linux-kernel - Re: [GIT PULL] more s390 updates for 6.6 merge window

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <ZRKIJJNnOVNzyP/J@infradead.org>
Date:   Tue, 26 Sep 2023 00:28:36 -0700
From:   Christoph Hellwig <hch@...radead.org>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Heiko Carstens <hca@...ux.ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>,
        Alexander Gordeev <agordeev@...ux.ibm.com>,
        linux-s390@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [GIT PULL] more s390 updates for 6.6 merge window

On Thu, Sep 07, 2023 at 11:02:23AM -0700, Linus Torvalds wrote:
> On Thu, 7 Sept 2023 at 04:27, Heiko Carstens <hca@...ux.ibm.com> wrote:
> >
> >   Note that the set_memory_xx() family comes with an int num_pages
> >   parameter which overflows with 8TB. This could be addressed by
> >   changing the num_pages parameter to unsigned long, however requires
> >   to change all architectures, since the module code expects an int
> >   parameter (see module_set_memory()).
> 
> Bah. We should just do this. Change the 'start' to a pointer, and
> change the page count to 'unsigned long'.
> 
> Changing module_set_memory() to take the right kind of function
> pointer looks trivial too.
> 
> I'm not even sure why we did that "unsigned long addr" thing, but I
> assume it is because of our (very very _very_ old) historical
> get_free_pages() model.
> 
> We should probably change get_free_pages() too, but that literally
> goes back to linux-0.01. It's horrendously bad, but it's _very_
> traditional, and it comes from the VM code historically using bitops
> on the page pointers.

Yes, please.  And maybe pick a better name for the void * returning
version.  kmalloc_pages/kfree_pages?  or is that to close to
kmalloc/kfree?