On 04/28/2015 09:49 AM, William Roberts wrote: > > On Apr 28, 2015 5:49 AM, "Stephen Smalley" <[email protected] > <mailto:[email protected]>> wrote: >> >> On 04/27/2015 07:44 PM, [email protected] > <mailto:[email protected]> wrote: >> > Jumping in off of this old thread: >> > http://marc.info/?l=seandroid-list&m=140560177420732&w=2 >> > >> > Their seems to be 3 approaches that centered around resolving sysfs > entry >> > labeling that I have found, they are: >> > 1. Fixup ueventd to add "online" event support for sysfs_fixup_perms() >> > calls and update the uevent config file >> > >> > 2. Add a restorecon_recursive in the init.rc for the offending path >> > >> > 3. Correct sysfs so new nodes coming in inherent parent labels. >> > >> > Whats the status on 3, was that ever taken up? >> >> Not AFAIK. On #1, I don't believe we get "online" events for these >> files in ueventd; the ueventd code already includes support for "online" >> events and calls sysfs_fixup_perms() in > > Well support was there for create. So if you have an entry in uevent > config file then it should call fixup.
handle_device_event() calls fixup_sys_perms() on "add", "change", or "online". fixup_sys_perms() sets owner/mode based on ueventd.rc, but unconditionally calls restorecon_recursive() on the node these days even without an entry in ueventd.rc, see https://android-review.googlesource.com/#/c/100249/ But for these particular nodes, the kernel is not generating any such event AFAICS. _______________________________________________ Seandroid-list mailing list [email protected] To unsubscribe, send email to [email protected]. To get help, send an email containing "help" to [email protected].
