From mboxrd@z Thu Jan 1 00:00:00 1970 From: Borislav Petkov Subject: Re: [patch v2] arcmsr: buffer overflow in arcmsr_iop_message_xfer() Date: Fri, 23 Sep 2016 13:22:26 +0200 Message-ID: <20160923112226.rteir5ivjou64ffh@pd.tnic> References: <20160915134456.GA30277@mwanda> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Return-path: Received: from mail.skyhub.de ([78.46.96.112]:60181 "EHLO mail.skyhub.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755175AbcIWLWf (ORCPT ); Fri, 23 Sep 2016 07:22:35 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: "Martin K. Petersen" Cc: Dan Carpenter , "James E.J. Bottomley" , Ching Huang , Hannes Reinicke , Johannes Thumshirn , Tomas Henzl , linux-scsi@vger.kernel.org, security@kernel.org On Thu, Sep 15, 2016 at 09:59:01AM -0400, Martin K. Petersen wrote: > >>>>> "Dan" == Dan Carpenter writes: > > Dan> We need to put an upper bound on "user_len" so the memcpy() doesn't > Dan> overflow. > > Applied to 4.9/scsi-queue. Yap, Tomas said the kfree was missing on the error path but can we simplify this further by doing the user_len check first so that the kfree() is not even needed? Patch ontop: --- From: Borislav Petkov Date: Fri, 23 Sep 2016 13:04:51 +0200 Subject: [PATCH] scsi: arcmsr: Simplify user_len checking Do the user_len check first and then the ver_addr allocation so that we can save us the kfree() on the error path when user_len is > ARCMSR_API_DATA_BUFLEN. Signed-off-by: Borislav Petkov Cc: Marco Grassi Cc: Dan Carpenter Cc: Tomas Henzl Cc: Martin K. Petersen --- drivers/scsi/arcmsr/arcmsr_hba.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/arcmsr/arcmsr_hba.c b/drivers/scsi/arcmsr/arcmsr_hba.c index 110eca9eaca0..3d53d636b17b 100644 --- a/drivers/scsi/arcmsr/arcmsr_hba.c +++ b/drivers/scsi/arcmsr/arcmsr_hba.c @@ -2391,18 +2391,20 @@ static int arcmsr_iop_message_xfer(struct AdapterControlBlock *acb, uint32_t user_len; int32_t cnt2end; uint8_t *pQbuffer, *ptmpuserbuffer; - ver_addr = kmalloc(ARCMSR_API_DATA_BUFLEN, GFP_ATOMIC); - if (!ver_addr) { + + user_len = pcmdmessagefld->cmdmessage.Length; + if (user_len > ARCMSR_API_DATA_BUFLEN) { retvalue = ARCMSR_MESSAGE_FAIL; goto message_out; } - ptmpuserbuffer = ver_addr; - user_len = pcmdmessagefld->cmdmessage.Length; - if (user_len > ARCMSR_API_DATA_BUFLEN) { + + ver_addr = kmalloc(ARCMSR_API_DATA_BUFLEN, GFP_ATOMIC); + if (!ver_addr) { retvalue = ARCMSR_MESSAGE_FAIL; - kfree(ver_addr); goto message_out; } + ptmpuserbuffer = ver_addr; + memcpy(ptmpuserbuffer, pcmdmessagefld->messagedatabuffer, user_len); spin_lock_irqsave(&acb->wqbuffer_lock, flags); -- 1.8.5.2 -- Regards/Gruss, Boris. ECO tip #101: Trim your mails when you reply.