I have a set of minor changes I've made while developing the Multisession and Leading-login support that I think may be beneficial to the open-iscsi project as well.
Repo: git://repo.or.cz/open-iscsi/multisession.git Tag: submitted/minor_fixes_v1 I hope they should be self-explanatory from the git commit messages, but I'm also including a quick summary here: 1/7: Fix dcb_app.c compile error with old kernels Allows the HEAD of the tree to compile on systems with 2.6.28 or earlier. 2/7: Check all ifaces during discovery even if some timeout There is a bug in the current discovery logic that will cause ALL interfaces to fail to complete their discovery operation if even ONE of them cannot reach the given target portal. This change only addresses transport errors such as connection timeouts, since other ifaces may succeed if they are on different subnets or different physical links. Other errors, such as out-of-memory or invalid login credentials that are not likely to succeed on other ifaces if one fails, should probably remain fatal to the entire discovery process. It should be easy to add other error cases to the list that do not warrant aborting all discoveries. 3/7: iscsid: In foreground mode, treat SIGINT like SIGTERM This allows ^C to exit iscsid but only when it is running in foreground mode, which is useful for testing. 4/7: Revise bind_conn_to_iface logic Cleans up what I found to be mildly confusing logic - Basically when you have both iface->hwaddress and iface->netdev set, netdev trumps hwaddress even though at a quick glance the code seemed to intend otherwise. With this change, hwaddress takes precedence first, then netdev. 5/7: Fixing compiler warnings in iscsi tools Compiler warnings are a pet peeve of mine, since my editor has a tendency to jump to them after compiling. All of the warnings I found here were fairly trivial, but I've always approached compiler warnings with the philosophy of: You need to track them down since some are important and may lead to ugly runtime complications, and if you go through the work to find out they are unimportant it's trivial to fix them so they don't warn any more. One I could not fix - a link-time warning about the fact that with glibc you can't truly statically link any app that uses 'getaddrinfo' due to the way glibc uses dynamic loading within nsswitch to do address resolution. 6/7 and 7/7: Fixing compiler warnings in fwparam_ibft and open-isns: These ones I'm less sure about submitting here, since I'm unclear as to whether or not the fwparam_ibft or open-isns code is still being maintained externally or just in this codebase. I also do not have had the facilities to test these changes, though I believe they should be functionally neutral. 6/7: fwparam_ibft - I updated prom_lex.c using flex-2.5.35 which, coupled with a couple minor changes in prom_lex.l and fwparam_ppc.c, removed the compiler warnings. But beware that prom_lex.c does not adhere to the normal coding standards, being an auto-generated file. 7/7: open-isns - Fairly minor fixes here. -- Jim Ramsay -- You received this message because you are subscribed to the Google Groups "open-iscsi" group. To post to this group, send email to open-iscsi@googlegroups.com. To unsubscribe from this group, send email to open-iscsi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/open-iscsi?hl=en.
