[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <wfl47fj3l4xhffrwuqfn5pgtrrn3s64lxxodnz5forx7d4x443@spsi3sx33lnf>
Date: Mon, 12 Jan 2026 13:46:44 +0100
From: Michal Koutný <mkoutny@...e.com>
To: Andrei Vagin <avagin@...gle.com>
Cc: Kees Cook <kees@...nel.org>, linux-kernel@...r.kernel.org,
linux-fsdevel@...r.kernel.org, linux-mm@...ck.org, criu@...ts.linux.dev,
Andrew Morton <akpm@...ux-foundation.org>, Chen Ridong <chenridong@...wei.com>,
Christian Brauner <brauner@...nel.org>, David Hildenbrand <david@...nel.org>,
Eric Biederman <ebiederm@...ssion.com>, Lorenzo Stoakes <lorenzo.stoakes@...cle.com>
Subject: Re: [PATCH 2/3] exec: inherit HWCAPs from the parent process
On Thu, Jan 08, 2026 at 05:07:47AM +0000, Andrei Vagin <avagin@...gle.com> wrote:
> @@ -1780,6 +1791,50 @@ static int bprm_execve(struct linux_binprm *bprm)
> return retval;
> }
>
> +static void inherit_hwcap(struct linux_binprm *bprm)
> +{
> + int i, n;
> +
> +#ifdef ELF_HWCAP4
> + n = 4;
> +#elif defined(ELF_HWCAP3)
> + n = 3;
> +#elif defined(ELF_HWCAP2)
> + n = 2;
> +#else
> + n = 1;
> +#endif
Is it guaranteed that HWCAP n+1 exists only when n does?
(To make this work.)
> +
> + for (i = 0; n && i < AT_VECTOR_SIZE; i += 2) {
> + long val = current->mm->saved_auxv[i + 1];
> +
> + switch (current->mm->saved_auxv[i]) {
> + case AT_HWCAP:
> + bprm->hwcap = val & ELF_HWCAP;
> + break;
> +#ifdef ELF_HWCAP2
> + case AT_HWCAP2:
> + bprm->hwcap2 = val & ELF_HWCAP2;
> + break;
> +#endif
> +#ifdef ELF_HWCAP3
> + case AT_HWCAP3:
> + bprm->hwcap3 = val & ELF_HWCAP3;
> + break;
> +#endif
> +#ifdef ELF_HWCAP4
> + case AT_HWCAP4:
> + bprm->hwcap4 = val & ELF_HWCAP4;
> + break;
> +#endif
> + default:
> + continue;
> + }
> + n--;
> + }
> + mm_flags_set(MMF_USER_HWCAP, bprm->mm);
Will this work when mm->saved_auxv isn't set by the prctl (it is
zeroes?)?
Thanks,
Michal
Download attachment "signature.asc" of type "application/pgp-signature" (266 bytes)
Powered by blists - more mailing lists