This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH][GCC][mid-end] Allow larger copies when target supports unaligned access [Patch (1/2)]
- From: Jeff Law <law at redhat dot com>
- To: Peter Bergner <bergner at vnet dot ibm dot com>, Tamar Christina <Tamar dot Christina at arm dot com>
- Cc: Alan Modra <amodra at gmail dot com>, Richard Biener <rguenther at suse dot de>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, nd <nd at arm dot com>, "ian at airs dot com" <ian at airs dot com>
- Date: Mon, 9 Apr 2018 16:47:31 -0600
- Subject: Re: [PATCH][GCC][mid-end] Allow larger copies when target supports unaligned access [Patch (1/2)]
- Autocrypt: addr=law at redhat dot com; prefer-encrypt=mutual; keydata= xsBNBFkbIO8BCACVIqDhDVh9ur8C+zNV1J/cXfwvVDAUcphDEFl4jyHqZORK4Pd3Db8oWqLm Q8lOCr/VOS7lrCtdpVMQkLGOGA16oJ8g7hzhnojpjY09UjsoUiG7oKacuxj8skfp6SIx93Zl +iNYPRa4S+za6nY8qiVjyUuiyX04ZPZMrKp2c2sGi+HnBKUZXGhrz/Jdzdox3tjajWZnObyy nhEN6hn9L3KawTtGPE/R6A/1RhHTD9FQmIWIeucpaY5c6GNKXTFpj2VYx57LY5hve1R5vhrJ IZcgwZAiOtmik5lVi96glY5h6bugRwpexjhwORTLPBCkwiYotSxX99mWd6EHL576i5CNABEB AAHNGUplZmYgTGF3IDxsYXdAcmVkaGF0LmNvbT7CwI4EEwEIADgWIQR+niGjtnP5P/8PpRq8 fP682pgzWwUCWRsg7wIbAwULCQgHAgYVCAkKCwIEFgIDAQIeAQIXgAAKCRC8fP682pgzW5QG B/9VATJmx5235RB+8jiDYGXQf3vd9gBfPy/l1tsaK400eFAevDzfGvKmeCKe+uGnlrH3vyT8 rg9zqH+s5a1Y+lDXPOpJAFmmzbOLU4FW4ucbawmtYvBL65PqpQneCTYnC802/OAcxjm/Onem HlgeK6WicNsBTPwYN/0araDFUejyYBIFi9CNqqflwk5Z3brKbQ9bAYIkysVLC/c3njKPmM0c WPFHG91ubLbWCHwTIK0+mAL714eTD74dXzOjO2ZDBPLGlFN/kO3+YjaO6UOD2O8acvAMCivT kWLr7JwRgLIQDN2DkhQDd3LTPqQE/yOcMcXBTO+fxm8KG0iKQBqWMyGJzsBNBFkbIO8BCACy qbOsv7XegSeea8XORt5zMaBVWKoSyhmmcCmlxZFS2cuYOBt79MO13lZE2DlO3Lv5IKikj/D4 ketGVO4+h5psEMH5Yz5P8bx0TmgwbK1GxPZrzeXozUFJDvvCDbIlT0v0pwUXuK3hg8Ieo2h5 uTed/cn1OjySXW5BqLxN0cyr5hL+J6dcsHvKLT/N3nTgCQhoJXK2MrEMhAGgF3jKpMn3CoS4 i/ZbNI2MQR6LWHwdZ95f0fI8NzHSfVzeLtzCKQec7nr9fgd6Ylk1ZpGWQUPlQmKjzYgeCeTK NO04cwt20WIrQWeWiZFPA0U86NDBdSBrYp4kG3dfIXE+wSSvE7qPABEBAAHCwHYEGAEIACAW IQR+niGjtnP5P/8PpRq8fP682pgzWwUCWRsg7wIbDAAKCRC8fP682pgzW3REB/9cT7iKRPg/ OK9bpLlllIEDM90IaKC79DQrv+fRudOR78cdV4XUwPSFnyHUsP3VJ4lDy5FhiKCwGie0BK53 EsxgMrLy1L8hboFdTE4Vi0xzCheMaMVp4hATDU29k1cuxu1VPpCa8E3mYeHjNV7ip0HN5L4D rfs8lRPJE/oM1vGs9DgQFZrCPPNRNGKC97BH+DHccesEJr7tSsQrkPkt0z/FTKr5wIM02vSx OJjgmcVbGB7dc2j/Sx8loXmuKnuKtM35668kUG8jeJvSQk3o/VHpD27bhl0rR68R2jN6G6kQ egMVb6dPu1Ius8rBE5rFw88J4JEb5q4hMNClWWUFHIdP
- Openpgp: preference=signencrypt
- References: <20171115155237.GA30502@arm.com> <alpine.LSU.2.20.1711160858100.12252@zhemvz.fhfr.qr> <20171116114254.GA11570@arm.com> <alpine.LSU.2.20.1711161254130.12252@zhemvz.fhfr.qr> <20171116152701.GA15229@arm.com> <20180329143525.GA3812@bubble.grove.modra.org> <64772721-85a9-96f7-62ef-53c7f31329ed@vnet.ibm.com> <alpine.LSU.2.20.1804031255190.18265@zhemvz.fhfr.qr> <20180403122505.GJ3812@bubble.grove.modra.org> <alpine.LSU.2.20.1804031429420.18265@zhemvz.fhfr.qr> <20180403131610.GK3812@bubble.grove.modra.org> <DB6PR0802MB25044486ABAA6A2E57E58FC5FFA40@DB6PR0802MB2504.eurprd08.prod.outlook.com> <c170a40d-1650-14c7-32f4-b67cad14bc83@vnet.ibm.com>
On 04/04/2018 06:43 AM, Peter Bergner wrote:
> On 4/4/18 4:06 AM, Tamar Christina wrote:
>> Now that I know how the loads are done, I have a patch should be both correct and generate better code in most cases.
>> It just calculates bitsize inside the loop and does the copying in the largest mode possible that's equal or less than the bits
>> That are to be copied. This avoids the issue with reading too much, honors padding and alignment and still generates better code
>> In most cases.
>>
>> I'm running the regression tests and should have a final version soon.
>
> If you give me your patch when it's ready, I'll do bootstrap and regression
> testing on powerpc*-linux and verify it fixes the issues we hit.
Similarly, I've got a jenkins instance here were I can get a bootstrap
and regression test on the usual targets like aarch64, armv7, i686,
powerpc (32, 64, 64le), s390x, sparc64, x86_64. But more interestingly
it'll also do a bootstrap test on alpha, hppa, m68k, sh4 and other
oddballs like aarch64-be.
A patch for the tip of the trunk is all I need.
It doesn't run the testsuite, but the ability to bootstrap on the lesser
used targets gives a level of code generator validation that is helpful.
Takes about 24hrs to cycle through everything...
jeff