From owner-svn-src-all@FreeBSD.ORG Fri Oct 29 14:17:24 2010 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D42291065672; Fri, 29 Oct 2010 14:17:24 +0000 (UTC) (envelope-from asmrookie@gmail.com) Received: from mail-qy0-f182.google.com (mail-qy0-f182.google.com [209.85.216.182]) by mx1.freebsd.org (Postfix) with ESMTP id 42E2B8FC12; Fri, 29 Oct 2010 14:17:23 +0000 (UTC) Received: by qyk2 with SMTP id 2so2110490qyk.13 for ; Fri, 29 Oct 2010 07:17:23 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:sender:received :in-reply-to:references:date:x-google-sender-auth:message-id:subject :from:to:cc:content-type:content-transfer-encoding; bh=KdEu707QFwEZVYQJHFIjZ0jvM6iiLVY9Qub9ZjCi7lw=; b=UNWrHjafRAdV5F6+y7Fy9A0XlBTGgdXfgyd4Vf0+FRpUbrAI/ihXpCSkmQtU/TTnfV Vo3ZxViEbqRCqbA++dM1b84TpKcmU5GSBMtgpjgffG/Kr1EzMoyK1FWxlsw5QpRfd9cr ocsMKN1DdgTfnRyPalnTYohEa34eT59RxfMeE= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=B9ZF1HYUwjIaCwcj5YYE313s8lZ3capCnOTQh286VHwoD6o7CUXAsdrQ/TCNz6yhyK GdIPkwan2uXFm4r6xkMDKZrm2vprQG+2FYQRwoqrQ9evDYKqU5ECY0a11GG366wt2NCm dM5QROebiy5qLnV0LF9xwb+TeC0YYIpKc+DJQ= MIME-Version: 1.0 Received: by 10.229.240.213 with SMTP id lb21mr11753797qcb.185.1288361843381; Fri, 29 Oct 2010 07:17:23 -0700 (PDT) Sender: asmrookie@gmail.com Received: by 10.229.237.9 with HTTP; Fri, 29 Oct 2010 07:17:23 -0700 (PDT) In-Reply-To: <201010290944.53628.jhb@freebsd.org> References: <201010281631.o9SGVdtZ014923@svn.freebsd.org> <201010281638.14043.jhb@freebsd.org> <201010290944.53628.jhb@freebsd.org> Date: Fri, 29 Oct 2010 16:17:23 +0200 X-Google-Sender-Auth: _fNL_4Q-S5v50NB8uHPlnX_6f7U Message-ID: From: Attilio Rao To: John Baldwin Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r214457 - in head/sys: amd64/amd64 conf i386/i386 x86/x86 X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 29 Oct 2010 14:17:24 -0000 2010/10/29 John Baldwin : > On Thursday, October 28, 2010 8:25:38 pm Attilio Rao wrote: >> 2010/10/28 John Baldwin : >> > On Thursday, October 28, 2010 4:28:12 pm Attilio Rao wrote: >> >> 2010/10/28 John Baldwin : >> >> > On Thursday, October 28, 2010 1:21:34 pm Attilio Rao wrote: >> >> >> 2010/10/28 John Baldwin : >> >> >> > On Thursday, October 28, 2010 12:31:39 pm Attilio Rao wrote: >> >> >> >> Author: attilio >> >> >> >> Date: Thu Oct 28 16:31:39 2010 >> >> >> >> New Revision: 214457 >> >> >> >> URL: http://svn.freebsd.org/changeset/base/214457 >> >> >> >> >> >> >> >> Log: >> >> >> >> =C2=A0 Merge nexus.c from amd64 and i386 to x86 subtree. >> >> >> >> >> >> >> >> =C2=A0 Sponsored by: =C2=A0 =C2=A0 =C2=A0 Sandvine Incorporated >> >> >> >> =C2=A0 Tested by: =C2=A0gianni >> >> >> >> >> >> >> > >> >> >> > It would be better to merge these two routines. =C2=A0The loader= now passes >> > the >> >> >> > smap to i386 kernels as well, so ram_attach() should probably be >> > changed to >> >> >> > try the amd64 approach first and if that fails fall back to usin= g the >> >> >> > phys_avail[] array instead. >> >> >> >> >> >> What do you think about this patch?: >> >> >> Index: nexus.c >> >> >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> >> >> --- nexus.c =C2=A0 =C2=A0 (revision 214457) >> >> >> +++ nexus.c =C2=A0 =C2=A0 (working copy) >> >> >> @@ -52,9 +52,7 @@ >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> -#ifdef __amd64__ >> >> >> =C2=A0#include >> >> >> -#endif >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> @@ -67,12 +65,10 @@ >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> >> >> >> -#ifdef __amd64__ >> >> >> =C2=A0#include >> >> >> -#include >> >> >> -#endif >> >> >> =C2=A0#include >> >> >> =C2=A0#include >> >> >> +#include >> >> >> >> >> >> =C2=A0#ifdef DEV_APIC >> >> >> =C2=A0#include "pcib_if.h" >> >> >> @@ -89,11 +85,13 @@ >> >> >> =C2=A0#include >> >> >> >> >> >> =C2=A0#ifdef __amd64__ >> >> >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_IO =C2=A0 =C2= =A0 =C2=A0 AMD64_BUS_SPACE_IO >> >> >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_MEM =C2=A0 =C2= =A0 =C2=A0AMD64_BUS_SPACE_MEM >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0X86_BUS_SPACE_IO =C2=A0 =C2=A0= =C2=A0 =C2=A0AMD64_BUS_SPACE_IO >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0X86_BUS_SPACE_MEM =C2=A0 =C2= =A0 =C2=A0 AMD64_BUS_SPACE_MEM >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0ELF_KERN_STR =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0"elf64 