From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id 6CBD63858D1E for ; Mon, 19 Sep 2022 14:49:01 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org 6CBD63858D1E Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=redhat.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1663598941; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=MINt8wTePVpKLApf1bT20BLTVqjZ4P0eDNNKKu9RI4I=; b=VxMnLbwNemiVDf1X4y2MBE+3be51jhgn9FOGaqxKbo/qsHHFLLNmjz3nX7DjwH9luBQUux WiVCXBHSAWRnt2VBTxKC2aVUkmuSS+xv+74YMKomKX8ecRzbz1CGGtRolWHKJPRgG8Cvi2 p/+fyGXtTUw+caMzG5AFbwMi4LXdu0A= Received: from mail-yw1-f199.google.com (mail-yw1-f199.google.com [209.85.128.199]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_128_GCM_SHA256) id us-mta-673-1SQDXHyGMn-tlrua-exLxg-1; Mon, 19 Sep 2022 10:48:59 -0400 X-MC-Unique: 1SQDXHyGMn-tlrua-exLxg-1 Received: by mail-yw1-f199.google.com with SMTP id 00721157ae682-3452b545544so252924037b3.3 for ; Mon, 19 Sep 2022 07:48:59 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=cc:to:subject:message-id:date:from:in-reply-to:references :mime-version:x-gm-message-state:from:to:cc:subject:date; bh=MINt8wTePVpKLApf1bT20BLTVqjZ4P0eDNNKKu9RI4I=; b=KPI30t/ICFEcyhqZr0vn3L55SP0GOrWL2+ZYoPsGWCAgFf0F/At0vfWRAEVMyXxr1l wpKP/0sCOT5OVHmoc1FDvr4M6+gjSa678JlbFk+CFcU/saUPGYhrvALdD+iw5KGCxqmF O3opFnTPVwbkxZpmNHWrHv3lAmYUF/B8ZeJuVBYPkPt5b8wwZfv1u+d2dusBMZrT+/cn W3COevqTqY5Sc7eMgRZ6CvXneU+4og4ISdadXMkpmQc+mVxmK81bDxuK8syUrFiA2Lnq tkCoEiS02CZxIQHtmTdvB9Nq2FEkmBKhnwgVLKZ4dSjR7ds44mCd+v6973bUKHKNXQBe /jrw== X-Gm-Message-State: ACrzQf2oM+fWur3BDfAysT9d3VFi5OWIzBstElkDq8wPQPW3pF24VFvU k65ZA+MLgSpojNsk7w2v94voSG3kQmOEzJmVNmiSSHndi3aRaVYPgMqq0tQQJxJJiarGtpS6pxD PmiJ6qpvuiG9tDBKBBOKxCWOgUSaSqRw= X-Received: by 2002:a25:4e0b:0:b0:6a8:ec00:c1c5 with SMTP id c11-20020a254e0b000000b006a8ec00c1c5mr14338478ybb.525.1663598938713; Mon, 19 Sep 2022 07:48:58 -0700 (PDT) X-Google-Smtp-Source: AMsMyM6640RtGrk124Wnq9xoVkshLFYfDLslONd3vqWHAr2U5E03zbgS0HJzjwKemjkX0GTsL+pl5pSo5FfOCFrPePg= X-Received: by 2002:a25:4e0b:0:b0:6a8:ec00:c1c5 with SMTP id c11-20020a254e0b000000b006a8ec00c1c5mr14338428ybb.525.1663598937731; Mon, 19 Sep 2022 07:48:57 -0700 (PDT) MIME-Version: 1.0 References: In-Reply-To: From: Jeff Johnston Date: Mon, 19 Sep 2022 10:48:46 -0400 Message-ID: Subject: Re: Malloc: unusable area at the end of the heap section To: Torbjorn SVENSSON Cc: Jerome Leroux , "newlib@sourceware.org" , Niklas DAHLQUIST X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: multipart/alternative; boundary="000000000000bb860505e908cf3f" X-Spam-Status: No, score=-5.0 required=5.0 tests=BAYES_00,DKIMWL_WL_HIGH,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,HTML_MESSAGE,KAM_LOTSOFHASH,RCVD_IN_DNSWL_LOW,SPF_HELO_NONE,SPF_NONE,TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org List-Id: --000000000000bb860505e908cf3f Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Hi Torbjorn, Adding the small memory logic is trivial so I'll add it and then push the change. Thanks, -- Jeff J. On Sat, Sep 17, 2022 at 5:01 AM Torbjorn SVENSSON wrote: > Hello, > > > > The patch looks good to me. > > Thank you Jeff for adding the HAVE_SYSCONF_PAGESIZE logic! > > > > > > One minor question. > > When HAVE_SYSCONF_PAGESIZE is not set, there would be 2 possible values > (128 or 4096). Would it make sense to also have that ifdef part in the > sysconf function? I know that my original