From owner-linux-arm-kernel@lists.arm.linux.org.uk Wed Dec 1 18:47:26 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id SAA05493 for linux-arm-kernel-outgoing; Wed, 1 Dec 1999 18:47:26 GMT Received: from xanadu.vipswitch.com (generic199.197.205.205.IN-ADDR.ARPA [205.205.197.199] (may be forged)) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id SAA05483 for ; Wed, 1 Dec 1999 18:47:15 GMT Date: Wed, 1 Dec 1999 12:24:47 -0500 (EST) From: Nicolas Pitre To: rmk@arm.linux.org.uk cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: fix for serial boot in bios-1.06 Message-ID: MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="-842151995-1273825200-944068880=:866" Content-ID: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. Send mail to mime@docserver.cac.washington.edu for more info. ---842151995-1273825200-944068880=:866 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII Content-ID: Unfortunately, the serial boot is subtily broken in bios-1.06. Here is a patch in attachment that fixes the problem. Enjoy! Nicolas ---842151995-1273825200-944068880=:866 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII; NAME="bios-1.06.diff" Content-Transfer-Encoding: BASE64 Content-ID: Content-Description: Content-Disposition: ATTACHMENT; FILENAME="bios-1.06.diff" ZGlmZiAtdXJOIDEuMDYvYmlvcy9kcml2ZXJzL2NoYXIvcnkuYyBiaW9zL2Ry aXZlcnMvY2hhci9yeS5jDQotLS0gMS4wNi9iaW9zL2RyaXZlcnMvY2hhci9y eS5jCVN1biBOb3YgMjggMTY6MTY6NTEgMTk5OQ0KKysrIGJpb3MvZHJpdmVy cy9jaGFyL3J5LmMJV2VkIERlYyAgMSAxMjowNzozMSAxOTk5DQpAQCAtNDEs NyArNDEsNyBAQA0KIA0KIHN0YXRpYyB2b2lkIHR4KCBpbnQgYyApDQogew0K LSAgd2hpbGUgKCooKHZvbGF0aWxlIGludCAqKShTRVJCQVNFICsgMHgxOCkp ICYgOCk7DQorICB3aGlsZSAoKigodm9sYXRpbGUgaW50ICopKFNFUkJBU0Ug KyAweDE4KSkgJiAoMTw8NSkpOw0KICAgKigodm9sYXRpbGUgaW50ICopKFNF UkJBU0UpKSA9IGM7DQogfQ0KIA0KQEAgLTUxLDkgKzUxLDkgQEANCiAgIHVu c2lnbmVkIGludCBleHBpcmUgPSBjZW50aXNlY3MgKyB0aW1lb3V0ICogMTAw Ow0KIA0KICAgd2hpbGUoIGNlbnRpc2VjcyA8PSBleHBpcmUgKXsNCi0gICAg aWYgKCAhICooKHZvbGF0aWxlIHVuc2lnbmVkIGludCAqKShTRVJCQVNFICsg MHgxOCkpICYgMTYpIHsNCisgICAgaWYgKCAhICgqKCh2b2xhdGlsZSBpbnQg KikoU0VSQkFTRSArIDB4MTgpKSAmICgxPDw0KSkgKSB7DQogICAgICAgYyA9 ICooKHZvbGF0aWxlIGludCAqKShTRVJCQVNFKSk7DQotICAgICAgcmV0dXJu IGM7DQorICAgICAgcmV0dXJuIGMgJiAweGZmOw0KICAgICB9DQogICB9DQog ICByZXR1cm4gVElNRU9VVDsNCkBAIC0xOTMsMTAgKzE5MywxMSBAQA0KIHsN CiAgIGludCByZXM7DQogDQotICBwcmludGYgKCJOb3cgc2VuZCBmaWxlIHdp dGggeW1vZGVtLi4uXG4iICk7DQorICBwcmludGYoICJcbk5vdyBzZW5kIGZp bGUgd2l0aCB5LW1vZGVtICheWCB0byBhYm9ydCkgLi4uXG4iICk7DQogICBw b3V0ID0gKHVuc2lnbmVkIGxvbmcgKilsb2FkX2FkZHI7DQogICByZXMgPSB3 Y3JlY2VpdmUoKTsNCi0gIHByaW50ZiggIlRyYW5zZmVyICVzLCBwcmVzcyBh bnkga2V5LlxuIiwgDQorICB3YWl0X2NzKCA1MCApOw0KKyAgcHJpbnRmKCAi XG5UcmFuc2ZlciAlcywgcHJlc3MgYW55IGtleS5cbiIsIA0KICAgICAgICAg ICAocmVzID09IE9LKSA/ICJjb21wbGV0ZSIgOiAiZmFpbGVkIiApOw0KICAg Z2V0YygpOw0KICAgcmV0dXJuIHJlczsNCkBAIC0yMDgsNyArMjA5LDcgQEAN CiB9DQogDQogc3RydWN0IGJvb3RkZXYgYm9vdF9zZXJpYWwgPSB7DQotICAg ICAgICAic2VyaWFsIiwNCisgICAgICAgICJzZXJpYWwgcG9ydCIsDQogICAg ICAgICBzZXJpYWxfZHVtbXksDQogICAgICAgICBzZXJpYWxfZHVtbXksDQog ICAgICAgICBzZXJpYWxfbG9hZCwNCg== ---842151995-1273825200-944068880=:866-- unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Fri Dec 3 01:49:31 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id BAA05907 for linux-arm-kernel-outgoing; Fri, 3 Dec 1999 01:49:31 GMT Received: from mail12.wlv.netzero.net (mail12.wlv.netzero.net [209.247.163.62]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with SMTP id BAA05903 for ; Fri, 3 Dec 1999 01:49:28 GMT Received: (qmail 12921 invoked by uid 0); 3 Dec 1999 01:29:25 -0000 Received: from pppa43-minneapolisb22-3r5050.saturn.bbn.com (HELO netzero.net) (4.12.5.233) by mail12.wlv.netzero.net with SMTP; 3 Dec 1999 01:29:25 -0000 Message-ID: <38471D49.58801135@netzero.net> Date: Thu, 02 Dec 1999 19:30:50 -0600 From: Gaixia Zhang X-Mailer: Mozilla 4.61 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 To: linux-arm-kernel@lists.arm.linux.org.uk Subject: EBSA285: RTNL: assertion failed at devinet.c(784):inetdev_event Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Hi, I am using BIOS 1.05 to bootup Linux on EBSA285 with eepro100.c With D_Cache and Write Buffer off, I got the following msg on 2.3.18-rmk, Dec 31 19:01:14 walker kernel: eepro100.c:v1.09j 7/27/99 Donald Becker http://cesdis.gsfc.nasa.gov/linux/drivers/eepro100.html Dec 31 19:01:14 walker kernel: RTNL: assertion failed at devinet.c(784):inetdev_event Dec 31 19:01:14 walker kernel: IP-Config: Incomplete network configuration information. Dec 31 19:01:14 walker kernel: RTNL: assertion failed at devinet.c(784):inetdev_event The RTNL failed error does not show up with 2.2.12-rmk, but the system is dramatically slow, telnet always times out before I was able to input userid and passwd, devinet.c is in net/ipv4. Anybody has this problem before? Do I need to do sth in BIOS? Thanks a lot, Gaixia __________________________________________ NetZero - Defenders of the Free World Get your FREE Internet Access and Email at http://www.netzero.net/download/index.html unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Fri Dec 3 07:58:44 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id HAA14823 for linux-arm-kernel-outgoing; Fri, 3 Dec 1999 07:58:44 GMT Received: from caramon.arm.linux.org.uk (root@p50-magpie-gui.tch.enablis.net [194.168.180.50]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id HAA14819 for ; Fri, 3 Dec 1999 07:58:41 GMT Received: from flint.arm.linux.org.uk (root@flint [192.168.0.4]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id HAA27052; Fri, 3 Dec 1999 07:48:15 GMT Received: (from linux@localhost) by flint.arm.linux.org.uk (8.9.3/8.9.3) id HAA26119; Fri, 3 Dec 1999 07:47:18 GMT From: Russell King - ARM Linux Admin Message-Id: <199912030747.HAA26119@flint.arm.linux.org.uk> Subject: Re: EBSA285: RTNL: assertion failed at devinet.c(784):inetdev_event To: gaixiaz@netzero.net (Gaixia Zhang) Date: Fri, 3 Dec 1999 07:47:18 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: <38471D49.58801135@netzero.net> from "Gaixia Zhang" at Dec 2, 99 07:30:50 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Gaixia Zhang writes: > Dec 31 19:01:14 walker kernel: RTNL: assertion failed at > devinet.c(784):inetdev_event > Dec 31 19:01:14 walker kernel: IP-Config: Incomplete network > configuration > information. > Dec 31 19:01:14 walker kernel: RTNL: assertion failed at > devinet.c(784):inetdev_event > > The RTNL failed error does not show up with 2.2.12-rmk, but the system > is dramatically slow, telnet always times out before I was able to input I always get one RTNL error with 2.3.xx kernels, which doesn't have any detrimental effects, so I'd suggest ignoring that one for now. I would however suggest that you check your bootp or rarp settings - the IP-Config message suggests that something is wrong here. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Fri Dec 3 17:12:25 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id RAA12189 for linux-arm-kernel-outgoing; Fri, 3 Dec 1999 17:12:25 GMT Received: from praim.com (mail.praim.com [194.185.227.2]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id RAA12181 for ; Fri, 3 Dec 1999 17:12:07 GMT Received: from flavio_nt ([195.120.60.89]) by praim.com (Post.Office MTA v3.1.2 release (PO205-101c) ID# 167-44480U100L100S0) with SMTP id AAA199 for ; Fri, 3 Dec 1999 18:10:38 +0100 Received: by localhost with Microsoft MAPI; Fri, 3 Dec 1999 18:06:46 +0100 Message-ID: <01BF3DB9.298417B0.flaviog@praim.com> From: fabiog@praim.com (Fabio Gobber) Reply-To: "flaviog@praim.com" To: "'linux-arm-kernel@lists.arm.linux.org.uk'" Subject: Porting of Linux on an ARM CPU without MMU (ARM7TDMI). Date: Fri, 3 Dec 1999 18:06:42 +0100 Organization: Praim S.p.A. X-Mailer: Posta elettronica Internet di Microsoft/MAPI - 8.0.0.4211 Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Good day, My name is Flavio Gobber, and I am working in Praim S.p.A with the responsibility of design of embedded product. Praim, an Italian Company, designs, produces and sells innovative peripheral for the connection to the IBM's Hosts. My company has decided to develop a Printer Server that allows the connection to the Host by means of the Lan network, and at the same time the translation of printing data stream IBM towards standard printers. The first idea has been: because not to use Linux that has already all which we need. All good but we must use a CPU that costs little and that has integrated the ethernet controller. After a search of the several solutions, the best solution is to use a chip of Samsung, based on Core Arm7TDMI, that it has integrated the controller the Ethernet and cost less than 10$. Unfortunately this CPU doesn't have the MMU, that is used by the Kernel Linux standard. Panic!!!. Now I am looking for possible collaborations