Some potential progress: https://bugzilla.kernel.org/show_bug.cgi?id=203573
On 5/30/19 2:50 PM, Rolf Fokkens wrote:
Not being sure if people here follow the issue at Fedora I'd like to
pass a suggestion:
https://bugzilla.redhat.com/show_bug.cgi?id=1708315#c27
On 5/22/19 1:44 AM, Nix wrote:
On 21 May 2019, Coly Li uttered the following:
Also I try to analyze the assemble code of bcache, just find out the
generated assembly code between gcc9 and gcc7 is quite different. For
gcc9 there is a XXXX.cold part. So far I can not tell where the problem
is from yet.
This is hot/cold partitioning. You can turn it off with
-fno-reorder-blocks-and-partition and see if that helps things (and if
it doesn't, it should at least make stuff easier to compare).
