Addressed review comments.
Updated an explanation in a comment.

Addressed review comments.

Abandoned because evaluateRelocExpr() needs to be rewritten anyway.

Abandoned due to existing problems with evaluateRelocExpr().

Rafael, are the common-code additions ok with you ?

Fixed a dumb bug in my original {set,clear}ModuleFeatureBits():
*The if was preventing the setting of the front options.
Added my proposed test case.
Added dependency on D10657.

Fixed restoration of previous hardfloat settings:

• The settings aren't actually changed, it's just that when in soft-float they are ignored.
• emitDirectiveModule{SoftFloat,HardFloat}() doesn't change the MipsABIFlagsSection data anymore, it only prints it.
• That data is changed by setting/clearing the soft-float bit in STI.FeatureBits and then calling updateABIInfo().
• updateABIInfo() uses the update mechanism from MipsABIFlagsSection to properly update.

Changed test to use -mips-abi-flags feature of llvm-readobj.
Made the module-hardfloat.s test more thorough.
Rebased on top of D10641, D10642, and D10643, which fix the other .module directives.

I was thinking of the following test:

In D6231#191632, @dsanders wrote:

That's ok for this patch, but following the principle above I'd suggest asserting that LargeGOT is not in use.

Fixed the RuntimeDyld test.
Addressed the other review comments.

Rebased.

Right now, there still are some missing pieces:

Addressed review comments.
Changed behaviour to more closely match GAS.
Refactored if-else control flow to early returns.
Added test cases for DLA with $at as an operand.
Added more test cases for unusual behaviours of (D)LA.
Added test cases for mips-expansions-bad.s.
Removed some unnecessary generation checks to fix mips-expansions-bad.s.
Removed unnecessary patch dependecies.
Rebased on top of D10568.

Merged expandUlw() and expandUlhu() together.
Rebased top of the updated D9671.

Addressed review comments (one of them resulted in D10568).
Added a code comment with an example of why we add the SrcReg locally.
Fixed mistake in test/MC/Mips/set-nomacro.s caused by a previous rebase.

Outlined the (inPicMode() && !(isABI_N32() || isABI_N64())) condition into a separate function.
Switched to using !isInt<16>() in createCpRestoreMemOp() if condition.
Moved forbidModuleDirective() calls to the MipsTargetStreamer base class.
Removed redundant addition to the .set reorder NOP-creation condition.
Improved code comments to better explain why we need a NOP after the JALR in .set noreorder.

In D6231#187264, @dsanders wrote:

Looking back through the comments, I believe the remaining issues are:

• Duplicated nops

Improved the code which gets the JalSym from the original JalExpr.
Moved relocation creation before MCInst creation in order to improve readability. (NFC)
Added FIXME comments for the remaining problems.

Daniel, is shortening IntermediateDstReg to TmpReg OK with you ?
You pointed that out in private, but since I ended up making the change, your position should be included in the public review.

I think this is the final version of this patch, as all of its dependencies have been committed.
Daniel, could take another look at this please ?

Addressed LGTM nit.

Addressed LGTM style nit.
Updated the ATReg-related code comment, as a result of this refactoring.

Ran clang-format on changes.

Added test case with MIPS-specific expression.

Addressed LGTM comment.

Moved ATReg inside the if block, as it is not used again after that.
Shortened the name of IntermediateDstReg to TmpReg.
Rebased on top of the new D9523, D9348 and D9366.
Removed unnecessary dependencies on other patches.

Moved ATReg inside the if block, as it is not used again after that.

Replaced the UseAT variable with its corresponding condition, as it was being used only once.
Put the "SrcReg != Mips::NoRegister" condition in a variable, as it is now used twice.
Shortened the name of IntermediateDstReg to TmpReg.
Rebased on top of the new D9523 and D9348.
Removed unnecessary dependencies on other patches.

Rebased on top of the new D9523 and made changes accordingly.
Removed unnecessary dependencies on other patches.

Removed all dependencies on other patches, as they were unnecessary.

Rebased.

Rebased.

Rebased (for Sean Bruno's convenience).

Rebased (for Sean Bruno's convenience).

Rebased (for Sean Bruno's convenience).

Will address review comments and other issues soon.

Rebased (for Sean Bruno's convenience).

Addressed review comments.
Rebased (for Sean Bruno's convenience).