patch did not include this, but > maybe it would be beneficial to have sysconf(_SC_PAGESIZE) return 128 when > SMALL_MEMORY is defined and sysconf is not overridden by the application? > > > > Kind regards, > > Torbj=C3=B6rn > > > > > > ST Restricted > > *From:* Jeff Johnston > *Sent:* den 16 september 2022 22:14 > *To:* Torbjorn SVENSSON > *Cc:* Jerome Leroux ; newlib@sourceware.org > *Subject:* Re: Malloc: unusable area at the end of the heap section > > > > Ok, > > > > I am attaching the modified patch after discussing with Torbjorn about the > licensing. Torbjorn, please > > review and let me know if there are changes you would like, otherwise, I > will push. > > > > -- Jeff J. > > > > > > On Wed, Sep 14, 2022 at 4:52 PM Jeff Johnston wrote: > > Hi Torbjorn, > > > > I took a look at what would be needed to make this more generic. I have a > patch almost ready, but > > the one issue is that your libc/sys/arm/sysconf.c file does not have a > license. Could you please append a > > newer version of the file with a license and I will complete the patch for > you to review? > > > > I decided to go with a simple HAVE_xxxx macro rather than a configure > option so any platform can > > simply set the flag in configure.host. > > > > -- Jeff J. > > > > On Tue, Jun 7, 2022 at 12:03 PM Jeff Johnston wrote: > > Hi Torbjorn, > > > > I think it would be useful. Do you want to modify the patch to be more > generic? I am thinking of a var set in configure.host that sets a compile > flag at the end such as _USE_SYSCONF_FOR_PAGESIZE. Then, > > the default sysconf.c can be put in the newlib/libc/unix directory. It is > then a straight-forward exercise to add this as an enablement configurati= on > option. What do you think? > > > > -- Jeff J. > > > > On Tue, Jun 7, 2022 at 2:18 AM Torbjorn SVENSSON via Newlib < > newlib@sourceware.org> wrote: > > Hello, > > A while back, I provided a patch[1] to newlib that would allow the > application to override the pagesize for malloc, but the patch got stalle= d. > Maybe this would be a good time to take another look at the patch and see > if it would actually fix the generic newlib usage or if it's still > something that is only applicable for small embedded targets. > > [1] https://ecos.sourceware.org/ml/newlib/current/017616.html > > Kind regards, > Torbj=C3=B6rn > > > -----Original Message----- > > From: Newlib > bounces+torbjorn.svensson=3Dst.com@sourceware.org> On Behalf Of Jerome > > Leroux > > Sent: den 6 juni 2022 23:18 > > To: newlib@sourceware.org > > Subject: Malloc: unusable area at the end of the heap section > > > > Hello Newlib developers, > > > > I am a user of Newlib in a project that runs on an NXP MCU. > > I am using MCUXpressoIDE_11.3.0_5180_prc3, which comes with GCC =E2=80= =9Carm- > > none-eabi-gcc.exe (GNU Arm Embedded Toolchain > > 9-2020-q2-update) 9.3.1 20200408 (release)=E2=80=9D and Newlib 3.3.0. > > > > I have identified an issue in malloc, and I think the problem is still > present in > > the latest version of Newlib. I could > > not see any changes in the incriminated code since Newlib 3.3.0. > > > > I noticed this issue only in the standard malloc implementation and not > in the > > nano-malloc version. > > > > Here is a description of the problem: > > The allocator splits the heap into pages. When a page is full, it > increases the > > heap size by reserving a new page in the > > heap