|
From: Florian K. <fl...@ei...> - 2015-09-20 11:22:12
|
On 19.09.2015 15:40, Mark Wielaard wrote: > On Mon, 2015-09-07 at 15:44 +0200, Julian Seward wrote: >> I plan to branch for 3.11.0 in the next 24 hours or so. Please >> yell ASAP if you believe there to be big breakage on the trunk >> that should be fixed before that. >> >> If the branch seems stable then I propose to let the branch stabilise >> until Mon 21 Sept and release at that point. > > I think the branch and valgrind-3.11.0.TEST1 are pretty stable and good. > I have been testing a bit with some additional patches to get slightly > less test failures and suppress some warnings. These are not urgently > needed, but would be nice to have in the final release: > > r15647 > libvex_test: Use arm64_[di]MinLine_lg2_szB values that make libvex > happy. > > r15650 > Fix various compiler warnings for the arm architecture. > > r15651 > Do not compile with -Wcast-align on arm. > > r15653 > Add Power PC ISA check to the vbit-test (bug #352765) > > r15654 > Filter out glibc warning messages about unexpected futex results in > tests. > > Additionally there are the following 3 VEX revisions for some issues on > ppc: > > r3188 > guard dis_dfp_fmt_conv and dis_dfp_exponent_test with allow_DFP > > r3189 > Add support for the Power PC Program Priority Register (bug #352769) > > =Note= there is one open review comment about this patch from Matthias. > In the ppc64 case the guest_PSPB is initialized to zero. The comment and > the ppc32 case explicitly say it should be non-zero. > > r3190 > Add support for the Power PC mbar instruction (bug #352768) > > With or without these patches I think valgrind 3.11.0 will be a great > release. > We should also consider the small patch set by Jean Delvare which was posted to valgrind-developers sep 9: "VKI_I2C_SMBUS: Fixes and cleanups". This patch set relates to r14908 which is the changeset for BZ #342603. This is a bug that we say was fixed in 3.11. So we should fix it properly. I cannot tell whether the patch is correct or not. But as the person who provided r14908 was on copy and has not spoken up since, I presume this is good to go in. I'll commit to HEAD shortly. Florian |