From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from az33egw02.freescale.net (az33egw02.freescale.net [192.88.158.103]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "az33egw02.freescale.net", Issuer "Thawte Premium Server CA" (verified OK)) by ozlabs.org (Postfix) with ESMTPS id 19FFAB6EEA for ; Fri, 9 Jul 2010 00:28:05 +1000 (EST) Subject: Re: [1/2] powerpc/crashdump: Fix issues with kexec and 36bit physical addr From: Matthew McClintock To: Milton Miller In-Reply-To: <20100708-mitltonm-msm-reply@mdm.bga.com> References: <1278535881-6463-1-git-send-email-msm@freescale.com> <20100708-mitltonm-msm-reply@mdm.bga.com> Content-Type: text/plain; charset="UTF-8" Date: Thu, 08 Jul 2010 09:27:15 -0500 Message-ID: <1278599235.9343.7.camel@localhost> Mime-Version: 1.0 Cc: kumar.gala@freescale.com, linuxppc-dev@lists.ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Thu, 2010-07-08 at 04:07 -0500, Milton Miller wrote: > On Wed, 07 Jul 2010 around 10:51:20 -0000 Matthew McClintock wrote: > > > > Fix sizes of variables so correct values are exported via /proc. > > Cast variable in comparison to avoid compiler error. > > > > Signed-off-by: Matthew McClintock > > > > > > - csize = min(csize, PAGE_SIZE); > > + csize = min(csize, (size_t)PAGE_SIZE); > > no use min_t Ok > > > > > - if (pfn < max_pfn) { > > + if ((min_low_pfn < pfn) && (pfn < max_pfn)) { > > vaddr = __va(pfn << PAGE_SHIFT); > > csize = copy_oldmem_vaddr(vaddr, buf, csize, offset, userbuf); > > } else { > > diff --git a/arch/powerpc/kernel/machine_kexec.c b/arch/powerpc/kernel/machine_kexec.c > > index bb3d893..ec7f054 100644 > > --- a/arch/powerpc/kernel/machine_kexec.c > > +++ b/arch/powerpc/kernel/machine_kexec.c > > @@ -145,6 +145,7 @@ int overlaps_crashkernel(unsigned long start, unsigned long size) > > > > /* Values we need to export to the second kernel via the device tree. */ > > static unsigned long kernel_end; > > +static unsigned long crashk_start; > > static unsigned long crashk_size; > > > > static struct property kernel_end_prop = { > > @@ -156,7 +157,7 @@ static struct property kernel_end_prop = { > > static struct property crashk_base_prop = { > > .name = "linux,crashkernel-base", > > .length = sizeof(unsigned long), > > - .value = &crashk_res.start, > > + .value = &crashk_start, > > }; > > > > This is wrong, its truncating the start and size. > > Change the variables to be physaddr_t and the length to be sizeof(physaddr_t). > > Since these properites only contain one variable, the number of cells > can be inferred from the property size like we do for reading the initrd-size. > > Technically they should be an array of be32 but we can make that a comment. I don't disagree but this can break kexec if phys_addr_t != unsigned long. Also, doesn't the crash kernel have to live below 2GB so unsigned long is always fine? Will still change unless I hear otherwise. > > By the way, why does 32 bit care about the running kernel's size? aka > linux,kernel-end? 64 bit book 3S needs it because we use mmu hooks > to copy the pages to their destination, but I thought ppc32 was using > a relocatable copy routine. Are we missing the code to create > temp ref tlbs on the fly for book 3E? This is not really in this patch or did I miss something? Kexec uses kernel-end just to add as a invalid region. Not crucial though for 32 bit. > > > static struct property crashk_size_prop = { > > @@ -180,6 +181,7 @@ static void __init export_crashk_values(struct device_node *node) > > prom_remove_property(node, prop); > > > > if (crashk_res.start != 0) { > > + crashk_start = crashk_res.start; > > prom_add_property(node, &crashk_base_prop); > > crashk_size = crashk_res.end - crashk_res.start + 1; > > prom_add_property(node, &crashk_size_prop); > > I guess we use the reuse of the resources varables, but such is > common code vs userspace. > > milton