section. When reserving a new page, the allocator keeps the page e= nd > > address aligned with malloc_getpagesize, which > > is set to 4096 by default. If there is not enough space to reserve the > full page, > > the allocation fails even if there is > > enough space in the heap to allocate the chunk of memory. > > Because the issue is related to the heap end address and how the linker > > positions the heap, the same sequence of > > allocations may lead to different results (failure or success) depending > on the > > location of the heap, even if the heap > > size is constant. Typically, adding a new C global variable can shift > the start > > address of the heap section and cause a > > malloc error. > > > > For example, with a heap section of 4096 bytes (0x1000 bytes): > > If the heap section address is 0x20100-0x21100, during the > initialization, the > > page end address is set to 0x21000 > > (aligned on 4096). We will be able to allocate until the address > 0x21000. After > > that, the allocator will try to reserve > > a new page, but it will fail because it won=E2=80=99t be able to reserv= e a 4096 > bytes > > page from 0x21000 to 0x22000. The > > following allocations will fail. The usable heap size is 3840 bytes > (0x21000 - > > 0x20100) instead of 4096. > > If the heap section address is 0x20F00-0x21F00 (same size), with the sa= me > > scenario, the usable heap size is 256 bytes > > (0x21000 - 0x20F00). > > Here are two examples of heap configurations: > > https://gist.github.com/jerome- > > leroux/759159fbd3e7bb5e189dbceb04636914?permalink_comment_id=3D4191 > > 266#gistcomment-4191266 > > > > I did not dig into the implementation so much. From my understanding, t= he > > problem comes from the usage of > > "malloc_getpagesize" (see > > https://github.com/bminor/newlib/blob/830a9b707caa5e343b6ffce7fcb2d3c > > a97e3259c/newlib/libc/stdlib/_mallocr.c#L198) in > > "malloc_extend_top" (probably here > > https://github.com/bminor/newlib/blob/830a9b707caa5e343b6ffce7fcb2d3c > > a97e3259c/newlib/libc/stdlib/_mallocr.c#L2166). > > I can understand it makes sense to keep the pages aligned when running > in a > > system that implements virtual memory. > > Still, on an MCU, the heap is just a contiguous chunk of memory > allocated at > > link time. Furthermore, the heap size is > > usually pretty small (a few kilobytes), so potentially wasting 4 KB of > memory > > is unacceptable. Using the default > > implementation of "sbrk" documented at > > https://sourceware.org/newlib/libc.html#index-sbrk will lead to the > > problem. > > > > I have written a simple example that demonstrates the issue (see > > https://gist.github.com/jerome- > > leroux/759159fbd3e7bb5e189dbceb04636914 ). To reproduce the problem, > > define the macros > > HEAP_SECTION_START_SYMBOL and HEAP_SECTION_END_SYMBOL, which > > are specific to your environment. Then call the function > > "test_malloc()". > > > > I tried to find someone with the same issue, but I couldn=E2=80=99t. Th= e related > > commits/discussions I found are: > > - > > https://github.com/bminor/newlib/commit/4a3d0a5a5d829c05868a34658eb > > 45731dbb5112b > > - https://stackoverflow.com/questions/39088598/malloc-in-newlib-does-it- > > waste-memory-after-one-big-failure-allocation > > > > Can anyone confirm what I have noticed? > > > > Thanks. > > > > -- > > Jerome Leroux > > --000000000000bb860505e908cf3f--