Date: Wed, 06 Jun 2018 07:35:12 -0400 From: Ravi Pokala <rpokala@freebsd.org> To: Mateusz Guzik <mjg@FreeBSD.org>, <src-committers@freebsd.org>, <svn-src-all@freebsd.org>, <svn-src-head@freebsd.org> Subject: Re: svn commit: r334702 - head/sys/sys Message-ID: <6E6E92B2-7536-4281-8EAF-72823E84902E@panasas.com> In-Reply-To: <201806060508.w56586c9053686@repo.freebsd.org> References: <201806060508.w56586c9053686@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Mateusz, =EF=BB=BF-----Original Message----- From: <owner-src-committers@freebsd.org> on behalf of Mateusz Guzik <mjg@Fr= eeBSD.org> Date: 2018-06-06, Wednesday at 01:08 To: <src-committers@freebsd.org>, <svn-src-all@freebsd.org>, <svn-src-head@= freebsd.org> Subject: svn commit: r334702 - head/sys/sys > Author: mjg > Date: Wed Jun 6 05:08:05 2018 > New Revision: 334702 > URL: https://svnweb.freebsd.org/changeset/base/334702 >=20 > Log: > malloc: elaborate on r334545 due to frequent questions I was one of the questioners. :-) Thank you for explaining the conditional = logic. > ... > + * Passing the flag down requires malloc to blindly zero the entire obje= ct. > + * In practice a lot of the zeroing can be avoided if most of the object > + * gets explicitly initialized after the allocation. Letting the compile= r > + * zero in place gives it the opportunity to take advantage of this stat= e. This part, I still don't understand. :-( The call to bzero() is still for the full length passed in, so how does thi= s help? > ... > + * _malloc_item =3D malloc(_size, type, (flags) &~ M_ZERO); > + * if (((flags) & M_WAITOK) !=3D 0 || _malloc_item !=3D NULL) > + * bzero(_malloc_item, _size); > + * > + * If the flag is set, the compiler knows the left side is always true, > + * therefore the entire statement is true and the callsite is: I think you mean "... the *right* side is always true ...", since the left = side is the check for the flag being set. "If the flag is set, compiler know= s (the check for the flag being set) is always true" is tautological. > ... > + * If the flag is not set, the compiler knows the left size is always fa= lse > + * and the NULL check is needed, therefore the callsite is: Same issue here. > ... > #ifdef _KERNEL > #define malloc(size, type, flags) ({ \ Now that I'm taking another look at this, I'm confused as to why the entire= macro expansion is inside parentheses? (The braces make sense, since this i= s a block with local variables which need to be contained.) > void *_malloc_item; \ > @@ -193,7 +228,8 @@ void *malloc(size_t size, struct malloc_type *type, i= n > if (__builtin_constant_p(size) && __builtin_constant_p(flags) &&\ > ((flags) & M_ZERO) !=3D 0) { \ > _malloc_item =3D malloc(_size, type, (flags) &~ M_ZERO); \ > - if (((flags) & M_WAITOK) !=3D 0 || _malloc_item !=3D NULL) \ > + if (((flags) & M_WAITOK) !=3D 0 || \ > + __predict_true(_malloc_item !=3D NULL)) \ > bzero(_malloc_item, _size); \ > } else { \ > _malloc_item =3D malloc(_size, type, flags); \ This confuses me too. If the constant-size/constant-flags/M_ZERO-is-set tes= t fails, then it falls down to calling malloc(). Which we are in the middle = of defining. So what does that expand to? Thanks, Ravi (rpokala@)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6E6E92B2-7536-4281-8EAF-72823E84902E>