From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 78413 invoked by alias); 24 Aug 2017 20:27:28 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Received: (qmail 78403 invoked by uid 89); 24 Aug 2017 20:27:27 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,RP_MATCHES_RCVD,SPF_PASS autolearn=ham version=3.3.2 spammy= X-HELO: aserp1040.oracle.com Received: from aserp1040.oracle.com (HELO aserp1040.oracle.com) (141.146.126.69) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 24 Aug 2017 20:27:26 +0000 Received: from userv0022.oracle.com (userv0022.oracle.com [156.151.31.74]) by aserp1040.oracle.com (Sentrion-MTA-4.3.2/Sentrion-MTA-4.3.2) with ESMTP id v7OKRMUm028495 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Thu, 24 Aug 2017 20:27:23 GMT Received: from userv0122.oracle.com (userv0122.oracle.com [156.151.31.75]) by userv0022.oracle.com (8.14.4/8.14.4) with ESMTP id v7OKRMvm012837 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Thu, 24 Aug 2017 20:27:22 GMT Received: from abhmp0008.oracle.com (abhmp0008.oracle.com [141.146.116.14]) by userv0122.oracle.com (8.14.4/8.14.4) with ESMTP id v7OKRMuf010659; Thu, 24 Aug 2017 20:27:22 GMT Received: from [10.159.146.13] (/10.159.146.13) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 24 Aug 2017 13:27:22 -0700 Subject: Re: [PATCH] break gdb build on 32-bit host with ADI support To: Pedro Alves , gdb-patches@sourceware.org References: <1503595405-89600-1-git-send-email-weimin.pan@oracle.com> <479506e1-9478-1a38-d15c-3df13a817fff@redhat.com> <19930243-307d-b126-d4f2-d83eea74c3a4@redhat.com> From: Wei-min Pan Message-ID: <1ed3ad56-1f9e-a6a7-7736-741989c5ed86@oracle.com> Date: Thu, 24 Aug 2017 20:27:00 -0000 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.3.0 MIME-Version: 1.0 In-Reply-To: <19930243-307d-b126-d4f2-d83eea74c3a4@redhat.com> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 8bit X-SW-Source: 2017-08/txt/msg00479.txt.bz2 On 8/24/2017 12:23 PM, Pedro Alves wrote: > On 08/24/2017 08:09 PM, Wei-min Pan wrote: >> >> On 8/24/2017 11:07 AM, Pedro Alves wrote: >>> On 08/24/2017 06:23 PM, Weimin Pan wrote: >>> >>>> diff --git a/gdb/sparc64-tdep.c b/gdb/sparc64-tdep.c >>>> index 6f4fca7..0da2ae5 100644 >>>> --- a/gdb/sparc64-tdep.c >>>> +++ b/gdb/sparc64-tdep.c >>>> @@ -90,12 +90,12 @@ static struct cmd_list_element *sparc64adilist = >>>> NULL; >>>> typedef struct >>>> { >>>> /* The ADI block size. */ >>>> - unsigned long blksize; >>>> + unsigned long long blksize; >>>> /* Number of bits used for an ADI version tag which can be >>>> * used together with the shift value for an ADI version tag >>>> * to encode or extract the ADI version value in a pointer. */ >>>> - unsigned long nbits; >>>> + unsigned long long nbits; >>> Do you really need to count 64-bit bits? :-P :-) >> Since the value of either nbits or blksize is between 0 and 64, >> no, I really don't. But without a 32-bit host, I'm simply play >> safe here so that the compiler won't bark. >>> (Formatting of comment is incorrect for GNU code, BTW. No '*' >>> on each line.) >> Corrected. >> >>>> /* The maximum ADI version tag value supported. */ >>>> int max_version; >>>> @@ -223,9 +223,10 @@ adi_available (void) >>>> proc->stat.checked_avail = true; >>>> if (target_auxv_search (¤t_target, AT_ADI_BLKSZ, >>>> - &proc->stat.blksize) <= 0) >>>> + (CORE_ADDR *)&proc->stat.blksize) <= 0) >>> Please don't introduce potential aliasing problems. Also, missing >>> space before &. >>> >>> Either make blksize really be a CORE_ADDR or do >>> >>> CORE_ADDR value; >>> if (target_auxv_search (¤t_target, AT_ADI_BLKSZ, &value) <= 0) >>> return false; >>> proc->stat.blksize = value; >> Since neither blksize nor nbits is a CORE_ADDR, I'm taking your second >> suggestion. > Then you don't need the > > - unsigned long nbits; > + unsigned long long nbits; > > change anymore.. > >>>> @@ -346,7 +347,8 @@ adi_read_versions (CORE_ADDR vaddr, size_t size, >>>> unsigned char *tags) >>>> if (!adi_is_addr_mapped (vaddr, size)) >>>> { >>>> adi_stat_t ast = get_adi_info (ptid_get_pid (inferior_ptid)); >>>> - error(_("Address at 0x%lx is not in ADI maps"), >>>> vaddr*ast.blksize); >>>> + error(_("Address at 0x%llx is not in ADI maps"), >>>> + (long long)(vaddr*ast.blksize)); >>>> } >>> Use paddress instead? Also spaces around '*' and after the cast. >> Where is paddress defined? I tried casting to "uint64" which yields to >> "unsigned long" on a 64-bit host and didn't bode well with %llx. > $ grep paddress *.h > utils.h:extern const char *paddress (struct gdbarch *gdbarch, CORE_ADDR addr); On a 64-bit host: ... sparc64-tdep.c: In function 'void do_assign(CORE_ADDR, size_t, int)': sparc64-tdep.c:449:56: error: expected ')' before 'vaddr'      error(_("No ADI information at 0x%llx"), (paddress)vaddr);                                                         ^~~~~ sparc64-tdep.c:449:61: error: format '%llx' expects argument of type 'long long unsigned int', but argument 2 has type 'const char* (*)(gdbarch*, CORE_ADDR) {ak a const char* (*)(gdbarch*, long unsigned int)}' [-Werror=format=]      error(_("No ADI information at 0x%llx"), (paddress)vaddr);                                                              ^ cc1plus: all warnings being treated as errors make: *** [sparc64-tdep.o] Error 1 >>>> @@ -387,7 +390,7 @@ adi_print_versions (CORE_ADDR vaddr, size_t cnt, >>>> unsigned char *tags) >>>> while (cnt > 0) >>>> { >>>> QUIT; >>>> - printf_filtered ("0x%016lx:\t", vaddr * adi_stat.blksize); >>>> + printf_filtered ("0x%016llx:\t", (long >>>> long)(vaddr*adi_stat.blksize)); >>> paddress / hex_string / phex_nz ? >> ?? > Try grepping for those things. > > >>> static CORE_ADDR >>> adi_normalize_address (CORE_ADDR addr) >>> { >>> adi_stat_t ast = get_adi_info (ptid_get_pid (inferior_ptid)); >>> >>> if (ast.nbits) >>> return ((CORE_ADDR)(((long)addr << ast.nbits) >> ast.nbits)); >>> return addr; >>> } >>> >>> looks suspiciously bogus to me. Consider a 32-bit host >>> remote/cross debugging a SPARC64 target machine. Also consider >>> a Win64-hosted GDB. >> Good point. Changing it to: >> >> return ((long long)(((long long)addr << ast.nbits) >> ast.nbits)); >> >> Thanks. > Still looks odd to me. > Why are you shifting signed types, for instance? > Any why do you need the casts in the first place, BTW? We need to sign extend to get a normalized address, based on the definitions in ADI space:         ADI versioned address - a VA with ADI bits (63-60) set         Normalized address - a VA with bit 59 sign extended into ADI bits Thanks. > > Thanks, > Pedro Alves