From vamsi.krishnak at gmail.com Wed Jun 1 19:00:56 2005 From: vamsi.krishnak at gmail.com (vamsi krishna) Date: Thu, 2 Jun 2005 00:30:56 +0530 Subject: pagefault resulting in SIGBUS Message-ID: <3faf0568050601120043b955a1@mail.gmail.com> Hello, I have a program which is trying to access its virtual memory segments (listed in /proc//maps file). But while reading the following virtual memory segment as below. <--------------------------------------------------------------------------------------------------------------------> 0x2a95556000 0x2a9566b000 0x115000 0 /lib64/ld-2.3.2.so 0x2a9566b000 0x2a9566c000 0x1000 0x15000 /lib64/ld-2.3.2.so <--------------------------------------------------------------------------------------------------------------------> generate SIGBUS with siginfo as BUS_ADRERR (Non-existent physical address), I traced back into the kernel code to get the detail on why this BUS_ADRERR is set and found that, if the kernel could not handle pagefault it generates BUS_ADRERR. In the file /usr/src/linux-2.4.21-20.EL/mm/memory.c at line 1746. <-------------------------------------------------------------------> new_page = vma->vm_ops->nopage(vma, address & PAGE_MASK, 0); if (new_page == NULL) /* no page was available -- SIGBUS */ return 0; if (new_page == NOPAGE_OOM) return -1; <-------------------------------------------------------------------> So is the kernel trying to get a page (map a page) into the swap space from file, and failing to map the file ? due to lack of swap space??. Currently the kernel runs on a amd64 machine with 5GB RAM and 33GB swap space. But if I run the same program(executable) on amd64 machine with kernel 2.4.21-4 with 8GB RAM and 16GB swap space, there were no page faults and I see only 1 virtual address mapping for /lib64/ld-2.3.2.so on this machine. I analyzed the strace on the machine failing with SIGBUS and found the following in the strace output <--------------------------------------------------------------------> mmap(NULL, 1061968, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) = 0x2a95687000 mprotect(0x2a9568b000, 1045584, PROT_NONE) = 0 mmap(0x2a95787000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x2a95787000 close(3) = 0 mprotect(0x7fbff09000, 4096, PROT_READ|PROT_WRITE|PROT_EXEC|PROT_GROWSDOWN) = -1 EINVAL (Invalid argument) mprotect(0x7fbff02000, 32768, PROT_READ|PROT_WRITE|PROT_EXEC) = -1 ENOMEM (Cannot allocate memory) mprotect(0x7fbff06000, 16384, PROT_READ|PROT_WRITE|PROT_EXEC) = -1 ENOMEM (Cannot allocate memory) mprotect(0x7fbff08000, 8192, PROT_READ|PROT_WRITE|PROT_EXEC) = 0 mprotect(0x7fbff06000, 8192, PROT_READ|PROT_WRITE|PROT_EXEC) = -1 ENOMEM (Cannot allocate memory) mprotect(0x7fbff07000, 4096, PROT_READ|PROT_WRITE|PROT_EXEC) = -1 ENOMEM (Cannot allocate memory) <--------------------------------------------------------------------> I see some system calls failing with 'Cannot allocate memory', is this because that I use only 5GB of RAM for a 64-bit machine? Can some mm hackers kindly give some inputs in this? Thanks in advance. Best, Vamsi kundeti. From jay at scherrer.com Mon Jun 6 03:39:18 2005 From: jay at scherrer.com (Jay Scherrer) Date: Sun, 05 Jun 2005 20:39:18 -0700 Subject: were's the source? Message-ID: <1118029158.5052.6.camel@gimly> List, I've always took this for granted, but were are the source files for FC3 x86_64? I usually would find them in the /usr/src/redhat/linux*** directory with Redhat or Fedora core 1, but they