Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 2 Aug 2018 06:58:51 +0200
From:   Dominique Martinet <asmadeus@codewreck.org>
To:     v9fs-developer@lists.sourceforge.net
Cc:     Dominique Martinet <dominique.martinet@cea.fr>,
        linux-kernel@vger.kernel.org, Matthew Wilcox <willy@infradead.org>,
        Greg Kurz <groug@kaod.org>, linux-fsdevel@vger.kernel.org
Subject: Re: [V9fs-developer] [PATCH v2 2/2] net/9p: add a per-client fcall
 kmem_cache
Message-ID: <20180802045851.GA15564@nautica>
References: <1532943263-24378-1-git-send-email-asmadeus@codewreck.org>
 <1533177452-2165-1-git-send-email-asmadeus@codewreck.org>
 <1533177452-2165-2-git-send-email-asmadeus@codewreck.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <1533177452-2165-2-git-send-email-asmadeus@codewreck.org>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Dominique Martinet wrote on Thu, Aug 02, 2018:
> [...]
> +	clnt->fcall_cache = kmem_cache_create("9p-fcall-cache", clnt->msize,
> +					      0, 0, NULL);


Well, my gut feeling that I'd need a v3 was right, after a bit more time
testing (slightly different setup), I got this warning:
Bad or missing usercopy whitelist? Kernel memory overwrite attempt
detected to SLUB object '9p-fcall-cache' (offset 23, size 55078)!

So this kmem_cache_create needs to change to kmem_cache_create_usercopy,
but I'm not sure how to best specify the range -- this warning was about
writing data to the buffer so a TWRITE:
size[4] Twrite tag[2] fid[4] offset[8] count[4] data[count]
(the data[count] part comes from user - this matches 4+1+2+4+8+4 = 23)

Similarily RREAD has data that can be copied to userspace, which has a
similar check:
size[4] Rread tag[2] count[4] data[count]

So I could hardcode offset = 4+1+2+4=11, usercopy size = msize - 11...


We have some P9_*HDR*SZ macros but none are really appropriate:
---
/* ample room for Twrite/Rread header */
#define P9_IOHDRSZ      24

/* Room for readdir header */
#define P9_READDIRHDRSZ 24

/* size of header for zero copy read/write */
#define P9_ZC_HDR_SZ 4096
---

It's actually been bugging me for a while that I see hardcoded '7's for
9p main header size (len + msg type + tag) everywhere, I could add a
first P9_HDRSZ of 7 and use that + 4?

-- 
Dominique