[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <fded1845-98db-f0fb-9320-842c9fdc30dc@vivier.eu>
Date: Fri, 6 Mar 2020 13:07:35 +0100
From: Laurent Vivier <laurent@...ier.eu>
To: YunQiang Su <syq@...ian.org>, Florian Weimer <fw@...eb.enyo.de>
Cc: torvalds@...ux-foundation.org,
Greg KH <gregkh@...uxfoundation.org>,
akpm@...ux-foundation.org, Al Viro <viro@...iv.linux.org.uk>,
James Bottomley <James.Bottomley@...senpartnership.com>,
linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
linux-arch@...r.kernel.org, linux-api@...r.kernel.org,
libc-alpha@...rceware.org
Subject: Re: [PATCH] binfmt_misc: pass binfmt_misc P flag to the interpreter
Le 06/03/2020 à 12:48, YunQiang Su a écrit :
> Florian Weimer <fw@...eb.enyo.de> 于2020年3月6日周五 下午7:42写道:
>>
>> * YunQiang Su:
>>
>>> AT_* only has 32 slot and now. I was afraid that maybe we shouldn't take one.
>>> /* AT_* values 18 through 22 are reserved */
>>> 27,28,29,30 are not used now.
>>> Which should we use?
>>
>> Where does this limit of 32 tags come from? I don't see it from a
>> userspace perspective.
>
> Sorry it is my mistake: In linux/auxvec.h, I saw
>
> #define AT_RANDOM 25 /* address of 16 random bytes */
> #define AT_HWCAP2 26 /* extension of AT_HWCAP */
>
> #define AT_EXECFN 31 /* filename of program */
>
> The number jump to 31 from 26.
>
> It is my fault: in x86_64-linux-gnu/bits/auxv.h, the max number is 47 now.
>
Numbers starting from 32 are arch specific.
18 to 22 are also reserved (and used by some archs).
(linux/arch/*/include/uapi/asm/auxvec.h)
So there is only 4 entries available (27 to 30)
(linux/include/uapi/linux/auxvec.h)
Do we want to waste one more entry whereas we can use an unused one?
Thanks,
Laurent
Powered by blists - more mailing lists