auto f = [](auto &&m) {
    enum { _,e3,e2,e1,C4,C3,C2,C1 };
    static constexpr int x_coeffs[3][4] = {
        {e1,C2,C3,C4},
        {e2,C1,C3,C4},
        {e3,C1,C2,C4},
    };
};

int main() {
    f(0);
}



<source>: In instantiation of '<lambda(auto:1&&)> [with auto:1 = int]':
<source>:11:6:   required from here
<source>:4:10: internal compiler error: in tsubst_copy, at cp/pt.cc:16910
    4 |         {e1,C2,C3,C4},
      |          ^~

Confirmed.  Started with r11-7993-g9f4c41147a41d0

Looking into it..

The patch above changed

-      if (TREE_CODE (template_type) == ENUMERAL_TYPE && !is_dependent_type
+      if (TREE_CODE (template_type) == ENUMERAL_TYPE
+         && !uses_template_parms (current_nonlambda_scope ())
          && !DECL_ALIAS_TEMPLATE_P (gen_tmpl))

but here current_nonlambda_scope () is ::, a NAMESPACE_DECL, which doesn't have a type, so is considered type-dependent.  So we don't call tsubst_enum.  That doesn't look right.

(In reply to Marek Polacek from comment #4)
> The patch above changed
> 
> -      if (TREE_CODE (template_type) == ENUMERAL_TYPE && !is_dependent_type
> +      if (TREE_CODE (template_type) == ENUMERAL_TYPE
> +         && !uses_template_parms (current_nonlambda_scope ())
>           && !DECL_ALIAS_TEMPLATE_P (gen_tmpl))
> 
> but here current_nonlambda_scope () is ::, a NAMESPACE_DECL, which doesn't
> have a type, so is considered type-dependent.  So we don't call tsubst_enum.
> That doesn't look right.

Whoops, sorry Marek, didn't know you were already looking into this.  Your analysis makes sense to me :)

Testing a fix.

The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:409edcca331296b53842c50d3b789e1b1ccc05e5

commit r12-8290-g409edcca331296b53842c50d3b789e1b1ccc05e5
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Apr 26 16:12:58 2022 -0400

    c++: enum in generic lambda at global scope [PR105398]
    
    We crash compiling this test since r11-7993 which changed
    lookup_template_class_1 so that we only call tsubst_enum when
    
      !uses_template_parms (current_nonlambda_scope ())
    
    But here current_nonlambda_scope () is the global NAMESPACE_DECL ::, which
    doesn't have a type, therefore is considered type-dependent.  So we don't
    call tsubst_enum, and crash in tsubst_copy/CONST_DECL because we didn't
    find the e1 enumerator.
    
    I don't think any namespace can depend on any template parameter, so
    this patch tweaks uses_template_parms.
    
            PR c++/105398
    
    gcc/cp/ChangeLog:
    
            * pt.cc (uses_template_parms): Return false for any NAMESPACE_DECL.
    
    gcc/testsuite/ChangeLog:
    
            * g++.dg/cpp1y/lambda-generic-enum2.C: New test.

Fixed on trunk so far, will backport to 11.4.

The releases/gcc-11 branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:3a1358e5f38864bf5688a7b6db1fda482321a77a

commit r11-9941-g3a1358e5f38864bf5688a7b6db1fda482321a77a
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Apr 26 16:12:58 2022 -0400

    c++: enum in generic lambda at global scope [PR105398]
    
    We crash compiling this test since r11-7993 which changed
    lookup_template_class_1 so that we only call tsubst_enum when
    
      !uses_template_parms (current_nonlambda_scope ())
    
    But here current_nonlambda_scope () is the global NAMESPACE_DECL ::, which
    doesn't have a type, therefore is considered type-dependent.  So we don't
    call tsubst_enum, and crash in tsubst_copy/CONST_DECL because we didn't
    find the e1 enumerator.
    
    I don't think any namespace can depend on any template parameter, so
    this patch tweaks uses_template_parms.
    
            PR c++/105398
    
    gcc/cp/ChangeLog:
    
            * pt.c (uses_template_parms): Return false for any NAMESPACE_DECL.
    
    gcc/testsuite/ChangeLog:
    
            * g++.dg/cpp1y/lambda-generic-enum2.C: New test.
    
    (cherry picked from commit 409edcca331296b53842c50d3b789e1b1ccc05e5)

Fixed.