Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 4BDF7C87FCC for ; Thu, 31 Jul 2025 09:02:54 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 1160410E73D; Thu, 31 Jul 2025 09:02:54 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="TppFAEU6"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.16]) by gabe.freedesktop.org (Postfix) with ESMTPS id 7EAD310E73D for ; Thu, 31 Jul 2025 09:02:52 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1753952572; x=1785488572; h=message-id:date:mime-version:subject:to:cc:references: from:in-reply-to:content-transfer-encoding; bh=tZ8o2Bj4yaEWgGcPj/VUxG+reLNJz6trABnet/XRoIw=; b=TppFAEU6VcNa23lLdGQ4QOuWqmvjKo7Whmcy26RQ7xbXsrwJBwUmhQjs kkB9i7+ywuHwU+A+MPcN+4orRU4j3/HXNa8R484a6kmcD9s7R3Gqy80WL 8B9IsrDOMEh+pCbaAgaV+d0pnzmvOyNt/LoP8f3ohj3yYmwzhL5eWXt17 V6uHTtvyo1fDhh4YBzgl5h60uUVfsfhVE0zpsHunPQJeRpHwWuNg5X9/H r9oc2nGF24QlxsoPi47tza53jgsK6GpOPYZJq/uUpaFFyEHAjBHVu8Go5 IPBMeLB9R/PymIB6e49TMTxBFX3hqBEt4uh/NhYojkud44/J9FycyVpMN Q==; X-CSE-ConnectionGUID: JTbKi1n/QvKrCkoJTOivcQ== X-CSE-MsgGUID: n1r7kU4zQjuqzMIg0JVykA== X-IronPort-AV: E=McAfee;i="6800,10657,11507"; a="43873949" X-IronPort-AV: E=Sophos;i="6.16,353,1744095600"; d="scan'208";a="43873949" Received: from orviesa006.jf.intel.com ([10.64.159.146]) by fmvoesa110.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 31 Jul 2025 02:02:52 -0700 X-CSE-ConnectionGUID: 0ZYejaqOQGu6Jx7Z1+7oyw== X-CSE-MsgGUID: ACDiSOlaQKyfzavsNbAVag== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.16,353,1744095600"; d="scan'208";a="162489307" Received: from sschumil-mobl2.ger.corp.intel.com (HELO [10.245.245.250]) ([10.245.245.250]) by orviesa006-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 31 Jul 2025 02:02:50 -0700 Message-ID: <9ee7071f-5d88-4b0e-89b4-b50792fa42de@intel.com> Date: Thu, 31 Jul 2025 10:02:48 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH 1/3] drm/xe/migrate: prevent infinite recursion To: "Summers, Stuart" , "intel-xe@lists.freedesktop.org" Cc: "Patelczyk, Maciej" , "Brost, Matthew" References: <20250730143335.174373-5-matthew.auld@intel.com> <20250730143335.174373-6-matthew.auld@intel.com> <8947a74a8e0d73a50e29c20b5d6093d22e93d620.camel@intel.com> Content-Language: en-GB From: Matthew Auld In-Reply-To: <8947a74a8e0d73a50e29c20b5d6093d22e93d620.camel@intel.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-BeenThere: intel-xe@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel Xe graphics driver List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" On 30/07/2025 19:11, Summers, Stuart wrote: > On Wed, 2025-07-30 at 15:33 +0100, Matthew Auld wrote: >> If the buf + offset is not aligned to XE_CAHELINE_BYTES we fallback >> to >> using a bounce buffer. However the bounce buffer here is allocated on >> the stack, and the only alignment requirement here is that it's >> naturally aligned to u8, and not XE_CACHELINE_BYTES. If the bounce >> buffer is also misaligned we then recurse back into the function >> again, >> however the new bounce buffer might also not be aligned, and might >> never >> be until we eventually blow through the stack, as we keep recursing. >> >> Instead of using the stack use kmalloc, which should respect the >> power-of-two alignment request here. Fixes a kernel panic when > > Should we add a build bug on (or similar) to guarantee that the power- > of-2 stays here? My understanding is this alignment isn't guaranteed > otherwise. Of course the chance of the cacheline being a non-power-of-2 > seems low... Agreed, will add. > >> triggering this path through eudebug. >> >> Fixes: 270172f64b11 ("drm/xe: Update xe_ttm_access_memory to use GPU >> for non-visible