in order to try to know how modifying the kernel so that it can work without MMU. To help me in this in this activity I have already acquired a board of the Rt-Control, with a version of Linux "uCLinux" that works with a CPU 68000 without MMU, while for the porting of Arm I have already ordered a thin client based on Strong arm of the NetWinder. I would be much pleasing if someone could help me with some opinion on the this activity that I have decided to undertake and on possible collaboration. My company has decided to assing this activity to someone with the payment for it. I would like to know if someone is interested in this activity and the forecast of how much does it cost. Thank you very much for your cooperation. Best regards Flavio unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Fri Dec 3 23:58:36 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id XAA31402 for linux-arm-kernel-outgoing; Fri, 3 Dec 1999 23:58:36 GMT Received: from mail1.wlv.netzero.net (mail1.wlv.netzero.net [209.247.163.51]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with SMTP id XAA31396 for ; Fri, 3 Dec 1999 23:58:32 GMT Received: (qmail 17134 invoked by uid 0); 3 Dec 1999 23:58:30 -0000 Received: from pppa63-minneapolisb12-1r529.saturn.bbn.com (HELO netzero.net) (4.11.247.245) by mail1.wlv.netzero.net with SMTP; 3 Dec 1999 23:58:30 -0000 Message-ID: <3848597F.EB7BA466@netzero.net> Date: Fri, 03 Dec 1999 17:59:59 -0600 From: Gaixia Zhang X-Mailer: Mozilla 4.61 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 To: linux-arm-kernel@lists.arm.linux.org.uk, Benjamin Herrenschmidt Subject: Cache coherence: fixed sym53c8xx in the CVS? Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Hi, I remember Benjamin Herrenschmidt in an e-mail in Sep. 1999 that there is a sym53c8xx fixup for cache coherence for ARM. Where could I find more information about this fixup? Is there a website about it? Thanks a lot, Gaixia __________________________________________ NetZero - Defenders of the Free World Get your FREE Internet Access and Email at http://www.netzero.net/download/index.html unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 5 18:49:13 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id SAA27305 for linux-arm-kernel-outgoing; Sun, 5 Dec 1999 18:49:13 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id SAA27301 for ; Sun, 5 Dec 1999 18:49:10 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11ugi1-0007Ce-00 for linux-arm-kernel@lists.arm.linux.org.uk; Sun, 05 Dec 1999 18:48:33 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: linux-arm-kernel@lists.arm.linux.org.uk Subject: ebsa-285 i/o patch Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 05 Dec 1999 18:48:33 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk This patch is needed for various ISA drivers in 2.3.29. It's about time we got rid of IO_FUDGE_FACTOR in any case. p. diff -udpNr rmk/linux/include/asm-arm/arch-ebsa285/io.h linux-991205/include/asm-arm/arch-ebsa285/io.h --- rmk/linux/include/asm-arm/arch-ebsa285/io.h Wed Nov 3 19:07:36 1999 +++ linux-991205/include/asm-arm/arch-ebsa285/io.h Sun Dec 5 17:25:27 1999 @@ -55,14 +55,7 @@ extern __inline__ void __outw(unsigned i #define valid_ioaddr(iomem,size) ((iomem) < 0x80000000 && (iomem) + (size) <= 0x80000000) #define io_to_phys(iomem) ((iomem) + DC21285_PCI_MEM) -/* - * Fudge up IO addresses by this much. Once we're confident that nobody - * is using read*() and so on with addresses they didn't get from ioremap - * this can go away. - */ -#define IO_FUDGE_FACTOR PCIMEM_BASE - -#define __pci_mem_addr(x) ((void *)(IO_FUDGE_FACTOR + (unsigned long)(x))) +#define __pci_mem_addr(x) ((void *)(PCIMEM_BASE + (unsigned long)(x))) /* * ioremap takes a PCI memory address, as specified in @@ -75,8 +68,6 @@ extern __inline__ void __outw(unsigned i if (valid_ioaddr(_addr, _size)) { \ _addr = io_to_phys(_addr); \ _ret = __ioremap(_addr, _size, 0); \ - if (_ret) \ - _ret = (void *)((int) _ret - IO_FUDGE_FACTOR); \ } \ _ret; }) @@ -84,18 +75,46 @@ extern __inline__ void __outw(unsigned i #define iounmap(_addr) do { __iounmap(__pci_mem_addr((_addr))); } while (0) -#define readb(addr) (*(volatile unsigned char *)__pci_mem_addr(addr)) -#define readw(addr) (*(volatile unsigned short *)__pci_mem_addr(addr)) -#define readl(addr) (*(volatile unsigned long *)__pci_mem_addr(addr)) +#define readb(addr) (*(volatile unsigned char *)(addr)) +#define readw(addr) (*(volatile unsigned short *)(addr)) +#define readl(addr) (*(volatile unsigned long *)(addr)) +#define __raw_readb readb +#define __raw_readw readw +#define __raw_readl readl -#define writeb(b,addr) (*(volatile unsigned char *)__pci_mem_addr(addr) = (b)) -#define writew(b,addr) (*(volatile unsigned short *)__pci_mem_addr(addr) = (b)) -#define writel(b,addr) (*(volatile unsigned long *)__pci_mem_addr(addr) = (b)) +#define writeb(b,addr) (*(volatile unsigned char *)(addr) = (b)) +#define writew(b,addr) (*(volatile unsigned short *)(addr) = (b)) +#define writel(b,addr) (*(volatile unsigned long *)(addr) = (b)) +#define __raw_writeb writeb +#define __raw_writew writew +#define __raw_writel writel -#define memset_io(a,b,c) memset(__pci_mem_addr(a),(b),(c)) -#define memcpy_fromio(a,b,c) memcpy((a),__pci_mem_addr(b),(c)) -#define memcpy_toio(a,b,c) memcpy(__pci_mem_addr(a),(b),(c)) +#define memset_io(a,b,c) memset((a),(b),(c)) +#define memcpy_fromio(a,b,c) memcpy((a),(b),(c)) +#define memcpy_toio(a,b,c) memcpy((a),(b),(c)) -#define eth_io_copy_and_sum(a,b,c,d) eth_copy_and_sum((a),__pci_mem_addr(b),(c),(d)) +#define eth_io_copy_and_sum(a,b,c,d) eth_copy_and_sum((a),(b),(c),(d)) + +/* + * ISA space is 'always mapped' on a typical x86 system, no need to + * explicitly ioremap() it. The fact that the ISA IO space is mapped + * to PAGE_OFFSET is pure coincidence - it does not mean ISA values + * are physical addresses. The following constant pointer can be + * used as the IO-area pointer (it can be iounmapped as well, so the + * analogy with PCI is quite large): + */ +#define isa_readb(a) readb(__pci_mem_addr(a)) +#define isa_readw(a) readw(__pci_mem_addr(a)) +#define isa_readl(a) readl(__pci_mem_addr(a)) +#define isa_writeb(b,a) writeb(b, __pci_mem_addr(a)) +#define isa_writew(w,a) writew(w,__pci_mem_addr(a)) +#define isa_writel(l,a) writel(l,__pci_mem_addr(a)) +#define isa_memset_io(a,b,c) memset_io(__pci_mem_addr(a),(b),(c)) +#define isa_memcpy_fromio(a,b,c) memcpy_fromio((a),__pci_mem_addr(b),(c)) +#define isa_memcpy_toio(a,b,c) memcpy_toio(__pci_mem_addr(a),(b),(c)) + +#define isa_eth_io_copy_and_sum(a,b,c,d) eth_copy_and_sum((a),__pci_mem_addr(b),(c),(d)) + +#define isa_check_signature(a,b,c) check_signature(__pci_mem_addr(a),(b),(c)) #endif unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 5 19:53:21 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id TAA29416 for linux-arm-kernel-outgoing; Sun, 5 Dec 1999 19:53:21 GMT Received: from caramon.arm.linux.org.uk (root@p28-cordelia-gui.tch.enablis.net [212.250.233.28]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id TAA29412 for ; Sun, 5 Dec 1999 19:53:17 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id TAA02156; Sun, 5 Dec 1999 19:53:13 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id TAA03767; Sun, 5 Dec 1999 19:52:30 GMT Message-Id: <199912051952.TAA03767@raistlin.arm.linux.org.uk> Subject: Re: ebsa-285 i/o patch To: Philip.Blundell@pobox.com (Philip Blundell) Date: Sun, 5 Dec 1999 19:52:29 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 5, 99 06:48:33 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > This patch is needed for various ISA drivers in 2.3.29. It's about time we > got rid of IO_FUDGE_FACTOR in any case. Have you checked that the NetWinder tulip driver is still functional? I would think that this change makes the kernel Oops when you load the driver. If you send the patch system a patch for both this and the tulip, it may find its way into the next kernel release (whenever the mainstream kernel guys get themselves sorted out so that ARM kernels actually have some semblence of actually working on more than one architecture again). _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 5 20:05:33 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id UAA30089 for linux-arm-kernel-outgoing; Sun, 5 Dec 1999 20:05:33 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id UAA30082 for ; Sun, 5 Dec 1999 20:05:30 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11uhtv-0008SK-00; Sun, 05 Dec 1999 20:04:55 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: ebsa-285 i/o patch In-Reply-To: Message from Russell King - ARM Linux Admin of "Sun, 05 Dec 1999 19:52:29 GMT." <199912051952.TAA03767@raistlin.arm.linux.org.uk> References: <199912051952.TAA03767@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 05 Dec 1999 20:04:54 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >Have you checked that the NetWinder tulip driver is still functional? Yes; the version in CVS works on my CATS machine. I haven't tried a real NetWinder but I can't see any reason it should be different. