[Bug 71701] Regression, Dual-Head, i915 mode-setting driver, "pipe B stuck"

[bugzilla-daemon at bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=71701

Kris Karas <bugs-a13 at moonlit-rail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEEDINFO                    |VERIFIED
         Resolution|---                         |CODE_FIX

--- Comment #17 from Kris Karas <bugs-a13 at moonlit-rail.com> ---
As no activity has been seen on this bug in more than 2 weeks, I am taking the
liberty of submitting a tested patch, attachment 130941 above, which fixes the
regression.

Apparently, https://bugs.freedesktop.org/show_bug.cgi?id=71267 reported a case
wherein DisplayPort handshaking never completes for some combination of Haswell
chipset and a docking station.  The fix for that bug was to introduce a timeout
into the aux_native_read and aux_native_write routines.  Unfortunately, the
"fix" used 2,800 microseconds as a timeout, sufficient perhaps for the Haswell
chipset itself but far too aggressive to deal with other devices that may be
connected to the DisplayPort.  In particular, a Dell E176 VGA monitor connected
via a BizLink ORN699 DP->VGA adapter dongle required an average of 25
milliseconds for the ACK to be received.

The patch submitted above extends the timeout delay to 100ms, hopefully
sufficient to accommodate an esoteric assortment of downstream hardware.

Last but not least, as the cause for this bug is now well known, I am changing
the status of this patch to VERIFIED.

-- 
You are receiving this mail because:
You are on the CC list for the bug.