Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 3 Jul 2019 15:57:26 -0700 (PDT)
From:   Paul Walmsley <paul.walmsley@sifive.com>
To:     Alexandre Ghiti <alex@ghiti.fr>
cc:     Hanjun Guo <guohanjun@huawei.com>,
        Christoph Hellwig <hch@infradead.org>,
        Mike Kravetz <mike.kravetz@oracle.com>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will.deacon@arm.com>,
        Palmer Dabbelt <palmer@sifive.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        "H . Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
        linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-riscv@lists.infradead.org
Subject: Re: [PATCH v3 0/2] Hugetlbfs support for riscv
In-Reply-To: <20190701175900.4034-1-alex@ghiti.fr>
Message-ID: <alpine.DEB.2.21.9999.1907031344330.10620@viisi.sifive.com>
References: <20190701175900.4034-1-alex@ghiti.fr>
User-Agent: Alpine 2.21.9999 (DEB 301 2018-08-15)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Alex,

Thanks for writing and testing these patches, and thanks for your patience 
waiting for reviews and testing.

On Mon, 1 Jul 2019, Alexandre Ghiti wrote:

> This series introduces hugetlbfs support for both riscv 32/64. Riscv32           
> is architecturally limited to huge pages of size 4MB whereas riscv64 has         
> 2MB/1G huge pages support. Transparent huge page support is not                  
> implemented here, I will submit another series later.                            
>                                                                                  

[ ... ]

> This series was validated using libhugetlbfs testsuite ported to riscv64         
> without linker script support.                                                   
> (https://github.com/AlexGhiti/libhugetlbfs.git, branch dev/alex/riscv).          
>                                                                                  
> - libhugetlbfs testsuite on riscv64/2M:                                          
>   - brk_near_huge triggers an assert in malloc.c, does not on x86.               

I was able to reproduce the 2MB megapages test results on rv64 QEMU.  On a 
HiFive Unleashed, though, a few more tests fail:

- icache_hygiene fails ("icache unclean")

  # LD_LIBRARY_PATH=obj64 ./tests/obj64/icache-hygiene
  Starting testcase "./tests/obj64/icache-hygiene", pid 732
  SIGILL at 0x15559fff80 (sig_expected=0x15559fff80)
  SIGILL at 0x1555dfff80 (sig_expected=0x1555dfff80)
  SIGILL at 0x15561fff80 (sig_expected=0x15561fff80)
  SIGILL at 0x15565fff80 (sig_expected=0x15565fff80)
  SIGILL at 0x15569fff80 (sig_expected=0x15569fff80)
  SIGILL at 0x1556dfff80 (sig_expected=(nil))
  FAIL   SIGILL somewhere unexpected
  #

- One of the heapshrink tests fails ("Heap did not shrink")

  # LD_PRELOAD="obj64/libhugetlbfs_privutils.so obj64/libhugetlbfs.so tests/obj64/libheapshrink.so" HUGETLB_MORECORE_SHRINK=yes HUGETLB_MORECORE=yes tests/obj64/heapshrink
  Starting testcase "tests/obj64/heapshrink", pid 753
  FAIL    Heap did not shrink
  #

Some of these may be related to the top-down mmap work, but there might be 
more work to do on actual hardware.

> - libhugetlbfs testsuite on riscv64/1G:                                          
>   - brk_near_huge triggers an assert in malloc.c, does not on x86.               
>   - mmap-gettest, mmap-cow: testsuite passes the number of default free          
>     pages as parameters and then fails for 1G which is not the default.          
>     Otherwise succeeds when given the right number of pages.                     
>   - map_high_truncate_2 fails on x86 too: 0x60000000 is not 1G aligned           
>     and fails at line 694 of fs/hugetlbfs/inode.c.                               
>   - heapshrink on 1G fails on x86 too, not investigated.                         
>   - counters.sh on 1G fails on x86 too: alloc_surplus_huge_page returns          
>     NULL in case of gigantic pages.                                              
>   - icache-hygiene succeeds after patch #3 of this series which lowers           
>     the base address of mmap.                                                    
>   - fallocate_stress.sh on 1G never ends, on x86 too, not investigated.          

I can reproduce some of these here on QEMU.  But for reasons that are 
unclear to me, 1G gigapages aren't working on the HiFive Unleashed here.

In any case, these patches are clearly a good start, so I've queued 
them for v5.3.


- Paul