From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 584E6C43334 for ; Tue, 21 Jun 2022 19:14:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1352164AbiFUTOW (ORCPT ); Tue, 21 Jun 2022 15:14:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:35204 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229918AbiFUTOV (ORCPT ); Tue, 21 Jun 2022 15:14:21 -0400 Received: from mx01.omp.ru (mx01.omp.ru [90.154.21.10]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 1FC3E13E09 for ; Tue, 21 Jun 2022 12:14:17 -0700 (PDT) Received: from [192.168.1.103] (178.176.77.12) by msexch01.omp.ru (10.188.4.12) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384) id 15.2.986.14; Tue, 21 Jun 2022 22:14:06 +0300 Subject: Re: [PATCH] ata: libata-scsi: fix sloppy result type of ata_ioc32() To: Damien Le Moal , References: <4a695619-2de7-671e-7b67-2afddf22423f@omp.ru> <9c231ffe-df3c-9c36-4b8b-dc518009c3d2@opensource.wdc.com> From: Sergey Shtylyov Organization: Open Mobile Platform Message-ID: Date: Tue, 21 Jun 2022 22:14:06 +0300 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.10.1 MIME-Version: 1.0 In-Reply-To: <9c231ffe-df3c-9c36-4b8b-dc518009c3d2@opensource.wdc.com> Content-Type: text/plain; charset="utf-8" Content-Language: en-US Content-Transfer-Encoding: 7bit X-Originating-IP: [178.176.77.12] X-ClientProxiedBy: msexch01.omp.ru (10.188.4.12) To msexch01.omp.ru (10.188.4.12) X-KSE-ServerInfo: msexch01.omp.ru, 9 X-KSE-AntiSpam-Interceptor-Info: scan successful X-KSE-AntiSpam-Version: 5.9.20, Database issued on: 06/21/2022 18:56:16 X-KSE-AntiSpam-Status: KAS_STATUS_NOT_DETECTED X-KSE-AntiSpam-Method: none X-KSE-AntiSpam-Rate: 59 X-KSE-AntiSpam-Info: Lua profiles 171268 [Jun 21 2022] X-KSE-AntiSpam-Info: Version: 5.9.20.0 X-KSE-AntiSpam-Info: Envelope from: s.shtylyov@omp.ru X-KSE-AntiSpam-Info: LuaCore: 491 491 a718ef6dc942138335b0bcd7ab07f27b5c06005e X-KSE-AntiSpam-Info: {rep_avail} X-KSE-AntiSpam-Info: {Tracking_from_domain_doesnt_match_to} X-KSE-AntiSpam-Info: {relay has no DNS name} X-KSE-AntiSpam-Info: {SMTP from is not routable} X-KSE-AntiSpam-Info: {Found in DNSBL: 178.176.77.12 in (user) b.barracudacentral.org} X-KSE-AntiSpam-Info: {Found in DNSBL: 178.176.77.12 in (user) dbl.spamhaus.org} X-KSE-AntiSpam-Info: 127.0.0.199:7.1.2;d41d8cd98f00b204e9800998ecf8427e.com:7.1.1;omp.ru:7.1.1;178.176.77.12:7.7.3,7.7.1 X-KSE-AntiSpam-Info: ApMailHostAddress: 178.176.77.12 X-KSE-AntiSpam-Info: {DNS response errors} X-KSE-AntiSpam-Info: Rate: 59 X-KSE-AntiSpam-Info: Status: not_detected X-KSE-AntiSpam-Info: Method: none X-KSE-AntiSpam-Info: Auth:dmarc=temperror header.from=omp.ru;spf=temperror smtp.mailfrom=omp.ru;dkim=none X-KSE-Antiphishing-Info: Clean X-KSE-Antiphishing-ScanningType: Heuristic X-KSE-Antiphishing-Method: None X-KSE-Antiphishing-Bases: 06/21/2022 19:00:00 X-KSE-AttachmentFiltering-Interceptor-Info: protection disabled X-KSE-Antivirus-Interceptor-Info: scan successful X-KSE-Antivirus-Info: Clean, bases: 6/21/2022 4:24:00 PM X-KSE-BulkMessagesFiltering-Scan-Result: InTheLimit Precedence: bulk List-ID: X-Mailing-List: linux-ide@vger.kernel.org On 6/21/22 1:44 AM, Damien Le Moal wrote: [...] >>>> While ata_ioc32() returns *int*, its result gets assigned to and compared >>>> with the *unsigned long* variable 'val' in ata_sas_scsi_ioctl(), its only >>>> caller, which implies a problematic implicit cast -- fix that by returning >>>> *unsigned long* instead. >>>> >>>> Found by Linux Verification Center (linuxtesting.org) with the SVACE static >>>> analysis tool. >>>> >>>> Signed-off-by: Sergey Shtylyov >>>> >>>> --- >>>> This patch is against the 'for-next' branch of Damien's 'libata.git' repo. >>>> >>>> drivers/ata/libata-scsi.c | 2 +- >>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>> >>>> Index: libata/drivers/ata/libata-scsi.c >>>> =================================================================== >>>> --- libata.orig/drivers/ata/libata-scsi.c >>>> +++ libata/drivers/ata/libata-scsi.c >>>> @@ -539,7 +539,7 @@ int ata_task_ioctl(struct scsi_device *s >>>> return rc; >>>> } >>>> >>>> -static int ata_ioc32(struct ata_port *ap) >>>> +static unsigned long ata_ioc32(struct ata_port *ap) >>>> { >>>> if (ap->flags & ATA_FLAG_PIO_DMA) >>>> return 1; >> >>> Actually, this should be a bool I think and the ioctl code cleaned to use >> >> By the ioctl code you mean ata_sas_scsi_ioctl()? > > yes. > >>> that type since the val argument of the ioctl is also used as a bool. >> >> As for HDIO_SET_32BIT, that's prolly OK but what to do with HDIO_GET_32BIT >> (it calls put_user() on *unsigned long*)? > > Something like this should work fine: > > diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c > index 86dbb1cdfabd..ec7f79cbb135 100644 > --- a/drivers/ata/libata-scsi.c > +++ b/drivers/ata/libata-scsi.c [...] > @@ -571,16 +572,16 @@ int ata_sas_scsi_ioctl(struct ata_port *ap, struct > scsi_device *scsidev, > return put_user(val, (unsigned long __user *)arg); > > case HDIO_SET_32BIT: Hmm, I told you this *case* is prolly OK -- it was HDIO_GET_32BIT *case* that I was concerned about... So you mean that HDIO_GET_32BIT handling should remain intact? > - val = (unsigned long) arg; > + pio32 = !!((unsigned long) arg); > rc = 0; > spin_lock_irqsave(ap->lock, flags); > if (ap->pflags & ATA_PFLAG_PIO32CHANGE) { > - if (val) > + if (pio32) > ap->pflags |= ATA_PFLAG_PIO32; > else > ap->pflags &= ~ATA_PFLAG_PIO32; > } else { > - if (val != ata_ioc32(ap)) > + if (pio32 != ata_ioc32(ap)) > rc = -EINVAL; > } > spin_unlock_irqrestore(ap->lock, flags); Not really sure this is worth it... *unsigned long* result type for ata_ioc32() seems simpler. MBR, Sergey