[00:51:46] --- Simon Wilkinson has become available [01:09:41] --- Simon Wilkinson has left [01:43:14] --- abo has left [01:43:15] --- abo has become available [01:43:19] --- abo has left [01:43:19] --- abo has become available [02:10:42] --- abo has left [02:10:42] --- abo has become available [02:11:00] --- abo has left [02:11:01] --- abo has become available [02:13:10] --- abo has left [02:14:08] --- abo has become available [02:14:51] --- abo has left [02:14:51] --- abo has become available [02:15:19] --- abo has left [02:15:20] --- abo has become available [02:15:39] --- Simon Wilkinson has become available [02:19:14] --- abo has left [02:19:15] --- abo has become available [04:20:55] --- gendalia has become available [06:06:31] --- mvita has become available [06:25:40] --- meffie has become available [06:48:39] --- wiesand has become available [07:30:19] Marc: thanks for the pointer on that sha256-kernel issue [07:31:05] np. the build system is a bit convoluted there [08:26:14] --- Marc Dionne has left [08:27:28] --- Marc Dionne has become available [08:38:33] --- wiesand has left [10:14:03] --- deason has become available [10:43:21] --- ballbery has become available [10:57:44] --- ballbery has left [10:57:48] --- ballbery has become available [11:04:43] I'm a little confused at why it's okay to include afsconfig.h in the kernel build, but it seems pretty intentional (i.e., crypto/hcrypto/kernel/config.h) [11:06:00] Because afsconfig.h also includes user selected options, not just autoconf discovered ones [11:06:21] It also includes autoconf discovered kernel options (for example, for Linux compatibility tests) [11:06:40] The OS X build (on Mavericks with Xcode 5.0.2, at least) has a lot of warnings about implicit declarations which imply that we have some header issues, but I'm not sure that I'm going to try to track them down anytime soon. [11:07:33] I'm just imagining a case where a function is available to userspace but not to kernel consumers. Like, say ... open(). [11:07:55] I guess that's still not my actual point, hang on. [11:09:26] One of the warnings from the OSX build was about arc4random() in the rand-fortuna-kernel build, and that's what brought this to mind. Since there is this lore about systems that don't export randomness functions to kernel modules and all. (HAVE_ARC4RANDOM is defined in afsconfig.h based on a userspace check, I believe.) [11:13:03] If you're using fortuna on Mac OS X, you've already lost. [11:13:10] Just use the kernel RNG? [11:13:40] --- ballbery has left [11:15:32] --- ballbery has become available [11:15:39] Well, yeah; I'm not concerned about OS X itself. [11:15:48] It just popped up the warning that caused me to notice. [11:16:26] The only place this should be a problem is in the Heimdal code. Everything else (in theory) knows when it is being built in userspace and when it is being built for kernel [11:16:38] Okay. [11:16:50] I will proceed to not worry about it very much. [11:26:12] --- Simon Wilkinson has left [11:37:20] --- meffie has left [11:40:18] --- gendalia85502 has become available [11:40:18] --- gendalia85502 has left [11:40:18] --- gendalia87003 has become available [12:03:08] --- ballbery has left [12:03:59] --- ballbery has become available [13:51:00] --- ballbery has left [14:01:58] --- ballbery has become available [14:12:04] --- Marc Dionne has left [14:26:24] --- meffie has become available [14:53:28] --- ballbery has left [15:02:13] --- ballbery has become available [15:14:27] --- Simon Wilkinson has become available [15:42:53] --- Marc Dionne has become available [16:07:56] --- Simon Wilkinson has left [16:37:41] --- mvita has left [18:02:17] Thanks for looking into the reading-past-the-end-of-the-file question, Andrew. [18:23:03] --- deason has left [19:04:56] What is the difference between afspag.ko and libafs.ko? [19:33:15] I think afspag.ko is only for the NFS translator [19:40:19] Ah. Apparently it needs afs_osi.o, so I need to scope the initialization to just libafs.ko. [22:50:59] --- mvita has become available