[03:09:45] *** Joins: luyikei (~quassel@i255020.dynamic.ppp.asahi-net.or.jp) [05:15:52] *** Quits: [Franklin] (~franklin@unaffiliated/franklin) (Remote host closed the connection) [06:38:01] *** Quits: TheSeven (~quassel@rockbox/developer/TheSeven) (Ping timeout: 245 seconds) [06:39:44] *** Joins: TheSeven (~quassel@rockbox/developer/TheSeven) [06:56:42] *** Quits: luyikei (~quassel@i255020.dynamic.ppp.asahi-net.or.jp) (Remote host closed the connection) [09:41:50] *** Quits: [Saint] (~saint@rockbox/staff/saint) (Quit: Quit.) [09:42:16] *** Joins: [Saint] (~saint@rockbox/staff/saint) [11:08:14] *** Joins: goom (~go4m@cpe-66-25-153-174.satx.res.rr.com) [16:45:20] *** Quits: [Saint] (~saint@rockbox/staff/saint) (*.net *.split) [16:45:24] *** Joins: [Franklin] (~franklin@unaffiliated/franklin) [16:45:59] *** Joins: [Saint] (~saint@rockbox/staff/saint) [17:40:31] user890104: sure, if you know which code is causing that ;) [17:42:17] TheSeven: i'm going to remove your write-caching code, and leave only the read-caching [17:42:29] readahead* [17:43:43] feel free to try that, but I'm not convinced that it will solve the problem [17:44:28] there are several other conditions that could cause corruption (old data in read ahead buffer, race conditions in buffer/job management) [17:44:56] misalignment, cache issues [17:46:09] do you have a way to reproduce the problem? it might be helpful to have a tool that nails down the specific sequence of accesses that causes to corruption and the way in which the data was corrupted in order to both pinpoint the cause and check whether it's fixed [18:03:59] *** Quits: [Franklin] (~franklin@unaffiliated/franklin) (Quit: Lost terminal) [18:06:18] TheSeven: i remeber that my initial version didn't have these issues, but it was a bit slow [18:07:03] you refactored some parts of it, and implemented read/write ahead in a single commit, so i'm not able to test these two improvements separately [18:07:12] or maybe we just didn't notice :P [18:07:19] anyway, feel free to try :) [18:07:31] well, i copied about 30 GB of data without errors [18:07:53] do you have any information about the nature of the corruption that occurs? [18:08:00] skipping read/write-ahead at all should be more stable, right? [18:08:06] no, but i can make some tests [18:08:13] do sectors get written to wrong locations, not written at all, garbage data written, etc.? [18:08:26] or are there maybe readback errors which confuse the filesystem driver? [18:08:50] i can check that, but first i need a way to access the drive without getting wrong data :) [18:09:34] that's why i have decided to make a more simplified version, to minimise the chances of hitting a bug [18:09:48] we have a usb debug api for that [18:10:11] ah, right. it can read raw hdd sectors, right? [18:10:16] yes [18:10:39] ok, i'll share my findings after making some tests later [23:42:33] *** Joins: INeedToUninstall (444b1a05@gateway/web/freenode/ip.68.75.26.5) [23:43:41] Hi guys, I need to uninstall on my ipod 6g. [23:48:13] Thread: http://forums.rockbox.org/index.php/topic,50515.msg233552.html#msg233552 [23:48:33] Please email me jdblackiii@gmail.com