From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=56064 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1Pzrn0-0001yc-Mz for qemu-devel@nongnu.org; Wed, 16 Mar 2011 10:33:31 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Pzrmz-0006GK-FC for qemu-devel@nongnu.org; Wed, 16 Mar 2011 10:33:30 -0400 Received: from e34.co.us.ibm.com ([32.97.110.152]:43620) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Pzrmz-0006Fv-9m for qemu-devel@nongnu.org; Wed, 16 Mar 2011 10:33:29 -0400 Received: from d03relay02.boulder.ibm.com (d03relay02.boulder.ibm.com [9.17.195.227]) by e34.co.us.ibm.com (8.14.4/8.13.1) with ESMTP id p2GELgfg028054 for ; Wed, 16 Mar 2011 08:21:42 -0600 Received: from d03av02.boulder.ibm.com (d03av02.boulder.ibm.com [9.17.195.168]) by d03relay02.boulder.ibm.com (8.13.8/8.13.8/NCO v9.1) with ESMTP id p2GEXRYN047440 for ; Wed, 16 Mar 2011 08:33:27 -0600 Received: from d03av02.boulder.ibm.com (loopback [127.0.0.1]) by d03av02.boulder.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id p2GEXQKU021248 for ; Wed, 16 Mar 2011 08:33:26 -0600 Message-ID: <4D80CA34.6040205@linux.vnet.ibm.com> Date: Wed, 16 Mar 2011 07:33:24 -0700 From: "Venkateswararao Jujjuri (JV)" MIME-Version: 1.0 References: <20110315103453.GA23922@linux.vnet.ibm.com> <20110315103901.GD23922@linux.vnet.ibm.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: [Qemu-devel] Re: [v1 PATCH 3/3]: Convert v9fs_stat to threaded model. List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Stefan Hajnoczi Cc: arun@linux.vnet.ibm.com, aliguori@us.ibm.com, qemu-devel@nongnu.org, aneesh.kumar@linux.vnet.ibm.com On 3/16/2011 3:23 AM, Stefan Hajnoczi wrote: > On Tue, Mar 15, 2011 at 10:39 AM, Arun R Bharadwaj > wrote: >> -static void v9fs_stat_post_lstat(V9fsState *s, V9fsStatState *vs, int err) >> +static void v9fs_stat_post_lstat(void *opaque) >> { >> - if (err == -1) { >> - err = -errno; >> + V9fsStatState *vs = (V9fsStatState *)opaque; > > No need to cast void* in C. > >> + if (vs->err == -1) { >> + vs->err = -(vs->v9fs_errno); > > How about the thread worker function puts the -errno into a vs->ret field: > > static void v9fs_stat_do_lstat(V9fsRequest *request) > { > V9fsStatState *vs = container_of(request, V9fsStatState, request); > > vs->ret = v9fs_do_lstat(vs->s, &vs->fidp->fsmap.path, &vs->stbuf); > if (vs->ret != 0) { > vs->ret = -errno; > } > } > > Then v9fs_stat_post_lstat() can use vs->ret directly and does not need > to juggle around the two fields, vs->err and vs->v9fs_errno. > >> goto out; >> } >> >> - err = stat_to_v9stat(s, &vs->fidp->fsmap.path, &vs->stbuf, &vs->v9stat); >> - if (err) { >> + vs->err = stat_to_v9stat(vs->s, &vs->fidp->fsmap.path, &vs->stbuf, &vs->v9stat); > > This function can block in v9fs_do_readlink(). Needs to be done > asynchronously to avoid blocking QEMU. > >> + if (vs->err) { >> goto out; >> } >> vs->offset += pdu_marshal(vs->pdu, vs->offset, "wS", 0, &vs->v9stat); >> - err = vs->offset; >> + vs->err = vs->offset; >> >> out: >> - complete_pdu(s, vs->pdu, err); >> + complete_pdu(vs->s, vs->pdu, vs->err); >> v9fs_stat_free(&vs->v9stat); >> qemu_free(vs); >> } >> >> +static void v9fs_stat_do_lstat(V9fsRequest *request) >> +{ >> + V9fsStatState *vs = container_of(request, V9fsStatState, request); > > Nice. Could container_of() be used for v9fs_post_lstat() too? I'm > suggesting making post op functions take the V9fsRequest* instead of a > void* opaque pointer. > >> + >> + vs->err = v9fs_do_lstat(vs->s, &vs->fidp->fsmap.path, &vs->stbuf); > > This is not threadsafe since rpath still uses a static buffer in > qemu.git. Please ensure that rpath() is thread-safe before pushing > this patch. There is another patch on the internal list to make rpath thread safe. > >> + vs->v9fs_errno = errno; >> +} >> + >> static void v9fs_stat(V9fsState *s, V9fsPDU *pdu) >> { >> int32_t fid; >> @@ -1487,6 +1496,10 @@ static void v9fs_stat(V9fsState *s, V9fsPDU *pdu) >> vs = qemu_malloc(sizeof(*vs)); >> vs->pdu = pdu; >> vs->offset = 7; >> + vs->s = s; >> + vs->request.func = v9fs_stat_do_lstat; >> + vs->request.post_op.func = v9fs_stat_post_lstat; >> + vs->request.post_op.arg = vs; >> >> memset(&vs->v9stat, 0, sizeof(vs->v9stat)); >> >> @@ -1498,8 +1511,11 @@ static void v9fs_stat(V9fsState *s, V9fsPDU *pdu) >> goto out; >> } >> >> + /* >> err = v9fs_do_lstat(s, &vs->fidp->fsmap.path, &vs->stbuf); >> v9fs_stat_post_lstat(s, vs, err); >> + */ > > Please remove unused code, it quickly becomes out-of-date and confuses readers. > >> + v9fs_qemu_submit_request(&vs->request); > > What happens when another PDU is handled next that uses the same fid? > The worst case is if the client sends TCLUNK and fid is freed while > the worker thread and later the post op still access the memory. > There needs to be some kind of guard (like a reference count) to > prevent this. As per the protocol this should not happen. Client is the controls the fid, and the fid is created or destroyed per the directive of the client. It should not send clunk until the response is received on that fid based operation(if there is any). - JV > > Stefan