Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756464Ab0LJARO (ORCPT ); Thu, 9 Dec 2010 19:17:14 -0500 Received: from swampdragon.chaosbits.net ([90.184.90.115]:16105 "EHLO swampdragon.chaosbits.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756189Ab0LJARN (ORCPT ); Thu, 9 Dec 2010 19:17:13 -0500 Date: Fri, 10 Dec 2010 01:10:34 +0100 (CET) From: Jesper Juhl To: Ky Srinivasan cc: zbr@ioremap.net, devel@linuxdriverproject.org, virtualization@lists.osdl.org, hjanssen@microsoft.com, gregkh@suse.de, linux-kernel@vger.kernel.org, Haiyang Zhang Subject: Re: [PATCH 1/1] Properly check return values of kmalloc and vmbus_recvpacket In-Reply-To: <4D013908.E57C.0030.1@novell.com> Message-ID: References: <1291926209-17120-1-git-send-email-hjanssen@microsoft.com> <4D013908.E57C.0030.1@novell.com> User-Agent: Alpine 2.00 (LNX 1167 2008-08-23) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4954 Lines: 140 On Thu, 9 Dec 2010, Ky Srinivasan wrote: > > > >>> On 12/9/2010 at 3:23 PM, in message > <1291926209-17120-1-git-send-email-hjanssen@microsoft.com>, Hank Janssen > wrote: > > Correct ugly oversight, we need to check the return values of kmalloc > > and vmbus_recvpacket and return if they fail. I also tightened up the > > call to kmalloc. > > > > Thanks to Evgeniy Polyakov for pointing this out. > > > > Signed-off-by: Haiyang Zhang > > Signed-off-by: Hank Janssen > > > > --- > > drivers/staging/hv/hv_utils.c | 48 ++++++++++++++++++++++++++++------------ > > 1 files changed, 33 insertions(+), 15 deletions(-) > > > > diff --git a/drivers/staging/hv/hv_utils.c b/drivers/staging/hv/hv_utils.c > > index 53e1e29..ac68575 100644 > > --- a/drivers/staging/hv/hv_utils.c > > +++ b/drivers/staging/hv/hv_utils.c > > @@ -43,21 +43,27 @@ static void shutdown_onchannelcallback(void *context) > > { > > struct vmbus_channel *channel = context; > > u8 *buf; > > - u32 buflen, recvlen; > > + u32 recvlen; > > u64 requestid; > > u8 execute_shutdown = false; > > + int ret = 0; > > > > struct shutdown_msg_data *shutdown_msg; > > > > struct icmsg_hdr *icmsghdrp; > > struct icmsg_negotiate *negop = NULL; > > > > - buflen = PAGE_SIZE; > > - buf = kmalloc(buflen, GFP_ATOMIC); > > + buf = kmalloc(PAGE_SIZE, GFP_ATOMIC); > > > > - vmbus_recvpacket(channel, buf, buflen, &recvlen, &requestid); > > + if (!buf) { > > + printk(KERN_INFO > > + "Unable to allocate memory for shutdown_onchannelcallback"); > > + return; > > + } > > + > > + ret = vmbus_recvpacket(channel, buf, PAGE_SIZE, &recvlen, &requestid); > > > > - if (recvlen > 0) { > > + if (ret == 0 && recvlen > 0) { > > DPRINT_DBG(VMBUS, "shutdown packet: len=%d, requestid=%lld", > > recvlen, requestid); > > > > @@ -151,17 +157,23 @@ static void timesync_onchannelcallback(void *context) > > { > > struct vmbus_channel *channel = context; > > u8 *buf; > > - u32 buflen, recvlen; > > + u32 recvlen; > > u64 requestid; > > struct icmsg_hdr *icmsghdrp; > > struct ictimesync_data *timedatap; > > + int ret = 0; > > > > - buflen = PAGE_SIZE; > > - buf = kmalloc(buflen, GFP_ATOMIC); > > + buf = kmalloc(PAGE_SIZE, GFP_ATOMIC); > > > > - vmbus_recvpacket(channel, buf, buflen, &recvlen, &requestid); > > + if (!buf) { > > + printk(KERN_INFO > > + "Unable to allocate memory for timesync_onchannelcallback"); > > + return; > > + } > > > > - if (recvlen > 0) { > > + ret = vmbus_recvpacket(channel, buf, PAGE_SIZE, &recvlen, &requestid); > > + > > + if (ret == 0 && recvlen > 0) { > > DPRINT_DBG(VMBUS, "timesync packet: recvlen=%d, requestid=%lld", > > recvlen, requestid); > > > > @@ -197,17 +209,23 @@ static void heartbeat_onchannelcallback(void *context) > > { > > struct vmbus_channel *channel = context; > > u8 *buf; > > - u32 buflen, recvlen; > > + u32 recvlen; > > u64 requestid; > > struct icmsg_hdr *icmsghdrp; > > struct heartbeat_msg_data *heartbeat_msg; > > + int ret = 0; > > + > > + buf = kmalloc(PAGE_SIZE, GFP_ATOMIC); > > > > - buflen = PAGE_SIZE; > > - buf = kmalloc(buflen, GFP_ATOMIC); > > + if (!buf) { > > + printk(KERN_INFO > > + "Unable to allocate memory for heartbeat_onchannelcallback"); > > + return; > > + } > > > > - vmbus_recvpacket(channel, buf, buflen, &recvlen, &requestid); > > + ret = vmbus_recvpacket(channel, buf, PAGE_SIZE, &recvlen, &requestid); > > > > - if (recvlen > 0) { > > + if (ret == 0 && recvlen > 0) { > > DPRINT_DBG(VMBUS, "heartbeat packet: len=%d, requestid=%lld", > > recvlen, requestid); > > > I am wondering if it might be better to allocate a page during module > startup for dealing with some of these services. Since the protocol > guarantees that there cannot be more than one outstanding request from > the host side, having a pre-allocated receive buffer would be a more > robust solution - we don't have to allocate memory when we cannot afford > to sleep and thereby don't have to deal with a class of failure > conditions that are not easy to deal with. For instance not being able > to shut the guest down because of low memory situation would be bad. > IMVHO; nicer to have a module fail at load time with ENOMEM than having failures later. If memory requirement is known at load time (and is fairly low), just allocate what's needed then and then there's no unexpected failure later. -- Jesper Juhl http://www.chaosbits.net/ Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html Plain text mails only, please. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/