm has almost no maintenance cost and might should be kept |
| 04:35.09 | _av500_ | as said, we cant even test it |
| 04:35.13 | dalias | but if they could change, then keeping crusty asm is too costly |
| 04:35.17 | _av500_ | I asked renesas for a board to test it |
| 04:35.21 | dalias | :) |
| 04:35.24 | _av500_ | they had no interest |
| 04:35.29 | dalias | qemu could be used |
| 04:35.56 | _av500_ | not sure that has all the sh-4 simd stuff |
| 04:36.28 | _av500_ | or dsp-ish stuff |
| 04:37.02 | dalias | ah maybe not |
| 04:37.15 | dalias | i think current qemu supports sh4a |
| 04:37.46 | _av500_ | as usual, patches welcome |
| 04:38.17 | dalias | :) |
| 04:38.22 | _av500_ | we also dropped stuff like alpha |
| 04:38.33 | dalias | alpha is utterly useless |
| 04:38.44 | dalias | imo things like sh have _some_ use in supporting |
| 04:39.00 | dalias | you'll find them in various devices that might be worth hacking, reappropriating, etc. |
| 04:39.10 | _av500_ | I am sure all the libav/ffmpeg users on SH will cry out |
| 04:39.22 | dalias | but alpha only exists in obsolete server machines that use 10x as much power (or worse) than modern machines that are 10x faster |
| 04:39.44 | dalias | av500, all 3 or 4 users? ;-) |
| 04:39.49 | _av500_ | yes |
| 04:39.59 | dalias | seriously, if there are users, get them to test the asm |
| 04:40.07 | _av500_ | we dont know any users |
| 04:40.12 | _av500_ | that is the point |
| 04:40.15 | dalias | the easiest way to do so might be disabling the asm and waiting for someone to cry |
| 04:40.19 | _av500_ | yes |
| 04:40.24 | _av500_ | doing that :) |
| 04:40.26 | dalias | then asking them to test it |
| 04:40.45 | _av500_ | with their 3.x gcc :) |
| 04:40.49 | dalias | .... |
| 04:42.59 | _av500_ | in fact, I have a ST7109 SH4 in my office |
| 04:43.07 | _av500_ | just nobody who is willing to set it up |
| 04:43.17 | _av500_ | including me |
| 04:45.12 | dalias | :-p |
| 05:55.14 | *** join/#edev clio (~andrej@85.159.109.222) |
| 06:14.31 | *** join/#edev kuldeepdhaka (~kuldeepdh@unaffiliated/kuldeepdhaka) |
| 07:36.57 | *** join/#edev dstuxo (~Dragos@89.121.200.106) |
| 08:05.56 | calculus | dreamcast linux |
| 08:06.38 | calculus | I always wanted a dreamcast |
| 08:26.24 | *** join/#edev florian_kc (~fuchs@port-217-146-132-69.static.qsc.de) |
| 08:26.25 | *** join/#edev florian_kc (~fuchs@Maemo/community/contributor/florian) |
| 08:54.41 | *** join/#edev kuldeepdhaka (~kuldeepdh@unaffiliated/kuldeepdhaka) |
| 09:45.17 | *** join/#edev kuldeepdhaka (~kuldeepdh@unaffiliated/kuldeepdhaka) |
| 10:05.32 | *** join/#edev frank_ (2e903e49@gateway/web/freenode/ip.46.144.62.73) |
| 10:06.19 | *** join/#edev frankp (2e903e49@gateway/web/freenode/ip.46.144.62.73) |
| 11:31.25 | *** join/#edev cousin_luigi (~luigi@unaffiliated/cousinluigi/x-395723) |
| 11:33.35 | cousin_luigi | What device would you recommend to implement a vpn server? Probably just one or two clients at the time with <500kbps overall. |
| 11:33.56 | cousin_luigi | The cheaper the better (but of course I need two ethernet ports) |
| 11:33.58 | SpeedEvil | #openwrt |
| 11:34.23 | cousin_luigi | why? |
| 11:36.58 | SpeedEvil | Because they are likely to know about routery things. |
| 11:37.05 | SpeedEvil | And be able to suggest some. |
| 11:37.15 | cousin_luigi | What is this channel about then?:O |
| 11:55.23 | cousin_luigi | bbl |
| 11:55.24 | *** part/#edev cousin_luigi (~luigi@unaffiliated/cousinluigi/x-395723) |
| 12:13.18 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 14:12.32 | *** join/#edev landley (~landley@70.114.134.56) |
| 15:24.44 | *** join/#edev prpplague (~danders@rrcs-97-77-26-26.sw.biz.rr.com) |
| 15:24.44 | *** mode/#edev [+o prpplague] by ChanServ |
| 16:34.02 | *** join/#edev uRandomMM (40c71302@gateway/web/freenode/ip.64.199.19.2) |
| 16:35.27 | *** join/#edev g1powermac (~g1powerma@76.177.157.108) |
| 16:35.39 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 17:15.35 | *** join/#edev g1powermac (~g1powerma@rrcs-74-87-150-6.west.biz.rr.com) |
| 17:15.42 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 17:26.28 | *** join/#edev uRandomMM (40c71302@gateway/web/freenode/ip.64.199.19.2) |
| 17:48.33 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
| 18:32.31 | *** join/#edev NightMonkey (~NightrMon@173-164-139-193-SFBA.hfc.comcastbusiness.net) |
| 18:32.31 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 18:33.24 | *** join/#edev uRandomMM (40c71302@gateway/web/freenode/ip.64.199.19.2) |
| 18:49.09 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
| 18:50.04 | *** join/#edev florian_kc (~fuchs@Maemo/community/contributor/florian) |
| 19:16.09 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw1-58c385-172.dhcp.inet.fi) |
| 19:55.51 | *** join/#edev uRandomMM (40c71302@gateway/web/freenode/ip.64.199.19.2) |
| 20:16.39 | *** join/#edev NightMonkey (~NightrMon@173-164-139-193-SFBA.hfc.comcastbusiness.net) |
| 20:16.39 | *** join/#edev NightMonkey (~NightrMon@pdpc/supporter/professional/nightmonkey) |
| 23:17.00 | *** join/#edev florian_kc (~fuchs@Maemo/community/contributor/florian) |
| 23:26.07 | *** join/#edev Lars_G (~Lars@unaffiliated/lars-g/x-000001) |