are not there now. My initial install was "everything". Thanks, Jay Scherrer From berryja at gmail.com Mon Jun 6 04:14:03 2005 From: berryja at gmail.com (Jonathan Berry) Date: Sun, 5 Jun 2005 23:14:03 -0500 Subject: were's the source? In-Reply-To: <1118029158.5052.6.camel@gimly> References: <1118029158.5052.6.camel@gimly> Message-ID: <8767947e0506052114384c6cd2@mail.gmail.com> On 6/5/05, Jay Scherrer wrote: > List, > I've always took this for granted, but were are the source files for FC3 > x86_64? I usually would find them in the /usr/src/redhat/linux*** > directory with Redhat or Fedora core 1, but they are not there now. My > initial install was "everything". > > Thanks, > Jay Scherrer http://fedora.redhat.com/docs/release-notes/fc3/x86_64/ Look for the kernel section in the Package-Specific Notes section. Jonathan From yusufg at outblaze.com Fri Jun 10 02:18:06 2005 From: yusufg at outblaze.com (Yusuf Goolamabbas) Date: Fri, 10 Jun 2005 10:18:06 +0800 Subject: Mysql Opteron HOWTO Message-ID: <20050610021806.GB28178@outblaze.com> Came across this. Maybe of use to someone or somebody from RH can edit the wiki and clarify various issues raised. Might stop some urban legends http://hashmysql.org/index.php?title=Opteron_HOWTO -- Yusuf Goolamabbas yusufg at outblaze.com From ravichandra_bn at yahoo.com Mon Jun 13 07:00:09 2005 From: ravichandra_bn at yahoo.com (ravi chandra) Date: Mon, 13 Jun 2005 00:00:09 -0700 (PDT) Subject: 64 bit Message-ID: <20050613070010.5116.qmail@web50710.mail.yahoo.com> how to cross compile a 64 bit kernel on 32 bit environment __________________________________ Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone. http://mobile.yahoo.com/learn/mail From sbathe at gmail.com Mon Jun 13 07:37:14 2005 From: sbathe at gmail.com (Saurabh Bathe) Date: Mon, 13 Jun 2005 13:07:14 +0530 Subject: 64 bit In-Reply-To: <20050613070010.5116.qmail@web50710.mail.yahoo.com> References: <20050613070010.5116.qmail@web50710.mail.yahoo.com> Message-ID: <42AD37AA.9060109@gmail.com> ravi chandra wrote: >how to cross compile a 64 bit kernel on 32 bit environment > > > >__________________________________ >Yahoo! Mail Mobile >Take Yahoo! Mail with you! Check email on your mobile phone. >http://mobile.yahoo.com/learn/mail > > > Hi, This is not impossible but quite difficult to achieve. I have never done this myself, but what I know is that you need to recompile the complete gcc-toolchain so that it recognizes and compiles 64 bit. (this includes glibc and gcc and possibly kernel as well) Please correct me if I am wrong though. Infact the reverse is quite possible and simple. If possible I would prefer to have a 64 bit machine (AMD64 or Intel EM64T) which can do 32 bit compile easily just by installing the proper devel libs. Regards, Saurabh From vamsi.krishnak at gmail.com Mon Jun 13 07:41:47 2005 From: vamsi.krishnak at gmail.com (vamsi krishna) Date: Mon, 13 Jun 2005 13:11:47 +0530 Subject: 64 bit In-Reply-To: <42AD37AA.9060109@gmail.com> References: <20050613070010.5116.qmail@web50710.mail.yahoo.com> <42AD37AA.9060109@gmail.com> Message-ID: <3faf05680506130041939fd16@mail.gmail.com> Hello, Add the following flags for 32-bit compilation on amd64 bit machine. gcc: add option -m32 ld: add -melf_i386 I dont think you need to recompile any gcc/gnu software. Thanks Vamsi kundeti On 6/13/05, Saurabh Bathe wrote: > ravi chandra wrote: > > >how to cross compile a 64 bit kernel on 32 bit environment > > > > > > > >__________________________________ > >Yahoo! Mail Mobile > >Take Yahoo! Mail with you! Check email on your mobile phone. > >http://mobile.yahoo.com/learn/mail > > > > > > > Hi, > > This is not impossible but quite difficult to achieve. > I have never done this myself, but what I know is that you need to > recompile the complete gcc-toolchain so that it recognizes and compiles > 64 bit. (this includes glibc and gcc and possibly kernel as well) > Please correct me if I am wrong though. > > Infact the reverse is quite possible and simple. If possible I would > prefer to have a 64 bit machine (AMD64 or Intel EM64T) which can do 32 > bit compile easily just by installing the proper devel libs. > > Regards, > Saurabh > > -- > amd64-list mailing list > amd64-list at redhat.com > https://www.redhat.com/mailman/listinfo/amd64-list > From sbathe at gmail.com Mon Jun 13 08:30:22 2005 From: sbathe at gmail.com (Saurabh Bathe) Date: Mon, 13 Jun 2005 14:00:22 +0530 Subject: 64 bit In-Reply-To: <3faf05680506130041939fd16@mail.gmail.com> References: <20050613070010.5116.qmail@web50710.mail.yahoo.com> <42AD37AA.9060109@gmail.com> <3faf05680506130041939fd16@mail.gmail.com> Message-ID: <42AD441E.4010502@gmail.com> vamsi krishna wrote: >Hello, > >Add the following flags for 32-bit compilation on amd64 bit machine. > >gcc: add option -m32 >ld: add -melf_i386 > >I dont think you need to recompile any gcc/gnu software. > >Thanks >Vamsi kundeti > >On 6/13/05, Saurabh Bathe wrote: > > >>ravi chandra wrote: >> >> >> >>>how to cross compile a 64 bit kernel on 32 bit environment >>> >>> >>> >>>__________________________________ >>>Yahoo! Mail Mobile >>>Take Yahoo! Mail with you! Check email on your mobile phone. >>>http://mobile.yahoo.com/learn/mail >>> >>> >>> >>> >>> >>Hi, >> >>This is not impossible but quite difficult to achieve. >>I have never done this myself, but what I know is that you need to >>recompile the complete gcc-toolchain so that it recognizes and compiles >>64 bit. (this includes glibc and gcc and possibly kernel as well) >>Please correct me if I am wrong though. >> >>Infact the reverse is quite possible and simple. If possible I would >>prefer to have a 64 bit machine (AMD64 or Intel EM64T) which can do 32 >>bit compile easily just by installing the proper devel libs. >> >>Regards, >>Saurabh >> >>-- >>amd64-list mailing list >>amd64-list at redhat.com >>https://www.redhat.com/mailman/listinfo/amd64-list >> >> >> > > > Hi, I think the question was about compiling the kernel to run on 64 bit platform using a machine that is running on a 32 bit platform. Recompilation of gcc toolkit is required only for compiling 64 bit binary on 32 bit platform and not for the reverse. Regards, Saurabh From vamsi.krishnak at gmail.com Mon Jun 13 10:45:50 2005 From: vamsi.krishnak at gmail.com (vamsi krishna) Date: Mon, 13 Jun 2005 16:15:50 +0530 Subject: 64 bit In-Reply-To: <42AD441E.4010502@gmail.com> References: <20050613070010.5116.qmail@web50710.mail.yahoo.com> <42AD37AA.9060109@gmail.com> <3faf05680506130041939fd16@mail.gmail.com> <42AD441E.4010502@gmail.com> Message-ID: <3faf0568050613034533bde131@mail.gmail.com> Hello, gcc also has an option of -m64 which will generate 64-bit object file on 32-bit machines, but the feature stands unimplemented till gcc version 3.3.2. To check if your gcc support generation of 64-bit output. try gcc -m64 test.c, if this is a success then you can get a 64-bit kernel on a 32-bit machine. Try with latest gcc version on