On Mon, Oct 30, 2017 at 06:01:44PM +0100, Arnd Bergmann wrote: > Right, either fail the build or use the workaround from the gcc bugzilla, > passing -fno-store-merging to the broken compilers avoids the problem > reliably at the cost of slightly worse code. For the decompressor, we might > also get away with passing -march=armv5t (not armv5te), which has > experimentally been found to avoid the problem and produce better code > than "-march=armv6 -fno-store-merging" as it avoids the strd instruction > but not other store merging.
Depends how important the compilers are... We currently refuse to build using these compilers: gcc = 3.x where x < 3 gcc = 4.x where x < 3 if unwinding is enabled gcc = 4.8.x where x < 3 -- RMK's Patch system: http://www.armlinux.org.uk/developer/patches/ FTTC broadband for 0.8mile line in suburbia: sync at 8.8Mbps down 630kbps up According to speedtest.net: 8.21Mbps down 510kbps up