kernel" >> >> >> =C2=A0#else >> >> >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_IO =C2=A0 =C2= =A0 =C2=A0 I386_BUS_SPACE_IO >> >> >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_MEM =C2=A0 =C2= =A0 =C2=A0I386_BUS_SPACE_MEM >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0X86_BUS_SPACE_IO =C2=A0 =C2=A0= =C2=A0 =C2=A0I386_BUS_SPACE_IO >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0X86_BUS_SPACE_MEM =C2=A0 =C2= =A0 =C2=A0 I386_BUS_SPACE_MEM >> >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0ELF_KERN_STR =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0"elf32 kernel" >> >> >> =C2=A0#endif >> >> >> @@ -701,16 +699,11 @@ >> >> >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 panic("ram_attach: resource %d failed to attach", >> > rid); >> >> >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rid++; >> >> >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 } >> >> >> - =C2=A0 =C2=A0 =C2=A0 return (0); >> >> >> -} >> >> >> -#else >> >> >> -static int >> >> >> -ram_attach(device_t dev) >> >> >> -{ >> >> >> - =C2=A0 =C2=A0 =C2=A0 struct resource *res; >> >> >> - =C2=A0 =C2=A0 =C2=A0 vm_paddr_t *p; >> >> >> - =C2=A0 =C2=A0 =C2=A0 int error, i, rid; >> >> >> >> >> >> + =C2=A0 =C2=A0 =C2=A0 /* If at least one smap attached, return. *= / >> >> >> + =C2=A0 =C2=A0 =C2=A0 if (rid !=3D 0) >> >> >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 return (0); >> >> >> + >> >> > >> >> > Perhaps this instead: >> >> > >> >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0/* If we found an SMAP, return. */ >> >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (smapbase !=3D NULL) >> >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return (0); >> >> >> >> No, I don't think this check is right, smapbase will always be !=3D N= ULL >> >> (otherwise the code panics). >> > >> > Oh, that needs to be fixed then. =C2=A0It can be NULL on i386 with an = old loader >> > (or on a really old machine without an SMAP). =C2=A0The amd64 nexus co= de could >> > assume it would never be NULL, but i386 cannot. >> > >> > It should probably more closely match what i386 does during the memory= probe >> > which is: >> > >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0kmdp =3D search("elf kernel"); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (kmdp =3D=3D NULL) >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0kmdp =3D search= ("elfXX kernel"); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (kmdp !=3D NULL) >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0smapbase =3D pr= eload_search(...); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0else >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0smapbase =3D NU= LL; >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0if (smapbase !=3D NULL) { >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0for (smap =3D .= ..) { >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0} >> > >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return (0); >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0} >> > >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0/* fall through to old i386 code using phys= _avail[] */ >> >> This further patch should address your concerns about nexus.c: >> Index: nexus.c >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> --- nexus.c =C2=A0 =C2=A0 (revision 214457) >> +++ nexus.c =C2=A0 =C2=A0 (working copy) >> @@ -52,9 +52,7 @@ >> =C2=A0#include >> =C2=A0#include >> =C2=A0#include >> -#ifdef __amd64__ >> =C2=A0#include >> -#endif >> =C2=A0#include >> =C2=A0#include >> =C2=A0#include >> @@ -67,12 +65,10 @@ >> =C2=A0#include >> =C2=A0#include >> >> -#ifdef __amd64__ >> =C2=A0#include >> -#include >> -#endif >> =C2=A0#include >> =C2=A0#include >> +#include >> >> =C2=A0#ifdef DEV_APIC >> =C2=A0#include "pcib_if.h" >> @@ -89,13 +85,15 @@ >> =C2=A0#include >> >> =C2=A0#ifdef __amd64__ >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_IO =C2=A0 =C2=A0 =C2= =A0 AMD64_BUS_SPACE_IO >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_MEM =C2=A0 =C2=A0 =C2= =A0AMD64_BUS_SPACE_MEM >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0BUS_SPACE_IO =C2=A0 =C2=A0AMD64_BUS_= SPACE_IO >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0BUS_SPACE_MEM =C2=A0 AMD64_BUS_SPACE= _MEM >> =C2=A0#else >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_IO =C2=A0 =C2=A0 =C2= =A0 I386_BUS_SPACE_IO >> -#define =C2=A0 =C2=A0 =C2=A0 =C2=A0RMAN_BUS_SPACE_MEM =C2=A0 =C2=A0 =C2= =A0I386_BUS_SPACE_MEM >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0BUS_SPACE_IO =C2=A0 =C2=A0I386_BUS_S= PACE_IO >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0BUS_SPACE_MEM =C2=A0 I386_BUS_SPACE_= MEM >> =C2=A0#endif >> >> +#define =C2=A0 =C2=A0 =C2=A0 =C2=A0ELF_KERN_STR =C2=A0 =C2=A0("elf"__XS= TRING(__ELF_WORD_SIZE)" kernel") >> + >> =C2=A0static MALLOC_DEFINE(M_NEXUSDEV, "nexusdev", "Nexus device"); >> >> =C2=A0#define DEVTONX(dev) =C2=A0 ((struct nexus_device *)device_get_iva= rs(dev)) >> @@ -435,7 +433,7 @@ >> =C2=A0#else >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_bushand= le(r, rman_get_start(r)); >> =C2=A0#endif >> - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_bustag(r, RM= AN_BUS_SPACE_IO); >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_bustag(r, BU= S_SPACE_IO); >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 break; >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 case SYS_RES_MEMORY: >> =C2=A0#ifdef PC98 >> @@ -446,7 +444,7 @@ >> =C2=A0#endif >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 vaddr =3D pmap_m= apdev(rman_get_start(r), rman_get_size(r)); >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_virtual= (r, vaddr); >> - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_bustag(r, RM= AN_BUS_SPACE_MEM); >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 rman_set_bustag(r, BU= S_SPACE_MEM); >> =C2=A0#ifdef PC98 >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 /* PC-98: the ty= pe of bus_space_handle_t is the structure. */ >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 bh->bsh_base =3D= (bus_addr_t) vaddr; >> @@ -668,48 +666,45 @@ >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 return (0); >> =C2=A0} >> >> -#ifdef __amd64__ >> =C2=A0static int >> =C2=A0ram_attach(device_t dev) >> =C2=A0{ >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 struct bios_smap *smapbase, *smap, *smapend; >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 struct resource *res; >> + =C2=A0 =C2=A0 =C2=A0 vm_paddr_t *p; >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 caddr_t kmdp; >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 uint32_t smapsize; >> - =C2=A0 =C2=A0 =C2=A0 int error, rid; >> + =C2=A0 =C2=A0 =C2=A0 int error, i, rid; >> >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 /* Retrieve the system memory map from the l= oader. */ >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 kmdp =3D preload_search_by_type("elf kernel"= ); >> =C2=A0 =C2=A0 =C2=A0 =C2=A0 if (kmdp =3D=3D NULL) >> - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 kmdp =3D preload_sear= ch_by_type("elf64 kernel"); >> - =C2=A0 =C2=A0 =C2=A0 smapbase =3D (struct bios_smap *)preload_search_i= nfo(kmdp, >> - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 MODINFO_METADATA | MODINFOMD_SMAP); >> - =C2=A0 =C2=A0 =C2=A0 smapsize =3D *((u_int32_t *)smapbase - 1); >> - =C2=A0 =C2=A0 =C2=A0 smapend =3D (struct bios_smap *)((uintptr_t)smapb= ase + smapsize); >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 kmdp =3D preload_sear= ch_by_type(ELF_KERN_STR); >> + =C2=A0 =C2=A0 =C2=A0 if (kmdp !=3D NULL) { >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 smapbase =3D (struct = bios_smap *)preload_search_info(kmdp, >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 MODINFO= _METADATA | MODINFOMD_SMAP); >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 smapsize =3D *((u_int= 32_t *)smapbase - 1); >> + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 smapend =3D (struct b= ios_smap *)((uintptr_t)smapbase + smapsize); > > I think this doesn't check for the case that smapbase is NULL. =C2=A0I wo= uld > expect kmdp to always be non-NULL in practice (it just means we have a > kernel). =C2=A0However, the preload_search_info() can fail when booting o= n an old > machine without an SMAP or an old loader that doesn't pass the SMAP to i3= 86. I don't understand where is the problem. smap is only used when it is consistent in that patch, otherwise it is not used and fallsthrough the physavail[] approach. It doesn't need to be set to NULL or checked. I guess that you want to patch the file and see how it looks like, will be easier to review than inline .diff, in this case. Attilio --=20 Peace can only be achieved by understanding - A. Einstein