access") >> Signed-off-by: Matthew Auld >> Cc: Maciej Patelczyk >> Cc: Matthew Brost >> --- >>  drivers/gpu/drm/xe/xe_migrate.c | 31 +++++++++++++++++++------------ >>  1 file changed, 19 insertions(+), 12 deletions(-) >> >> diff --git a/drivers/gpu/drm/xe/xe_migrate.c >> b/drivers/gpu/drm/xe/xe_migrate.c >> index 3a276e2348a2..b6bdc9bd32b0 100644 >> --- a/drivers/gpu/drm/xe/xe_migrate.c >> +++ b/drivers/gpu/drm/xe/xe_migrate.c >> @@ -1987,15 +1987,21 @@ int xe_migrate_access_memory(struct >> xe_migrate *m, struct xe_bo *bo, >>         if (!IS_ALIGNED(len, XE_CACHELINE_BYTES) || >>             !IS_ALIGNED((unsigned long)buf + offset, >> XE_CACHELINE_BYTES)) { >>                 int buf_offset = 0; >> +               void *bounce; >> +               int err; >> + >> +               bounce = kmalloc(XE_CACHELINE_BYTES, GFP_KERNEL); >> +               if (!bounce) >> +                       return -EINVAL; > > ENOMEM? Oops. Thanks. > > Thanks, > Stuart > >> + >> +               xe_assert(xe, IS_ALIGNED((unsigned long)bounce, >> +                                        XE_CACHELINE_BYTES)); >> >>                 /* >>                  * Less than ideal for large unaligned access but >> this should be >>                  * fairly rare, can fixup if this becomes common. >>                  */ >>                 do { >> -                       u8 bounce[XE_CACHELINE_BYTES]; >> -                       void *ptr = (void *)bounce; >> -                       int err; >>                         int copy_bytes = min_t(int, bytes_left, >>                                                XE_CACHELINE_BYTES - >>                                                (offset & >> XE_CACHELINE_MASK)); >> @@ -2004,22 +2010,22 @@ int xe_migrate_access_memory(struct >> xe_migrate *m, struct xe_bo *bo, >>                         err = xe_migrate_access_memory(m, bo, >>                                                        offset & >> >> ~XE_CACHELINE_MASK, >> -                                                      (void *)ptr, >> - >> sizeof(bounce), 0); >> +                                                      bounce, >> + >> XE_CACHELINE_BYTES, 0); >>                         if (err) >> -                               return err; >> +                               break; >> >>                         if (write) { >> -                               memcpy(ptr + ptr_offset, buf + >> buf_offset, copy_bytes); >> +                               memcpy(bounce + ptr_offset, buf + >> buf_offset, copy_bytes); >> >>                                 err = xe_migrate_access_memory(m, bo, >>                                                                offset >> & ~XE_CACHELINE_MASK, >> -                                                              (void >> *)ptr, >> - >> sizeof(bounce), write); >> + >> bounce, >> + >> XE_CACHELINE_BYTES, write); >>                                 if (err) >> -                                       return err; >> +                                       break; >>                         } else { >> -                               memcpy(buf + buf_offset, ptr + >> ptr_offset, >> +                               memcpy(buf + buf_offset, bounce + >> ptr_offset, >>                                        copy_bytes); >>                         } >> >> @@ -2028,7 +2034,8 @@ int xe_migrate_access_memory(struct xe_migrate >> *m, struct xe_bo *bo, >>                         offset += copy_bytes; >>                 } while (bytes_left); >> >> -               return 0; >> +               kfree(bounce); >> +               return err; >>         } >> >>         dma_addr = xe_migrate_dma_map(xe, buf, len + page_offset, >> write); >