Fix Issues 23130, 19528 : Inline asm was too liberal about operand si…

[maxhaton]

…zes.

A 8-byte type was only considered 8 bytes in 64 bit code, for some reason. This led to inconsistent compilation of extended asm.

I tried to add an error message suggesting the use of an instruction like movsx but this code is spaghetti.

[dlang-bot]

Thanks for your pull request, @maxhaton!

Bugzilla references

Auto-close	Bugzilla	Severity	Description
✓	19528	blocker	asm ptoblem
✓	23130	normal	Inline asm lets you mov to half a variable.

Testing this PR locally

If you don't have a local development environment setup, you can use Digger to test this PR:

dub run digger -- build "stable + dmd#14152"

[thewilsonator]

Stable?

[maxhaton]

I can but the bug was probably misfiled.

[thewilsonator]

I can but the bug was probably misfiled.

OK, master is fine if it is easier to get green.

[ghost]

this breaks -m32 LEA and when performed on values with size > 4. This case should be supported as mov REG, dword ptr [value], i.e get address of value.

[maxhaton]

this breaks -m32 LEA and when performed on values with size > 4. This case should be supported as mov REG, dword ptr [value], i.e get address of value.

Let me take a look.

[RazvanN7]

@maxhaton any progress on this?

[ibuclaw]

@maxhaton still pursuing or close?

[maxhaton]

Still pursuing at some point although set back by the amount of work that needs doing to it overall

On Sun, 15 Jan 2023, 15:02 Iain Buclaw, @.***> wrote:

@maxhaton https://github.com/maxhaton still pursuing or close?

— Reply to this email directly, view it on GitHub https://github.com/dlang/dmd/pull/14152#issuecomment-1383173648, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABLI75HRQEMZR5RS6KPTL3TWSQGQFANCNFSM5WQ76CBA . You are receiving this because you were mentioned.Message ID: @.***>

[RazvanN7]

@maxhaton is this still on your radar?