Date: Tue, 06 Jan 2026 18:26:33 +0100 From: =?utf-8?Q?Dag-Erling_Sm=C3=B8rgrav?= <des@FreeBSD.org> To: John Baldwin <jhb@FreeBSD.org> Cc: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-main@FreeBSD.org Subject: Re: git: 27894e20f140 - main - libgeom: Fix segfault in 32-on-64 case Message-ID: <86eco2hcty.fsf@ltc.des.dev> In-Reply-To: <cce035dc-ca9b-4dc0-81f7-22b92da52217@FreeBSD.org> (John Baldwin's message of "Tue, 6 Jan 2026 08:37:12 -0600") References: <6958dd10.b4b9.2aebecda@gitrepo.freebsd.org> <cce035dc-ca9b-4dc0-81f7-22b92da52217@FreeBSD.org>
index | next in thread | previous in thread | raw e-mail
John Baldwin <jhb@FreeBSD.org> writes: > Should we perhaps not use pointers to hold the cookies? This is going to truncate > in the lib32 case which will probably still work in practice as the low 32 bits of > kernel object addresses are probably unique, but isn't foolproof. Perhaps the cookie > values should be stored as either kvaddr_t values, or uintmax_t? Unfortunately this is effectively part of the KBI, these values get cross-referenced against pointers in struct devstat returned by kern.devstat.all. I'm working on fixing that as well. Also, the way the code works, the “cookies” get stored in pointer members and then later resolved to the actual pointers to the corresponding struct, so no, we can't easily not use pointers. DES -- Dag-Erling Smørgrav - des@FreeBSD.orghome | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86eco2hcty.fsf>