From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tom Rini Date: Fri, 1 Mar 2013 09:20:32 -0500 Subject: [U-Boot] Please pull u-boot-x86.git In-Reply-To: References: <20130217205804.9E8232005E0@gemini.denx.de> <20130218225202.GD3722@bill-the-cat> <20130228230019.GD20569@bill-the-cat> Message-ID: <5130B930.2000705@ti.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/28/2013 10:55 PM, Simon Glass wrote: > Hi Tom, > > On Thu, Feb 28, 2013 at 4:22 PM, Simon Glass > wrote: >> Hi Tom, >> >> On Thu, Feb 28, 2013 at 3:00 PM, Tom Rini wrote: >>> On Wed, Feb 27, 2013 at 01:18:23PM -0800, Simon Glass wrote: >>> >>>> Hi Tom, >>>> >>>> I have pulled the latest series into a branch in the x86 >>>> tree. You can also get it from patchwork. If you are happy >>>> with it, please see below. I haven't seen any comments for a >>>> few days. >>> >>> OK, building with ELDK4.2 for a number of ARM boards such as >>> igep0030: cmd_mem.c: In function 'do_mem_mtest': cmd_mem.c:979: >>> warning: passing argument 1 of 'unmap_sysmem' discards >>> qualifiers from pointer target type cmd_mem.c:980: warning: >>> passing argument 1 of 'unmap_sysmem' discards qualifiers from >>> pointer target type >>> >> >> Strange - there is even an explicit cast,. But mine is gcc 4.4.1 >> so may be a bit later. I could just remove those two lines since >> they are only there for semantic correctness and compile to >> nothing anyway. But I will track down that tool chain and see if >> I can work out a fix. > > OK I have repeated this - it seems that the compiler does not like > a direct 'cast away' of volatile in a function argument. I have > added a work-around, and sent out an updated patch 20. You can > either pull this in from patchwork, or I have updated the pull > information below. Thanks, I'll give it a day or two in public before re-pulling. > Rather ominously this might mean that I need to start building > with multiple tool chains for each architecture. I was rather > hoping to avoid that... > > Thanks for spotting it. Sadly, it's a must. ELDK4.2 is the oldest toolchain we support, but we do support it for ARM (except when THUMB is required, those fail gracefully tho). My build loops include ELDK4.2/arm, ELDK5.2/arm and whatever Linaro PPA has up, for ARM (and eldk5.2 for powerpc). - -- Tom -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJRMLkvAAoJENk4IS6UOR1W5SUQAJ8a/lgm4/0IC3iQidTuhYr7 7/Xp3lwZkFvcIYYcV3hlMwejHwEZ6+YLynCNwRaRcLu2rPjrO2VrCyezFF0lXBRB +REFPzPelUWRmzNlJNE95rlf1R1Oj+NMalou/BLtP6UYP/BDsYqEO1DSg+b+xAzU 1ynx4Jbz/elszgII8On/rn6fjbLn1Nx8r+GvEd/EoZQNb4ykgp/P23FY5n1SpDav QgLsOBnoGNh3t+iM/d1Uqkp9SxaCmgVkgCVjWMYBsLrDB0lQKS92P3cXPQ7eYPPu wW3mlKxJI3E42X+XLxk961BmGvyzWDxs92/YW+ctTZ52jKmbBciD1+Yk7p7jlKLP inBPKmBChswzpZOyyg/aOCJ3DENc6ad8JQA5kELtn3wOlJJcwUDJdEzvaRFwvbQV qUqSO+aUjL6Kd7lvn72MI40sqgPxleP84fCy3ZjqFISGPtcFGrsc9NVteJM4k59R S6FEwYdWYarMhJq+1Q25tyfEMjUzvLLARdPx88nh97Ew4D7d3ZGt8ma6ClfO0/Ku dzS0bBhlRn2ZUfSHjPRl8SvoyjjRT19VO5zrucIupnU5Zx+wU4g1h/bMoGlDxTVw 3Yf8OhFCAXWnhPtsIQTj7KM5Asgl1waX5jc4afiEyTt+SqT2BNtcShjyfUR2c3gh pI2LEszwK0iD7K0E6p5m =oihH -----END PGP SIGNATURE-----