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 5 20:13:13 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id UAA30655 for linux-arm-kernel-outgoing; Sun, 5 Dec 1999 20:13:13 GMT Received: from caramon.arm.linux.org.uk (root@p28-cordelia-gui.tch.enablis.net [212.250.233.28]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id UAA30651 for ; Sun, 5 Dec 1999 20:13:10 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id UAA02288; Sun, 5 Dec 1999 20:13:08 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id UAA03887; Sun, 5 Dec 1999 20:12:24 GMT Message-Id: <199912052012.UAA03887@raistlin.arm.linux.org.uk> Subject: Re: ebsa-285 i/o patch To: Philip.Blundell@pobox.com (Philip Blundell) Date: Sun, 5 Dec 1999 20:12:24 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 5, 99 08:04:54 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > Yes; the version in CVS works on my CATS machine. I haven't tried a real > NetWinder but I can't see any reason it should be different. The NetWinder's tulip is unlike the usual tulips - its registers are PCI memory mapped rather than the more normal PCI IO mapping. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 5 20:33:00 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id UAA31975 for linux-arm-kernel-outgoing; Sun, 5 Dec 1999 20:33:00 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id UAA31967 for ; Sun, 5 Dec 1999 20:32:58 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11uiKw-0000tg-00; Sun, 05 Dec 1999 20:32:50 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: ebsa-285 i/o patch In-Reply-To: Message from Russell King - ARM Linux Admin of "Sun, 05 Dec 1999 20:12:24 GMT." <199912052012.UAA03887@raistlin.arm.linux.org.uk> References: <199912052012.UAA03887@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 05 Dec 1999 20:32:50 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >The NetWinder's tulip is unlike the usual tulips - its registers are PCI >memory mapped rather than the more normal PCI IO mapping. That is controlled statically by CONFIG_ARCH_NETWINDER (which is set in my configuration) rather than being determined by the actual hardware. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 6 04:23:56 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id EAA23739 for linux-arm-kernel-outgoing; Mon, 6 Dec 1999 04:23:56 GMT Received: from xanadu.gn.com (modemcable233.137-200-24.mtl.mc.videotron.net [24.200.137.233]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id EAA23717 for ; Mon, 6 Dec 1999 04:23:52 GMT Received: from localhost (nico@localhost) by xanadu.gn.com (8.9.3/8.9.3) with ESMTP id XAA25490; Sun, 5 Dec 1999 23:18:20 -0500 X-Authentication-Warning: xanadu.gn.com: nico owned process doing -bs Date: Sun, 5 Dec 1999 23:18:19 -0500 (EST) From: Nicolas Pitre X-Sender: nico@xanadu.gn.com To: linux-arm-kernel@lists.arm.linux.org.uk cc: linux-kernel@vger.rutgers.edu Subject: bootmem and mem_map allocation problem Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk I'm currently re-integrating the StrongARM SA1100 support into 2.3.29 but here is the problem: Physical RAM banks are located at 0xc0000000, 0xc8000000, 0xd0000000 and 0xd8000000. The system I have contains 4 megs for each bank, but the SA1100 can support many other configurations as well. The bootmem allocator keep track of all physical ram pages linearly, which is still OK since the holes between all banks are marked reserved. However this represents a total of ((0xd8000000 + 4MB - 0xc0000000) >> PAGE_SHIFT) = 99328 pages total. When it is time to allocate the mem_map array which is also indexing physical pages, we have this: /* * Some architectures (with lots of mem and discontinous memory * maps) have to search for a good mem_map area: */ map_size = totalpages*sizeof(struct page); mem_map = (struct page *) alloc_bootmem(map_size); map_size is then equal to 6754304 which is larger than a single ram bank. Then, alloc_bootmem() fails since it can't find enough contiguous memory. Even if the mem_map "holes" can be freed later, the valuable entries may not fit with the available memory layout, especially if we want to keep the virt_to_phys()/phys_to_virt() macros simple. Maybe the mem_map array could be indexed with virtual page location since the virtual memory space is being compressed a bit, thus reducing the number of entries required. However this would require modifications to the free_all_bootmem() function which isn't architecture specific. What should be the correct fix for this? Also I needed this patch to make things work on a machine with a non null physical memory offset: --- ../2.3.29/linux/mm/bootmem.c Thu Dec 2 22:08:08 1999 +++ linux/mm/bootmem.c Sat Dec 4 14:17:24 1999 @@ -38,7 +38,7 @@ unsigned long mapsize = (pages+7)/8; mem_start = mstart; - bootmem_map = phys_to_virt(start << PAGE_SHIFT); + bootmem_map = phys_to_virt((start << PAGE_SHIFT) + mem_start); max_low_pfn = pages; /* @@ -217,7 +217,7 @@ count++; ClearPageReserved(page); set_page_count(page, 1); - if (i >= (virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT)) + if (i >= MAP_NR(MAX_DMA_ADDRESS) ) clear_bit(PG_DMA, &page->flags); __free_page(page); } Thanks. Nicolas unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 6 09:47:02 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id JAA02076 for linux-arm-kernel-outgoing; Mon, 6 Dec 1999 09:47:02 GMT Received: from bacchus.fh-brandenburg.de ([195.37.1.36]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with SMTP id JAA02070 for ; Mon, 6 Dec 1999 09:46:48 GMT Received: (2105 bytes) by bacchus.fh-brandenburg.de via sendmail with P:stdio/R:match-inet-hosts/T:smtp (sender: ) id for ; Mon, 6 Dec 1999 10:46:11 +0100 (MET) (Smail-3.2.0.99 1997-Dec-4 #1 built DST-Apr-16) Received: from zeus(195.37.1.35) via SMTP by bacchus, id smtpdAAAa002JU; Mon Dec 6 10:46:06 1999 Date: Mon, 6 Dec 1999 10:44:28 +0100 (MET) From: Roman Zippel To: Nicolas Pitre cc: linux-arm-kernel@lists.arm.linux.org.uk, linux-kernel@vger.rutgers.edu Subject: Re: bootmem and mem_map allocation problem In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Hi, On Sun, 5 Dec 1999, Nicolas Pitre wrote: > Even if the mem_map "holes" can be freed later, the valuable > entries may not fit with the available memory layout, especially if we > want to keep the virt_to_phys()/phys_to_virt() macros simple. I used the patch below (for 2.3.28, but should still apply, I had time yet to update yet), but it removes the requirement for an 1:1 mapping of physical memory to the memmap. Instead the whole logic is in MAP_NR(), where an abitrary logic to convert a virtual address into a map nr can be implemented. bye, Roman --- linux-2.3.28.org/mm/bootmem.c Mon Nov 15 15:03:35 1999 +++ linux-2.3.28/mm/bootmem.c Mon Nov 15 19:07:35 1999 @@ -203,18 +203,19 @@ unsigned long __init free_all_bootmem (void) { struct page * page; - unsigned long i, count, total = 0; + unsigned long i, nr, count, total = 0; if (!bootmem_map) BUG(); - page = mem_map; count = 0; - for (i = 0; i < max_low_pfn; i++, page++) { + for (i = 0; i < max_low_pfn; i++) { if (!test_bit(i, bootmem_map)) { + nr = MAP_NR(phys_to_virt(i << PAGE_SHIFT)); + page = mem_map + nr; count++; ClearPageReserved(page); set_page_count(page, 1); - if (i >= (virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT)) + if (nr >= (virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT)) clear_bit(PG_DMA, &page->flags); __free_page(page); } unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 6 19:15:58 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id TAA29558 for linux-arm-kernel-outgoing; Mon, 6 Dec 1999 19:15:58 GMT Received: from caramon.arm.linux.org.uk (root@p58-magpie-gui.tch.enablis.net [194.168.180.58]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id TAA29554 for ; Mon, 6 Dec 1999 19:15:54 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id TAA06525; Mon, 6 Dec 1999 19:14:44 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id TAA00468; Mon, 6 Dec 1999 19:14:03 GMT Message-Id: <199912061914.TAA00468@raistlin.arm.linux.org.uk> Subject: Re: bootmem and mem_map allocation problem To: nico@CAM.ORG (Nicolas Pitre) Date: Mon, 6 Dec 1999 19:14:02 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk, linux-kernel@vger.rutgers.edu In-Reply-To: from "Nicolas Pitre" at Dec 5, 99 11:18:19 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Nicolas Pitre writes: > I'm currently re-integrating the StrongARM SA1100 support into 2.3.29 but > here is the problem: As I said quite prominently in the release notes for 2.3.29 for ARM, this is expected at the moment: > The 2.3.29 kernel patch is now on the FTP site. I can say as a certainty > that it will NOT work on any architecture which has a non-zero physical > memory offset. Depending on which patches get into Linus' kernel will > depend on when this becomes available again. > ... > Again, please note that this kernel will not work on some of the ARM > architectures (RiscPC for example). There are fixes in the pipeline > already. Ingo has a fix for this, and I am currently waiting for him to integrate it into the kernel; there is no point fixing something twice when it has already been fixed. Ingo's fix works well, and have tried to encourage him to forward them to Linus, but have not received any response to date. Unfortunately, Ingo's patch affects a number of files, which make it difficult to manage when his patch goes into the mainstream. If you're interested, I could redistribute the patch separately. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 6 21:02:03 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA02408 for linux-arm-kernel-outgoing; Mon, 6 Dec 1999 21:02:03 GMT Received: from xanadu.vipswitch.com (generic199.197.205.205.in-addr.arpa [205.205.197.199] (may be forged)) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA02402 for ; Mon, 6 Dec 1999 21:01:58 GMT Message-Id: <199912062059.PAA15574@xanadu.vipswitch.com> Date: Mon, 6 Dec 1999 14:50:39 -0500 (EST) From: Nicolas Pitre To: Russell King - ARM Linux Admin Cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: bootmem and mem_map allocation problem Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk On Mon, 6 Dec 1999, Russell King - ARM Linux Admin wrote: > Nicolas Pitre writes: > > I'm currently re-integrating the StrongARM SA1100 support into 2.3.29 but > > here is the problem: [...] > Unfortunately, Ingo's patch affects a number of files, which make it difficult > to manage when his patch goes into the mainstream. If you're interested, I > could redistribute the patch separately. Please do so. I was probing for an existing solution before I do my own. If there is a patch available, then I'll resume to the SA1100 specifics. Nicolas unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Wed Dec 8 00:28:04 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id AAA03722 for linux-arm-kernel-outgoing; Wed, 8 Dec 1999 00:28:04 GMT Received: from caramon.arm.linux.org.uk (root@p23-cordelia-gui.tch.enablis.net [212.250.233.23]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id AAA03718 for ; Wed, 8 Dec 1999 00:28:00 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id AAA12076; Wed, 8 Dec 1999 00:27:56 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id AAA01816; Wed, 8 Dec 1999 00:27:14 GMT Message-Id: <199912080027.AAA01816@raistlin.arm.linux.org.uk> Subject: Re: bootmem and mem_map allocation problem To: nico@CAM.ORG (Nicolas Pitre) Date: Wed, 8 Dec 1999 00:27:14 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: <199912062059.PAA15574@xanadu.vipswitch.com> from "Nicolas Pitre" at Dec 6, 99 02:50:39 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Ok, 2.3.30 is now out from Linus, and I'm now wondering why the hell I'm re-writing the boot-time memory initialisation on each and every kernel release since 2.3.23. It looks like yet another re-write for 2.3.30 is on the cards again. To this end, I shall be temporarily halting work on the 2.3 series kernel until things stabilise in this critical area again. Appologies for any inconvience. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 19:44:34 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id TAA31678 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 19:44:34 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id TAA31671 for ; Thu, 9 Dec 1999 19:44:24 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11w9SA-0001b0-00 for linux-arm-kernel@lists.arm.linux.org.uk; Thu, 09 Dec 1999 19:42:14 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: linux-arm-kernel@lists.arm.linux.org.uk Subject: memory mapping oddity Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 19:42:14 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk A while ago I noticed that under some circumstances NWFPE would generate an oops while trying to load an instruction from user memory. Apparently the undefined instruction trap was being called with LR pointing to a bad address. At the time I chose to ignore it and just added an exception wrapper to NWFPE to catch the bad access -- this shuts up the error messages but doesn't really address the problem. It now seems that the same thing can happen with the SWI vector. Again the effect is that the load of the SWI instruction aborts with a message like this: Unable to handle kernel paging request at virtual address 6776e2ac pgd = c2628000 *pgd = 00000000, *pmd = 00000000 Internal error: Oops: 2 CPU: 0 pc : [] lr : [<6776e2b0>] sp : c32f3fb8 ip : bffff9c4 fp : 00000000 r10: 4010188c r9 : 0200a3dc r8 : c32f3ff4 r7 : 0200a3f0 r6 : 20000010 r5 : 0200a3dc r4 : 00000015 r3 : 0200a3e4 r2 : 00000100 r1 : bffff804 r0 : 0080493c Flags: nzCv IRQs off FIQs on Mode SVC_32 Segment user Control: 262917D Table: 0262917D DAC: 00000015 The code in question is: c00324c0 : c00324c0: e24dd048 sub sp, sp, #72 c00324c4: e88d1fff stmia sp, {r0, r1, r2, r3, r4, r5, r6, r7, r8, r9, sl, fp, ip} c00324c8: e28d803c add r8, sp, #60 c00324cc: e9486000 stmdb r8, {sp, lr}^ c00324d0: e588e000 str lr, [r8] c00324d4: e14f6000 mrs r6, spsr c00324d8: e5886004 str r6, [r8, #4] c00324dc: e5880008 str r0, [r8, #8] c00324e0: e3a0b000 mov fp, #0 c00324e4: e51e6004 ldr r6, [lr, -#4] It seems hard to imagine what might cause this. I can just about conceive of some TLB synchronisation issue where a page has been unmapped in the D-TLB but not the I-TLB (because it's not marked as VM_EXEC), but I'm unconvinced. Does anyone have any thoughts? Again we could add an exception wrapper to stop the kernel actually coming to grief but I am not sure this is the right thing to do. Incidentally, someone changed the first instruction fetch in NWFPE to be an LDRT rather than a plain LDR; if this was actually necessary (I'm not sure it was) then perhaps the SWI handler should do the same. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 20:37:55 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id UAA02270 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 20:37:55 GMT Received: from caramon.arm.linux.org.uk (root@p42-magpie-gui.tch.enablis.net [194.168.180.42]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id UAA02265 for ; Thu, 9 Dec 1999 20:37:51 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id UAA20450; Thu, 9 Dec 1999 20:37:48 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id UAA00619; Thu, 9 Dec 1999 20:37:09 GMT Message-Id: <199912092037.UAA00619@raistlin.arm.linux.org.uk> Subject: Re: memory mapping oddity To: Philip.Blundell@pobox.com (Philip Blundell) Date: Thu, 9 Dec 1999 20:37:09 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 9, 99 07:42:14 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > thing to do. Incidentally, someone changed the first instruction fetch in > NWFPE to be an LDRT rather than a plain LDR; if this was actually necessary > (I'm not sure it was) then perhaps the SWI handler should do the same. No, SWIs can be called from kernel-space as well as user-space. An ldrt would fault if called from kernel-space. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:01:54 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA03373 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:01:54 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA03369 for ; Thu, 9 Dec 1999 21:01:49 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wAh7-0001p0-00; Thu, 09 Dec 1999 21:01:45 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: memory mapping oddity In-Reply-To: Message from Russell King - ARM Linux Admin of "Thu, 09 Dec 1999 20:37:09 GMT." <199912092037.UAA00619@raistlin.arm.linux.org.uk> References: <199912092037.UAA00619@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:01:45 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >No, SWIs can be called from kernel-space as well as user-space. An ldrt would >fault if called from kernel-space. Can they still? I thought all, or almost all, of the instances of that had been eliminated. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:07:44 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA03594 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:07:44 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA03590 for ; Thu, 9 Dec 1999 21:07:39 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wAmi-0001pw-00 for linux-arm-kernel@lists.arm.linux.org.uk; Thu, 09 Dec 1999 21:07:32 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: linux-arm-kernel@lists.arm.linux.org.uk Subject: AFLAGS lossage Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:07:32 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Despite what the name might suggest, $(AFLAGS) is only actually passed to gcc rather than the assembler. The patch below (for 2.3.29) changes this variable to contain flags that the compiler actually understands. This gives a measurable speedup on my machine. (I've only tested with egcs-1.1.1 and binutils 2.9.4.) p. Index: arch/arm/Makefile =================================================================== RCS file: /cvsroot/pub/kernel/armlinux/arch/arm/Makefile,v retrieving revision 1.20 retrieving revision 1.21 diff -u -r1.20 -r1.21 --- Makefile 1999/12/05 17:13:48 1.20 +++ Makefile 1999/12/09 21:03:35 1.21 @@ -83,10 +82,10 @@ TEXTADDR = 0xC0008000 ifeq ($(CONFIG_CPU_32v4),y) CFLAGS += $(CFLAGS_PROC_CPU_32v4) - AFLAGS += -mapcs-32 -marmv4 + AFLAGS += $(CFLAGS_PROC_CPU_32v4) else CFLAGS += $(CFLAGS_PROC_CPU_32v3) - AFLAGS += -mapcs-32 -marmv3m + AFLAGS += $(CFLAGS_PROC_CPU_32v3) endif # # Exactly one of the following must be selected unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:12:00 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA03787 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:12:00 GMT Received: from caramon.arm.linux.org.uk (root@p42-magpie-gui.tch.enablis.net [194.168.180.42]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA03783 for ; Thu, 9 Dec 1999 21:11:57 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id VAA20645; Thu, 9 Dec 1999 21:11:54 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id VAA00734; Thu, 9 Dec 1999 21:11:15 GMT Message-Id: <199912092111.VAA00734@raistlin.arm.linux.org.uk> Subject: Re: AFLAGS lossage To: Philip.Blundell@pobox.com (Philip Blundell) Date: Thu, 9 Dec 1999 21:11:14 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 9, 99 09:07:32 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > Despite what the name might suggest, $(AFLAGS) is only actually passed to gcc > rather than the assembler. The patch below (for 2.3.29) changes this > variable to contain flags that the compiler actually understands. This > gives a measurable speedup on my machine. (I've only tested with egcs-1.1.1 > and binutils 2.9.4.) I don't really believe the performance improvement, unless egcs somehow optimises assembler code... Before: AFLAGS = -mapcs-32 -marmv4 After: AFLAGS = -march=armv4 Please explain how this can speed up assembler code. (Also, it breaks the compile with the old GCC 2.7.2.2) _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:15:09 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA03972 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:15:09 GMT Received: from caramon.arm.linux.org.uk (root@p42-magpie-gui.tch.enablis.net [194.168.180.42]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA03966 for ; Thu, 9 Dec 1999 21:15:06 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id VAA20658; Thu, 9 Dec 1999 21:15:03 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id VAA00748; Thu, 9 Dec 1999 21:14:25 GMT Message-Id: <199912092114.VAA00748@raistlin.arm.linux.org.uk> Subject: Re: memory mapping oddity To: Philip.Blundell@pobox.com (Philip Blundell) Date: Thu, 9 Dec 1999 21:14:24 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 9, 99 07:42:14 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > A while ago I noticed that under some circumstances NWFPE would generate an > oops while trying to load an instruction from user memory. Apparently the > undefined instruction trap was being called with LR pointing to a bad address. > At the time I chose to ignore it and just added an exception wrapper to NWFPE > to catch the bad access -- this shuts up the error messages but doesn't really > address the problem. > > It now seems that the same thing can happen with the SWI vector. Again the > effect is that the load of the SWI instruction aborts with a message like this: Incidentally, Rebel.com just reported similar problems with 2.2.12. My kernels are still rock stable, so I can't help here. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:26:07 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA04405 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:26:07 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA04401 for ; Thu, 9 Dec 1999 21:26:04 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wB4W-0001tC-00; Thu, 09 Dec 1999 21:25:56 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: AFLAGS lossage In-Reply-To: Message from Russell King - ARM Linux Admin of "Thu, 09 Dec 1999 21:11:14 GMT." <199912092111.VAA00734@raistlin.arm.linux.org.uk> References: <199912092111.VAA00734@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:25:56 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >Before: > >AFLAGS = -mapcs-32 -marmv4 "-marmv4" is not a legal compiler option; GCC just ignores it. >AFLAGS = -march=armv4 This causes GCC to define __ARM_ARCH_4__, which various bits of assembler use when deciding what the best code sequence to use is. Compare the syscall entry path before and after, for example. >(Also, it breaks the compile with the old GCC 2.7.2.2) That's not too much of a surprise. I'll see if I can fix it up. By the way, "-mapcs-32" in CFLAGS probably does more harm than good; I think we want to get rid of that. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:28:29 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA04436 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:28:29 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA04432 for ; Thu, 9 Dec 1999 21:28:25 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wB6s-0001tT-00; Thu, 09 Dec 1999 21:28:22 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: memory mapping oddity In-Reply-To: Message from Russell King - ARM Linux Admin of "Thu, 09 Dec 1999 21:14:24 GMT." <199912092114.VAA00748@raistlin.arm.linux.org.uk> References: <199912092114.VAA00748@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:28:22 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >Incidentally, Rebel.com just reported similar problems with 2.2.12. > >My kernels are still rock stable, so I can't help here. It's not an instability as such; I only saw two instances of the SWI problem in 24 hours of running crashme, and the undefined instruction one was similarly rare. Neither seems to do any real harm. Unless you are testing under high load I wouldn't expect you to have seen the problem. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:31:24 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA04585 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:31:24 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA04579 for ; Thu, 9 Dec 1999 21:31:21 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wB9h-0001uy-00; Thu, 09 Dec 1999 21:31:17 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: AFLAGS lossage In-Reply-To: Message from Russell King - ARM Linux Admin of "Thu, 09 Dec 1999 21:11:14 GMT." <199912092111.VAA00734@raistlin.arm.linux.org.uk> References: <199912092111.VAA00734@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:31:17 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >(Also, it breaks the compile with the old GCC 2.7.2.2) I looked at the change again and I can't see what would cause this. I guess I don't know in enough detail what the old GCC accepts. Can you explain the problem a bit more? p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:32:16 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA04602 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:32:16 GMT Received: from caramon.arm.linux.org.uk (root@p42-magpie-gui.tch.enablis.net [194.168.180.42]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA04598 for ; Thu, 9 Dec 1999 21:32:13 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id VAA20755; Thu, 9 Dec 1999 21:31:47 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id VAA00921; Thu, 9 Dec 1999 21:31:08 GMT Message-Id: <199912092131.VAA00921@raistlin.arm.linux.org.uk> Subject: Re: AFLAGS lossage To: Philip.Blundell@pobox.com (Philip Blundell) Date: Thu, 9 Dec 1999 21:31:06 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 9, 99 09:25:56 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > By the way, "-mapcs-32" in CFLAGS probably does more harm than good; I think > we want to get rid of that. That's why it's not in CFLAGS. The alternative is - can you supply me the spec line from GCC for the assembler please? _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 21:39:12 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id VAA04831 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 21:39:12 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id VAA04827 for ; Thu, 9 Dec 1999 