From: Trond Myklebust <trond.myklebust@fys.uio.no>
To: Manfred Spraul <manfred@colorfullife.com>
Cc: Andrea Arcangeli <andrea@suse.de>,
Russell King <rmk@arm.linux.org.uk>,
Hubert Mantel <mantel@suse.de>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Alan Cox <alan@lxorguk.ukuu.org.uk>
Subject: Re: Compatibility issue with 2.2.19pre7
Date: Thu, 11 Jan 2001 18:44:04 +0100 (CET) [thread overview]
Message-ID: <14941.61668.697523.866481@charged.uio.no> (raw)
In-Reply-To: <3A5DDD09.C8C70D36@colorfullife.com>
In-Reply-To: <20010110013755.D13955@suse.de> <200101100654.f0A6sjJ02453@flint.arm.linux.org.uk> <20010110163158.F19503@athlon.random> <shszogy2jmr.fsf@charged.uio.no> <3A5DDD09.C8C70D36@colorfullife.com>
>>>>> " " == Manfred Spraul <manfred@colorfullife.com> writes:
> Trond Myklebust wrote:
>>
>>
>> As for the issue of casting 'fh->data' as a 'struct knfsd' then
>> that is a perfectly valid operation.
>>
> No it isn't.
> fh-> data is an array of characters, thus without any alignment
> restrictions. 'struct knfsd' begins with a pointer, thus it
> must be 4 or 8 byte aligned.
> The portable 'struct nfs_fh' structure would be
> #define NFS_HANDLESIZE 64
> struct nfs_fh {
> unsigned short len; void* data[NFS_HANDLESIZE/sizeof(void*)];
> };
Ok. I see your point now. How about the appended patch then? It means
an extra copy operation, but it should be a lot less ugly than doing
manual alignment...
Cheers,
Trond
diff -u --recursive --new-file linux-2.2.18/fs/lockd/svcsubs.c linux-2.2.18-fix_ppc/fs/lockd/svcsubs.c
--- linux-2.2.18/fs/lockd/svcsubs.c Mon Dec 11 01:49:44 2000
+++ linux-2.2.18-fix_ppc/fs/lockd/svcsubs.c Thu Jan 11 18:43:31 2001
@@ -49,34 +49,37 @@
nlm_lookup_file(struct svc_rqst *rqstp, struct nlm_file **result,
struct nfs_fh *f)
{
- struct knfs_fh *fh = (struct knfs_fh *) f->data;
+ struct knfs_fh fh;
struct nlm_file *file;
unsigned int hash;
u32 nfserr;
+ /* Copy filehandle to avoid pointer alignment issues */
+ memcpy(&fh, f->data, sizeof(fh));
+
dprintk("lockd: nlm_file_lookup(%s/%u)\n",
- kdevname(u32_to_kdev_t(fh->fh_dev)), fh->fh_ino);
+ kdevname(u32_to_kdev_t(fh.fh_dev)), fh.fh_ino);
- hash = file_hash(u32_to_kdev_t(fh->fh_dev), u32_to_ino_t(fh->fh_ino));
+ hash = file_hash(u32_to_kdev_t(fh.fh_dev), u32_to_ino_t(fh.fh_ino));
/* Lock file table */
down(&nlm_file_sema);
for (file = nlm_files[hash]; file; file = file->f_next) {
- if (file->f_handle.fh_dcookie == fh->fh_dcookie &&
- !memcmp(&file->f_handle, fh, sizeof(*fh)))
+ if (file->f_handle.fh_dcookie == fh.fh_dcookie &&
+ !memcmp(&file->f_handle, &fh, sizeof(fh)))
goto found;
}
dprintk("lockd: creating file for %s/%u\n",
- kdevname(u32_to_kdev_t(fh->fh_dev)), fh->fh_ino);
+ kdevname(u32_to_kdev_t(fh.fh_dev)), fh.fh_ino);
nfserr = nlm4_lck_denied_nolocks;
file = (struct nlm_file *) kmalloc(sizeof(*file), GFP_KERNEL);
if (!file)
goto out_unlock;
memset(file, 0, sizeof(*file));
- file->f_handle = *fh;
+ memcpy(&file->f_handle, &fh, sizeof(file->f_handle));
file->f_sema = MUTEX;
/* Open the file. Note that this must not sleep for too long, else
@@ -85,7 +88,7 @@
* We have to make sure we have the right credential to open
* the file.
*/
- if ((nfserr = nlmsvc_ops->fopen(rqstp, fh, &file->f_file)) != 0) {
+ if ((nfserr = nlmsvc_ops->fopen(rqstp, &fh, &file->f_file)) != 0) {
dprintk("lockd: open failed (nfserr %d)\n", ntohl(nfserr));
goto out_free;
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
next prev parent reply other threads:[~2001-01-11 17:48 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-01-10 0:37 Hubert Mantel
2001-01-10 0:43 ` Alan Cox
2001-01-10 6:54 ` Russell King
2001-01-10 15:31 ` Andrea Arcangeli
2001-01-10 22:09 ` Russell King
2001-01-10 23:59 ` Andrea Arcangeli
2001-01-11 7:34 ` Russell King
2001-01-11 10:33 ` Andi Kleen
2001-01-11 10:36 ` Russell King
2001-01-11 12:10 ` Manfred
2001-01-11 12:10 ` Andi Kleen
2001-01-11 12:29 ` Manfred
2001-01-11 13:27 ` Russell King
2001-01-24 7:51 ` Richard Henderson
2001-01-24 9:02 ` Andrea Arcangeli
2001-01-24 9:51 ` Richard Henderson
2001-01-24 12:21 ` Andrea Arcangeli
2001-01-24 17:49 ` Richard Henderson
2001-01-24 10:09 ` Thomas Pornin
2001-01-11 15:28 ` Trond Myklebust
2001-01-11 16:19 ` Manfred Spraul
2001-01-11 17:44 ` Trond Myklebust [this message]
2001-01-11 18:22 ` Trond Myklebust
2001-01-11 18:27 ` Andrea Arcangeli
2001-01-11 18:30 ` Trond Myklebust
2001-01-11 18:43 ` Andrea Arcangeli
2001-01-11 20:09 ` Russell King
2001-01-11 20:39 ` Trond Myklebust
2001-01-11 11:37 ` Trond Myklebust
2001-01-24 13:46 Jesse Pollard
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=14941.61668.697523.866481@charged.uio.no \
--to=trond.myklebust@fys.uio.no \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=andrea@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=manfred@colorfullife.com \
--cc=mantel@suse.de \
--cc=rmk@arm.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®