linux-kernel - Re: [PATCH v2 1/1] initrd: move initrd_start calculate within linear mapping range check

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190401153810.GC6884@fuggles.cambridge.arm.com>
Date:   Mon, 1 Apr 2019 16:38:10 +0100
From:   Will Deacon <will.deacon@....com>
To:     pierre kuo <vichy.kuo@...il.com>
Cc:     Catalin Marinas <catalin.marinas@....com>,
        Steven Price <steven.price@....com>,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        Florian Fainelli <f.fainelli@...il.com>,
        ard.biesheuvel@...aro.org
Subject: Re: [PATCH v2 1/1] initrd: move initrd_start calculate within linear
 mapping range check

[+Ard in case I'm missing something]

On Mon, Apr 01, 2019 at 10:59:53PM +0800, pierre kuo wrote:
> > > With CONFIG_RANDOMIZE_BASE we can get a further change to memstart_addr
> > > after the place where you moved the initrd_{start,end} setting, which
> > > would result in a different value for __phys_to_virt(phys_initrd_start).
> >
> > I found what you mean, since __phys_to_virt will use PHYS_OFFSET
> > (memstart_addr) for calculating.
> > How about moving CONFIG_RANDOMIZE_BASE part of code ahead of
> > CONFIG_BLK_DEV_INITRD checking?
> >
> > That means below (d) moving ahead of (c)
> > prvious:
> > if (memstart_addr + linear_region_size < memblock_end_of_DRAM()) {} ---(a)
> > if (memory_limit != PHYS_ADDR_MAX) {}            ---(b)
> > if (IS_ENABLED(CONFIG_BLK_DEV_INITRD) && phys_initrd_size) {} ---(c)
> > if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) {} ---(d)
> >
> > now:
> > if (memstart_addr + linear_region_size < memblock_end_of_DRAM()) { ---(a)
> > if (memory_limit != PHYS_ADDR_MAX) {}              ----------------(b)
> > if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) {}  --------------(d)
> > if (IS_ENABLED(CONFIG_BLK_DEV_INITRD) && phys_initrd_size) {}  ---(c)
> >
> 
> After tracing the kernel code,
> is it even possible to move CONFIG_RANDOMIZE_BASE part of code ahead
> of memory_limit?
> 
> that mean the flow may look like below:
> now2:
> if (memstart_addr + linear_region_size < memblock_end_of_DRAM()) {} ---(a)
> if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) {} ---(d)
> if (memory_limit != PHYS_ADDR_MAX) {}            ---(b)
> if (IS_ENABLED(CONFIG_BLK_DEV_INITRD) && phys_initrd_size) {} ---(c)
> 
> in (b), the result of __pa_symbol(_text), memory_limit,
> memblock_mem_limit_remove_map and  memblock_add
> are not depended on memsart_addr.
> So the now2 flow can grouping modification of memstart_address, put
> (a) and (d) together.

I'm afraid that you've lost me with this. Why isn't it just as simple as
the diff below?

Will

--->8

diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
index c29b17b520cd..ec3487c94b10 100644
--- a/arch/arm64/mm/init.c
+++ b/arch/arm64/mm/init.c
@@ -377,7 +377,7 @@ void __init arm64_memblock_init(void)
 			 base + size > memblock_start_of_DRAM() +
 				       linear_region_size,
 			"initrd not fully accessible via the linear mapping -- please check your bootloader ...\n")) {
-			initrd_start = 0;
+			phys_initrd_size = 0;
 		} else {
 			memblock_remove(base, size); /* clear MEMBLOCK_ flags */
 			memblock_add(base, size);