21:39:09 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11wBHD-0001w3-00; Thu, 09 Dec 1999 21:39:03 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: AFLAGS lossage In-Reply-To: Message from Russell King - ARM Linux Admin of "Thu, 09 Dec 1999 21:31:06 GMT." <199912092131.VAA00921@raistlin.arm.linux.org.uk> References: <199912092131.VAA00921@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 09 Dec 1999 21:39:03 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >That's why it's not in CFLAGS. Sorry, you're right. >The alternative is - can you supply me the spec line from GCC for the >assembler please? It varies from version to version. Here is the one I happen to have installed locally. Why do you want to know? *asm: %{mbig-endian:-EB} %{mapcs-float:-mfloat} p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 22:06:58 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id WAA05761 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 22:06:58 GMT Received: from www.netwinder.org (IDENT:root@netwinder.org [207.245.35.202]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id WAA05757 for ; Thu, 9 Dec 1999 22:06:55 GMT Received: from gw.rebel.com (gate.rebel.com [207.245.35.200]) by www.netwinder.org (8.8.7/8.8.7) with SMTP id RAA30070 for ; Thu, 9 Dec 1999 17:06:54 -0500 Message-ID: <3850270E.9DF821EA@netwinder.org> Received: from dhcp111.isa.ott.rebel.com by gw.rebel.com via smtpd (for netwinder.org [207.245.35.202]) with SMTP; 9 Dec 1999 22:06:10 UT Date: Thu, 09 Dec 1999 17:02:54 -0500 From: Scott Bambrough Organization: Rebel.com X-Mailer: Mozilla 4.7 [en] (Win95; I) X-Accept-Language: en MIME-Version: 1.0 To: Arm Linux Kernel List Subject: sys_cacheflush... Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk I'm having trouble calling sys_cacheflush from user mode. Can someone document the parameters for me. I thought they were: r0 starting VM address r1 length r2 boolean (1 if Icache flush desired). Scott -- Scott Bambrough - Software Engineer REBEL.COM http://www.rebel.com NetWinder http://www.netwinder.org unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 9 22:11:12 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id WAA05959 for linux-arm-kernel-outgoing; Thu, 9 Dec 1999 22:11:12 GMT Received: from caramon.arm.linux.org.uk (root@p17-robin-gui.tch.enablis.net [194.168.180.77]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id WAA05955 for ; Thu, 9 Dec 1999 22:11:09 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id WAA21048; Thu, 9 Dec 1999 22:11:05 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id WAA01449; Thu, 9 Dec 1999 22:10:26 GMT Message-Id: <199912092210.WAA01449@raistlin.arm.linux.org.uk> Subject: Re: sys_cacheflush... To: scottb@netwinder.org (Scott Bambrough) Date: Thu, 9 Dec 1999 22:10:25 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: <3850270E.9DF821EA@netwinder.org> from "Scott Bambrough" at Dec 9, 99 05:02:54 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Scott Bambrough writes: > I'm having trouble calling sys_cacheflush from user mode. Can someone > document the parameters for me. I thought they were: > > r0 starting VM address > r1 length > r2 boolean (1 if Icache flush desired). Looks right to me. What is the problem? _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Fri Dec 10 14:32:13 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id OAA17293 for linux-arm-kernel-outgoing; Fri, 10 Dec 1999 14:32:13 GMT Received: from salisbury.labs.futuretv.com (salisbury.futuretv.com [194.216.164.17]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id OAA17288 for ; Fri, 10 Dec 1999 14:31:53 GMT Received: from [192.0.0.67] (helo=zebra.labs.futuretv.com ident=mail) by salisbury.labs.futuretv.com with esmtp (Exim 3.03 #1) id 11wR6Q-0005wS-00 for linux-arm-kernel@lists.arm.linux.org.uk; Fri, 10 Dec 1999 14:32:58 +0000 Received: from [192.0.0.21] (helo=fountain.labs.futuretv.com ident=mail) by zebra.labs.futuretv.com with esmtp (Exim 3.03 #1) id 11wR5B-000628-00; Fri, 10 Dec 1999 14:31:41 +0000 Received: from [127.0.0.1] (helo=fountain ident=pb) by fountain.labs.futuretv.com with esmtp (Exim 3.03 #1) id 11wR5A-0000S5-00; Fri, 10 Dec 1999 14:31:40 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: scottb@netwinder.org Cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: sys_cacheflush... From: Philip Blundell Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 10 Dec 1999 14:31:40 +0000 Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Scott Bambrough wrote: > I'm having trouble calling sys_cacheflush from user mode. Can someone > document the parameters for me. I thought they were: > r0 starting VM address r1 length r2 boolean (1 if > Icache flush desired). I think R1 is the end address, not the length. The meaning for R2 has never been documented; GCC passes zero in that register and expects the Icache to be flushed. At the moment the kernel always flushes both caches and ignores R2 altogether. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 12 10:18:06 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id KAA12592 for linux-arm-kernel-outgoing; Sun, 12 Dec 1999 10:18:06 GMT Received: from caramon.arm.linux.org.uk (root@p06-robin-gui.tch.enablis.net [194.168.180.66]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id KAA12584 for ; Sun, 12 Dec 1999 10:17:59 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id KAA31852; Sun, 12 Dec 1999 10:17:42 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id KAA01590; Sun, 12 Dec 1999 10:17:04 GMT Message-Id: <199912121017.KAA01590@raistlin.arm.linux.org.uk> Subject: Re: ebsa-285 i/o patch To: Philip.Blundell@pobox.com (Philip Blundell) Date: Sun, 12 Dec 1999 10:17:03 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 5, 99 08:32:50 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > >The NetWinder's tulip is unlike the usual tulips - its registers are PCI > >memory mapped rather than the more normal PCI IO mapping. > > That is controlled statically by CONFIG_ARCH_NETWINDER (which is set in my > configuration) rather than being determined by the actual hardware. Err, so? I've just been looking into the change, and it will break the tulip driver (since the BASE_1 register is used directly in the readl macros). Unfortunately, the general architecture of the tulip driver does not lend itself to ioremap. If you'd like to resubmit the patch with the tulip driver fixed, I'll drop it into the 2.3 stuff. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 12 10:19:19 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id KAA12607 for linux-arm-kernel-outgoing; Sun, 12 Dec 1999 10:19:19 GMT Received: from caramon.arm.linux.org.uk (root@p06-robin-gui.tch.enablis.net [194.168.180.66]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id KAA12603 for ; Sun, 12 Dec 1999 10:19:16 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id KAA31796; Sun, 12 Dec 1999 10:08:00 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id KAA01545; Sun, 12 Dec 1999 10:07:22 GMT Message-Id: <199912121007.KAA01545@raistlin.arm.linux.org.uk> Subject: Re: memory mapping oddity To: Philip.Blundell@pobox.com (Philip Blundell) Date: Sun, 12 Dec 1999 10:07:21 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 9, 99 09:01:45 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > >No, SWIs can be called from kernel-space as well as user-space. An ldrt would > >fault if called from kernel-space. > > Can they still? I thought all, or almost all, of the instances of that > had been eliminated. A lot of the SWI calls in the kernel have been eliminated, but there are certain calls (eg, clone) which still need the SWI register stack layout and return structure to work. IIRC, clone is the only instance where this is used. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 12 11:32:26 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id LAA15191 for linux-arm-kernel-outgoing; Sun, 12 Dec 1999 11:32:26 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id LAA15187 for ; Sun, 12 Dec 1999 11:32:23 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11x7EZ-0000S6-00; Sun, 12 Dec 1999 11:32:11 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: ebsa-285 i/o patch In-Reply-To: Message from Russell King - ARM Linux Admin of "Sun, 12 Dec 1999 10:17:03 GMT." <199912121017.KAA01590@raistlin.arm.linux.org.uk> References: <199912121017.KAA01590@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 12 Dec 1999 11:32:11 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Russell King wrote: >Philip Blundell writes: >> >The NetWinder's tulip is unlike the usual tulips - its registers are PCI >> >memory mapped rather than the more normal PCI IO mapping. >> >> That is controlled statically by CONFIG_ARCH_NETWINDER (which is set in my >> configuration) rather than being determined by the actual hardware. > >Err, so? So it should make no difference whether I run it on an actual NetWinder or not. A driver compiled with CONFIG_ARCH_NETWINDER set should work the same on all machines. As I said, such a driver works fine on my CATS. >I've just been looking into the change, and it will break the >tulip driver (since the BASE_1 register is used directly in the readl >macros). I don't see this. Please be more specific about the problem. If you are seeing a crash (which it sounds like you are), what does the Oops message say? Are you sure you are looking at the same version of the tulip driver as me? This one has: #ifdef CONFIG_ARCH_NETWINDER #undef inl #undef outl #define inl(p) readl((p)) #define outl(v,p) writel((v),(p)) #define tulip_resource iomem_resource #else #define tulip_resource ioport_resource #endif and it does seem to use ioremap to get the base address. