| --- Log opened Fri Nov 07 00:00:17 2014 | ||
| stekern | poke53282: | 07:13 |
|---|---|---|
| stekern | ~ # ./pastie-9700857 | 07:13 |
| stekern | -1 | 07:13 |
| stekern | 9 | 07:13 |
| stekern | but I don't have the bugfix applied yet | 07:15 |
| poke53282 | I don't think the buxfix is important here. | 07:23 |
| poke53282 | the same as root? | 07:24 |
| poke53282 | do you get the same result on you x86 machine? | 07:24 |
| stekern | yeah, I don't think the bugfix is important neither | 07:28 |
| stekern | well, it's important, but not for this case ;) | 07:28 |
| stekern | I can't see the mkstemp trying to open the file when I strace it, so I assume it fails here: http://git.musl-libc.org/cgit/musl/tree/src/temp/mkostemps.c#n13 | 07:35 |
| stekern | if I add a printf of the name to your pastie, I get this: | 07:38 |
| stekern | name = /tmp/myTmpFile-XXXXXXBT | 07:38 |
| stekern | which would explain why that check fails | 07:38 |
| stekern | you have a off-by-one in your strncpy | 07:42 |
| stekern | it should be: strncpy(name,"/tmp/myTmpFile-XXXXXX",22); | 07:42 |
| stekern | to account for the \0 | 07:43 |
| stekern | poke53282: ^ | 07:49 |
| olofk | No more heartbleed or shellshock now | 08:53 |
| olofk | ysionneau: It would be great if you could find out which revision of DFI you are using | 09:37 |
| maxpaln | olofk: back again. How're you getting on? I have pretty much finished adding the random wait state functionality. | 15:24 |
| maxpaln | Looking at the code, there is no reason for the 2d arrays to be outside the wb_bfm_master module. In fact, the code would be neater if they were in there. | 15:25 |
| maxpaln | I will move them into the wb_bfm_master module and take them out of the task headers - it should make for more transportable code. | 15:26 |
| maxpaln | olofk: a new zip file is here: https://www.dropbox.com/s/xvwvo6va9cvpudr/backend.rar?dl=0 | 16:28 |
| poke53282 | Ok, thanks stekern | 16:33 |
| maxpaln | hold on - just spotted a bug, BRB | 16:33 |
| maxpaln | k, here's the link: https://www.dropbox.com/s/xvwvo6va9cvpudr/backend.rar?dl=0 | 16:37 |
| poke53282 | stekern: It works, unfortunately I get a "Operation not supported on transport endpoint" error on the second function. But this is the 9p filesystem. Maybe I should change musl, that it uses ftruncate instead. Otherwise wayland/weston fails. | 16:37 |
| maxpaln | I've fixed a bunch of bugs, removed the 2d arrays from the task headers, added random wait states (default 0 to 8 - but use the MAX_WAIT_STATES parameter to change), I've also added the ability to set the random seed via a plusarg on the simulation command line: vsim +seed=3 | 16:38 |
| maxpaln | I think this should be close to all the functionality I expected to add. See if it works as well on your design. | 16:39 |
| poke53282 | Finally jor1k supports UTF-8 for terminal and filesystem. During the development I realized that the swedish keyboard has a very strange sign at "4" which looks like a curved square. What's this? | 17:03 |
| stekern | poke53282: it = ? | 17:18 |
| poke53282 | Ahh, this is the generic currency sign | 17:28 |
| stekern | poke53282: what is the actual problem? is the actual code doing what you're example is doing? | 18:44 |
| stekern | if so, you could mount a tmpfs on /tmp | 18:44 |
| stekern | that you can fallocate on | 18:45 |
| poke53282 | Well, I had two problems. The one with mkstemp is solved. My fault. | 18:55 |
| poke53282 | The other is a limitation of the filesystem, or better a limitation of code using this fallocate function without proper error handling. | 18:56 |
| poke53282 | http://cgit.freedesktop.org/wayland/weston/tree/shared/os-compatibility.c#n143 | 19:00 |
| poke53282 | I have already a patch, that weston don't use fallocate. | 19:01 |
| poke53282 | But anyhow, using tmpfs is a good idea | 19:20 |
| --- Log closed Sat Nov 08 00:00:19 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!