http://gcc.gnu.org and see if -m64 option is implemented. If not then as saurabh told you may need to complie the gnu tools. Or if you have time hack gcc source and force that implementation. Best, Vamsi kundeti On 6/13/05, Saurabh Bathe wrote: > vamsi krishna wrote: > > >Hello, > > > >Add the following flags for 32-bit compilation on amd64 bit machine. > > > >gcc: add option -m32 > >ld: add -melf_i386 > > > >I dont think you need to recompile any gcc/gnu software. > > > >Thanks > >Vamsi kundeti > > > >On 6/13/05, Saurabh Bathe wrote: > > > > > >>ravi chandra wrote: > >> > >> > >> > >>>how to cross compile a 64 bit kernel on 32 bit environment > >>> > >>> > >>> > >>>__________________________________ > >>>Yahoo! Mail Mobile > >>>Take Yahoo! Mail with you! Check email on your mobile phone. > >>>http://mobile.yahoo.com/learn/mail > >>> > >>> > >>> > >>> > >>> > >>Hi, > >> > >>This is not impossible but quite difficult to achieve. > >>I have never done this myself, but what I know is that you need to > >>recompile the complete gcc-toolchain so that it recognizes and compiles > >>64 bit. (this includes glibc and gcc and possibly kernel as well) > >>Please correct me if I am wrong though. > >> > >>Infact the reverse is quite possible and simple. If possible I would > >>prefer to have a 64 bit machine (AMD64 or Intel EM64T) which can do 32 > >>bit compile easily just by installing the proper devel libs. > >> > >>Regards, > >>Saurabh > >> > >>-- > >>amd64-list mailing list > >>amd64-list at redhat.com > >>https://www.redhat.com/mailman/listinfo/amd64-list > >> > >> > >> > > > > > > > Hi, > > I think the question was about compiling the kernel to run on 64 bit > platform using a machine that is running on a 32 bit platform. > Recompilation of gcc toolkit is required only for compiling 64 bit > binary on 32 bit platform and not for the reverse. > > Regards, > Saurabh > From ravichandra_bn at yahoo.com Wed Jun 22 08:37:23 2005 From: ravichandra_bn at yahoo.com (ravi chandra) Date: Wed, 22 Jun 2005 01:37:23 -0700 (PDT) Subject: wireless_seq_show crashes Message-ID: <20050622083723.41623.qmail@web50704.mail.yahoo.com> I am porting a 32 bit driver to redhat 2.6.9-EL 64 bit kernel. when i type iwconfig on 64 bit kernel, the kernel displays message like unable derefernce null pointer, (wireless_show_seq+58)...... The iwconfig version is v16. ____________________________________________________ Yahoo! Sports Rekindle the Rivalries. Sign up for Fantasy Football http://football.fantasysports.yahoo.com From arjanv at redhat.com Sun Jun 26 10:08:24 2005 From: arjanv at redhat.com (Arjan van de Ven) Date: Sun, 26 Jun 2005 12:08:24 +0200 Subject: wireless_seq_show crashes In-Reply-To: <20050622083723.41623.qmail@web50704.mail.yahoo.com> References: <20050622083723.41623.qmail@web50704.mail.yahoo.com> Message-ID: <1119780504.3215.21.camel@laptopd505.fenrus.org> On Wed, 2005-06-22 at 01:37 -0700, ravi chandra wrote: > I am porting a 32 bit driver to redhat 2.6.9-EL > 64 bit kernel. > when i type iwconfig on 64 bit kernel, the kernel > displays message > like unable derefernce null pointer, > (wireless_show_seq+58)...... > > The iwconfig version is v16. you forgot to attach a tarbal of your sourcecode, or provide a pointer to it -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: From cfzeitler at yahoo.com Mon Jun 27 02:08:16 2005 From: cfzeitler at yahoo.com (charles f. zeitler) Date: Sun, 26 Jun 2005 19:08:16 -0700 (PDT) Subject: motherboards for fedora core Message-ID: <20050627020816.38981.qmail@web80104.mail.yahoo.com> the following are 939 based, dual channel, up to 4 gb, agp 8x (w/ one exception) left justified have been reported as works-with fedora. indented have the same or nearly same chipsets as the left justified above. ASUS A8V Deluxe ASUS A8V MSI K8N Neo2 Platinum GIGABYTE GA K8NS Ultra 939 GA K8NS 939 GA K8NSNXP 939 GA K8NSC 939 GA K8NS Pro 939 GA K8V Ultra 939 the following have the same specs, but i found no reports on them: ABIT AV8 AV8 3rd eye (these have the same chipset) MSI RS480M2-IL (built-in video, no agp slot) MSI K8T Neo2-F GIGABYTE GA K8U 939 thats it for now, thanks for any help. cfzeitler : Do What Thou Wilt : : Shall Be : : The Whole of The Law : From cfzeitler at yahoo.com Mon Jun 27 02:20:44 2005 From: cfzeitler at yahoo.com (charles f. zeitler) Date: Sun, 26 Jun 2005 19:20:44 -0700 (PDT) Subject: Fwd: motherboards for fedora core errata Message-ID: <20050627022045.40668.qmail@web80104.mail.yahoo.com> --- "charles f. zeitler" wrote: > Date: Sun, 26 Jun 2005 19:08:16 -0700 (PDT) > From: "charles f. zeitler" > Subject: motherboards for fedora core > To: amd_64 > > the following are 939 based, > dual channel, up to 4 gb, > agp 8x (w/ one exception) > > left justified have been > reported as works-with fedora. > indented have the same > or nearly same chipsets > as the left justified above. > > ASUS A8V Deluxe > ASUS A8V <-----where did this come from? > > MSI K8N Neo2 Platinum > > GIGABYTE GA K8NS Ultra 939 > GA K8NS 939 > GA K8NSNXP 939 > GA K8NSC 939 > GA K8NS Pro 939 > GA K8V Ultra 939 > > > the following have the same specs, > but i found no reports on them: > > ABIT AV8 > AV8 3rd eye > (these have the same chipset) > > MSI RS480M2-IL > (built-in video, no agp slot) > MSI K8T Neo2-F > > GIGABYTE GA K8U 939 > > thats it for now, > thanks for any help. > > cfzeitler > > > > > : Do What Thou Wilt > : > > : Shall Be > : > > : The Whole of The Law > : > : Do What Thou Wilt : : Shall Be : : The Whole of The Law : From cochranb at speakeasy.net Mon Jun 27 03:08:20 2005 From: cochranb at speakeasy.net (Robert L Cochran) Date: Sun, 26 Jun 2005 23:08:20 -0400 Subject: motherboards for fedora core In-Reply-To: <20050627020816.38981.qmail@web80104.mail.yahoo.com> References: <20050627020816.38981.qmail@web80104.mail.yahoo.com> Message-ID: <42BF6DA4.9090005@speakeasy.net> I'm starting to think the real question now is whether Fedora Core 4 in the x86_64 arch can take advantage of the dual-core Athlon X2 processors that are available right now. And I'm hoping that MSI will soon release a BIOS that supports the X2. I want to upgrade my K8N Neo2 Platinum to the X2. Bob Cochran charles f. zeitler wrote: >the following are 939 based, >dual channel, up to 4 gb, >agp 8x (w/ one exception) > >left justified have been >reported as works-with fedora. >indented have the same >or nearly same chipsets >as the left justified above. > >ASUS A8V Deluxe > ASUS A8V > >MSI K8N Neo2 Platinum > >GIGABYTE GA K8NS Ultra 939 > GA K8NS 939 > GA K8NSNXP 939 > GA K8NSC 939 > GA K8NS Pro 939 > GA K8V Ultra 939 > > >the following have the same specs, >but i found no reports on them: > >ABIT AV8 > AV8 3rd eye >(these have the same chipset) > >MSI RS480M2-IL >(built-in video, no agp slot) >MSI K8T Neo2-F > >GIGABYTE GA K8U 939 > >thats it for now, >thanks for any help. > >cfzeitler > > > > >: Do What Thou Wilt : > >: Shall Be : > >: The Whole of The Law : > > >