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 12 11:36:13 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id LAA15343 for linux-arm-kernel-outgoing; Sun, 12 Dec 1999 11:36:13 GMT Received: from caramon.arm.linux.org.uk (root@p57-magpie-gui.tch.enablis.net [194.168.180.57]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id LAA15338 for ; Sun, 12 Dec 1999 11:36:09 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id LAA32242; Sun, 12 Dec 1999 11:36:05 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id LAA02001; Sun, 12 Dec 1999 11:35:26 GMT Message-Id: <199912121135.LAA02001@raistlin.arm.linux.org.uk> Subject: Re: ebsa-285 i/o patch To: Philip.Blundell@pobox.com (Philip Blundell) Date: Sun, 12 Dec 1999 11:35:25 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 12, 99 11:32:11 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > and it does seem to use ioremap to get the base address. Ah, it does? In that case, it probably doesn't iounmap it actually, which means it'll waste memory if it's inserted and removed many times. (I forgot that I fixed it ;(). _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Sun Dec 12 12:03:53 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id MAA16088 for linux-arm-kernel-outgoing; Sun, 12 Dec 1999 12:03:53 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id MAA16084 for ; Sun, 12 Dec 1999 12:03:46 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11x7j0-0000Ul-00; Sun, 12 Dec 1999 12:03:38 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: ebsa-285 i/o patch In-Reply-To: Message from Russell King - ARM Linux Admin of "Sun, 12 Dec 1999 11:35:25 GMT." <199912121135.LAA02001@raistlin.arm.linux.org.uk> References: <199912121135.LAA02001@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 12 Dec 1999 12:03:38 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk >> and it does seem to use ioremap to get the base address. > >Ah, it does? In that case, it probably doesn't iounmap it actually, That does look like it's the case. It also looks suspiciously like it might be prepared to try and ioremap an I/O address, which isn't too good. p. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 13 19:13:58 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id TAA28118 for linux-arm-kernel-outgoing; Mon, 13 Dec 1999 19:13:58 GMT Received: from kings-cross.london.uk.eu.org (exim@tazenda.demon.co.uk [158.152.220.239]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id TAA28113 for ; Mon, 13 Dec 1999 19:13:54 GMT Received: from localhost ([::ffff:127.0.0.1] helo=kings-cross.london.uk.eu.org ident=phil) by kings-cross.london.uk.eu.org with esmtp (Exim 3.03 #1) id 11xaup-0003Uw-00; Mon, 13 Dec 1999 19:13:47 +0000 X-Mailer: exmh version 2.0.2 2/24/98 (debian) To: Russell King - ARM Linux Admin cc: linux-arm-kernel@lists.arm.linux.org.uk Subject: Re: ebsa-285 i/o patch In-Reply-To: Message from Russell King - ARM Linux Admin of "Sun, 12 Dec 1999 11:35:25 GMT." <199912121135.LAA02001@raistlin.arm.linux.org.uk> References: <199912121135.LAA02001@raistlin.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 13 Dec 1999 19:13:47 +0000 From: Philip Blundell Message-Id: Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Turns out the cyber2000fb driver was broken by that patch, though. Here's the obvious fix. p. Index: cyber2000fb.c =================================================================== RCS file: /cvsroot/pub/kernel/armlinux/drivers/video/cyber2000fb.c,v retrieving revision 1.11 diff -u -p -r1.11 cyber2000fb.c --- cyber2000fb.c 1999/12/05 17:16:10 1.11 +++ cyber2000fb.c 1999/12/13 19:13:05 @@ -1404,7 +1404,6 @@ int __init cyber2000fb_init(void) err = -ENOMEM; goto release_smem_resource; } -current_par.screen_base += PCIMEM_BASE; current_par.screen_size = smem_size; current_par.screen_base_p = smem_base + 0x80000000; current_par.regs_base_p = mmio_base + 0x80000000; unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Mon Dec 13 19:33:41 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id TAA28655 for linux-arm-kernel-outgoing; Mon, 13 Dec 1999 19:33:41 GMT Received: from caramon.arm.linux.org.uk (root@p27-robin-gui.tch.enablis.net [194.168.180.87]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id TAA28651 for ; Mon, 13 Dec 1999 19:33:37 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id TAA05892; Mon, 13 Dec 1999 19:33:23 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id TAA00932; Mon, 13 Dec 1999 19:32:46 GMT Message-Id: <199912131932.TAA00932@raistlin.arm.linux.org.uk> Subject: Re: ebsa-285 i/o patch To: Philip.Blundell@pobox.com (Philip Blundell) Date: Mon, 13 Dec 1999 19:32:43 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: from "Philip Blundell" at Dec 13, 99 07:13:47 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Philip Blundell writes: > Turns out the cyber2000fb driver was broken by that patch, though. Here's the > obvious fix. Already got that one. PS, please don't view linux-arm-kernel as a place to post patches to be integrated into the kernel - the kernel patch system still stands, and should most definitely be used for all kernel patches. You may post the patches here in addition to sending them to the patch system. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Tue Dec 14 02:33:03 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id CAA09380 for linux-arm-kernel-outgoing; Tue, 14 Dec 1999 02:33:03 GMT Received: from mail13.wlv.netzero.net (mail13.wlv.netzero.net [209.247.163.63]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with SMTP id CAA09372 for ; Tue, 14 Dec 1999 02:32:58 GMT Received: (qmail 2224 invoked by uid 0); 14 Dec 1999 02:32:53 -0000 Received: from pppa3-minneapolisb40-1r5052.saturn.bbn.com (HELO netzero.net) (4.12.3.129) by mail13.wlv.netzero.net with SMTP; 14 Dec 1999 02:32:53 -0000 Message-ID: <3855ACC7.637D9F73@netzero.net> Date: Mon, 13 Dec 1999 20:34:48 -0600 From: Gaixia Zhang X-Mailer: Mozilla 4.61 [en] (Win98; I) X-Accept-Language: en MIME-Version: 1.0 To: linux-arm-kernel@lists.arm.linux.org.uk Subject: eepro100.c: ftp can't upload file to EBSA285 server Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Hi, I have eepro100.c:v1.09j 7/27/99 with EBSA285. Ftp download from EBSA285 server works fine, but ftp upload to EBSA285 server doesn't work. I can upload only a tiny file (8 bytes), for bigger files (several hundred bytes or more), ftp just says "open Binary mode for ...." and then hangs, Also I noticed ftp on EBSA to some other server doesnot work either. An "ls" will hang ftp, While everything works fine for 3c590. This is kernel 2.3.18-rmk with D-cache and Write Buffer off. Any suggestions is greatly apprecaied, Thanks a lot, Gaixia __________________________________________ NetZero - Defenders of the Free World Get your FREE Internet Access and Email at http://www.netzero.net/download/index.html unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 16 12:31:35 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id MAA02400 for linux-arm-kernel-outgoing; Thu, 16 Dec 1999 12:31:35 GMT Received: from dhinfo_server.elim.net ([203.239.167.60]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id MAA02396 for ; Thu, 16 Dec 1999 12:30:53 GMT Received: from 捞切疙 by dhinfo_server.elim.net with SMTP (Microsoft Exchange Internet Mail Service Version 5.0.1457.7) id Y6NBV1QW; Thu, 16 Dec 1999 21:35:05 +0900 Message-ID: <3858DBE4.C9507368@donghoit.com> Date: Thu, 16 Dec 1999 21:32:36 +0900 From: hmlee X-Mailer: Mozilla 4.05 [en] (Win95; I) MIME-Version: 1.0 To: linux-arm-kernel@lists.arm.linux.org.uk Subject: "lost interrupt" Content-Type: text/plain; charset=iso-2022-kr Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk Hi all ! I have a problem to running an armlinux. We made a system as follows: - sa-110 - DC21285 ( PCI Arbitor mode ) - VIA82C686 ( South Bridge ) - Keyboard, Mouse etc.. - Linux-2.2.12 with rmk1 - Chips 64550 (vga) - bios-1.05 with some change ( to start at 0x0000 address) I had download image to rom emulator and run it. Following messages shown. Help me! What is the mean of the "lost interrupt". Please tell me what is occurred. Uncompressing Linux...done. Now booting the kernel Linux version 2.2.12-rmk1 (root@arm-linux.net) (gcc version egcs-2.91.66 19990325/philb (egcs-1.1.2 release)) #58 Thu Dec 16 21:07:16 EST 1999 <4>NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp. Console: colour VGA+ 80x25 Calibrating delay loop... 4.60 BogoMIPS Memory: 14796k/16M available (1136k code, 20k reserved, 416k data, 16k init) POSIX conformance testing by UNIFIX <7>PCI: DEC21285 revision 04 PCI: Probing PCI hardware <7>PCI: master abort pc=[] <7>PCI: [1106:0686] status = 2200 <3>IRQ LOCK: IRQ31 is locking the system, disabled <7>PCI: 00:38 [102c/00e0] on irq 21 <7>PCI: 00:40 [1106/0686] on irq 24 <7>PCI: 00:41 [1106/0571] on irq 14 <7>PCI: 00:42 [1106/3038] on irq 22 <7>PCI: 00:43 [1106/3038] on irq 22 <7>PCI: 00:44 [1106/3057] on irq 22 <7>PCI: 00:48 [10b7/5950] on irq 22 <6>Linux NET4.0 for Linux 2.2 <6>Based upon Swansea University Computer Society NET3.039 <6>NET4: Unix domain sockets 1.0 for Linux NET4.0. <6>NET4: Linux TCP/IP 1.0 for NET4.0 <6>IP Protocols: ICMP, UDP, TCP Starting kswapd v 1.5 <6>parport0: PC-style at 0x3bc [SPP] <6>parport1: PC-style at 0x378 [SPP] <6>parport2: PC-style at 0x278 [SPP] <6>Serial driver version 4.27 with MANY_PORTS SHARE_IRQ DETECT_IRQ enabled Software Watchdog Timer: 0.05, timer margin: 60 sec <6>Real Time Clock Driver v1.09 RAM disk driver initialized: 16 RAM disks of 4096K size <6>loop: registered device at major 7 VP_IDE: IDE controller on PCI bus 00 dev 41 VP_IDE: 100% native mode on irq 14 ide0: BM-DMA at 0xe800-0xe807, BIOS settings: hda:pio, hdb:pio ide0: VIA Bus-Master (U)DMA Timing Config Success hda: UUIISS PP3388AA, ATA DISK drive <---- Crashaed.(Fujitsu 8.4G hard disk) ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 <6>hda: UUIISS PP3388AA, 32255MB w/0kB Cache, CHS=65535/16/63 <6>Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M floppy0: Unable to grab DMA2 for the floppy driver <6>3c59x.c:v0.99H 11/17/98 Donald Becker http://cesdis.gsfc.nasa.gov/linux/drivers/vortex.html <6>eth0: 3Com 3c595 Vortex 100baseTx at 0x8000, 00:a0:24:71:d1:03, IRQ 22 <6> 8K byte-wide RAM 5:3 Rx:Tx split, 10baseT interface. <6>eth0: Overriding PCI latency timer (CFLT) setting of 32, new value is 248. <6>Partition check: <6> hda:hda: lost interrupt hda: lost interrupt <---- What ? hda: lost interrupt hda: lost interrupt hda: read_intr: status=0x50 { DriveReady SeekComplete } hda: lost interrupt unknown partition table VFS: Cannot open root device 00:00 <0>Kernel panic: VFS: Unable to mount root fs on 00:00 email : dhinfo@hitel.net tel : 822-402-6844 fax: 822-3432-6844 unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 16 14:51:58 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id OAA07702 for linux-arm-kernel-outgoing; Thu, 16 Dec 1999 14:51:58 GMT Received: from [62.161.177.33] (mailhost.mipsys.com [62.161.177.33]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id OAA07697 for ; Thu, 16 Dec 1999 14:51:53 GMT Received: from [192.168.1.10] by [192.168.1.2] with SMTP (QuickMail Pro Server for Mac 2.0); 16 DEC 99 17:01:53 UT In-Reply-To: <3858DBE4.C9507368@donghoit.com> X-Mailer: CTM PowerMail 2.4v7 Date: Thu, 16 Dec 1999 15:52:23 +0100 To: hmlee , linux-arm-kernel@lists.arm.linux.org.uk From: Benjamin Herrenschmidt Subject: Re: "lost interrupt" Message-Id: <19991216155223.017352@mailhost.mipsys.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk On Thu, Dec 16, 1999, hmlee wrote: ><6>Partition check: ><6> hda:hda: lost interrupt >hda: lost interrupt <---- What ? >hda: lost interrupt >hda: lost interrupt >hda: read_intr: status=0x50 { DriveReady SeekComplete } >hda: lost interrupt > unknown partition table >VFS: Cannot open root device 00:00 ><0>Kernel panic: VFS: Unable to mount root fs on 00:00 There are several possible causes to this problem: 1- Incorrect configuration of the interrupts. Make sure you adapted the VIA driver correctly to find the interrupt yu actually wired on your board, or that you correctly filled the PCI config interrupt number. 2- Make sure the interrupt code handles correctly the cascade (did you actually use a cascaded interrupt controller ? I don't know if the VIA contains one, most southbridges do) 3- Do not use IDE DMA transfers (I mean WIN_DMAREAD/WIN_DMAWRITE commands, not real bus mastering) unless you have done the appropriate IDE SET_FEATURE to the hard disk and correctly programmed the VIA timings. If you interrupt routing and PIC configuration is correct, this error usually means that there is a mismatch between the drive IDE timing configuration (PIO mode, DMA mode, ...) and the timings programmed in the IDE controller. Usually, most recent drives cope quite well with weird PIO timings but not with weird DMA timings. unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 16 22:53:42 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id WAA05189 for linux-arm-kernel-outgoing; Thu, 16 Dec 1999 22:53:42 GMT Received: from caramon.arm.linux.org.uk (root@p54-cordelia-gui.tch.enablis.net [212.250.233.54]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id WAA05179 for ; Thu, 16 Dec 1999 22:53:33 GMT Received: from raistlin.arm.linux.org.uk (linux@raistlin [192.168.0.3]) by caramon.arm.linux.org.uk (8.9.3/8.9.3) with ESMTP id WAA20544; Thu, 16 Dec 1999 22:52:45 GMT From: Russell King - ARM Linux Admin Received: (from linux@localhost) by raistlin.arm.linux.org.uk (8.7.4/8.7.3) id WAA00976; Thu, 16 Dec 1999 22:52:10 GMT Message-Id: <199912162252.WAA00976@raistlin.arm.linux.org.uk> Subject: Re: "lost interrupt" To: hmlee@donghoit.com (hmlee) Date: Thu, 16 Dec 1999 22:52:09 +0000 (GMT) Cc: linux-arm-kernel@lists.arm.linux.org.uk In-Reply-To: <3858DBE4.C9507368@donghoit.com> from "hmlee" at Dec 16, 99 09:32:36 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk hmlee writes: > Uncompressing Linux...done. > Now booting the kernel > Linux version 2.2.12-rmk1 (root@arm-linux.net) (gcc version egcs-2.91.66 2.2.12 is rather old now - you should upgrade to 2.2.13. > <7>PCI: DEC21285 revision 04 > PCI: Probing PCI hardware > <7>PCI: master abort pc=[] > <7>PCI: [1106:0686] status = 2200 > <3>IRQ LOCK: IRQ31 is locking the system, disabled Oops, looks like the PCI error interrupt (assuming that it is still IRQ31) is going permanently active. Check PCI bus connections. > <6>loop: registered device at major 7 > VP_IDE: IDE controller on PCI bus 00 dev 41 > VP_IDE: 100% native mode on irq 14 > ide0: BM-DMA at 0xe800-0xe807, BIOS settings: hda:pio, hdb:pio > ide0: VIA Bus-Master (U)DMA Timing Config Success > hda: UUIISS PP3388AA, ATA DISK drive <---- Crashaed.(Fujitsu 8.4G > hard disk) The name looks wrong - check the PCI data bus, byte enables, and the IDE interface connections. It looks like the data on d0-d7 is being replicated on d8-d15. This is probably also the cause of: > <6> hda:hda: lost interrupt > hda: lost interrupt <---- What ? > hda: lost interrupt > hda: lost interrupt > hda: read_intr: status=0x50 { DriveReady SeekComplete } > hda: lost interrupt I'd suggest an in-depth look at your hardware. _____ |_____| ------------------------------------------------- ---+---+- | | Russell King linux@arm.linux.org.uk --- --- | | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / | | +-+-+ --- -+- / | THE developer of ARM Linux |+| /|\ / | | | --- | +-+-+ ------------------------------------------------- /\\\ | unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk From owner-linux-arm-kernel@lists.arm.linux.org.uk Thu Dec 23 05:20:31 1999 Received: (from majordomo@localhost) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) id FAA09824 for linux-arm-kernel-outgoing; Thu, 23 Dec 1999 05:20:31 GMT Received: from mta3.263.net ([202.96.44.48]) by zen.ics.uwe.ac.uk (8.9.3/8.8.7) with ESMTP id FAA09820 for ; Thu, 23 Dec 1999 05:20:23 GMT From: daihb@263.net Received: by mta3.263.net (Postfix, from userid 60001) id 740AD1C624EDC; Thu, 23 Dec 1999 13:22:44 +0800 (CST) MIME-Version: 1.0 Message-Id: <3861B1A4.13063@mta6> Date: Thu, 23 Dec 1999 13:22:44 +0800 (CST) To: linux-arm-kernel@lists.arm.linux.org.uk Subject: can linxarm run on SA1110? X-Priority: 3 Sender: owner-linux-arm-kernel@lists.arm.linux.org.uk Precedence: bulk I'm a newbie on linux-arm,and i dont know much about this right mailing list,would somebody tell me something about the currenct status of these projects? Thanx. _____________________________________________ 首都在线--先进中国人的网上家园 http://www.263.net 免费邮箱 邮件杂志 签名邮件 邮件加密 邮件追身呼 搜索引擎 个人站点 在线游戏 网上聊天 网上挂号 金融王国 在线杀毒 跳蚤市场 软件下载 休闲娱乐 诺方安全，助您e路平安 unsubscribe: body of `unsubscribe linux-arm-kernel' to majordomo@lists.arm.linux.org.uk