From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from EUR05-VI1-obe.outbound.protection.outlook.com (mail-vi1eur05on2062.outbound.protection.outlook.com [40.107.21.62]) by sourceware.org (Postfix) with ESMTPS id C41B83858426 for ; Tue, 22 Aug 2023 11:21:50 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org C41B83858426 Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=arm.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=arm.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=armh.onmicrosoft.com; s=selector2-armh-onmicrosoft-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=P2ZBfnwiYwZZhA+Ye5n5liGl62yApGtAVaiZ+6AA0fg=; b=wnJ0WOaXsZEJULcFuD3FaihG2HzvQcDA0AFZltlgGSt8SHR8vvRRwfSsCMLPkUOvcxBtqPWHhDdPr5bUsfDL5KfTy8HjFBYsyqsJ4gIfA7wpCbwLm6x7sqF6tp0aA+rmrPw1TfOzIDIHajc37+PdXxJBN6keyHwFgd+NTbWMLUM= Received: from DUZPR01CA0107.eurprd01.prod.exchangelabs.com (2603:10a6:10:4bb::25) by VI1PR08MB5343.eurprd08.prod.outlook.com (2603:10a6:803:12d::21) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6699.20; Tue, 22 Aug 2023 11:21:47 +0000 Received: from DBAEUR03FT015.eop-EUR03.prod.protection.outlook.com (2603:10a6:10:4bb:cafe::4) by DUZPR01CA0107.outlook.office365.com (2603:10a6:10:4bb::25) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6699.20 via Frontend Transport; Tue, 22 Aug 2023 11:21:47 +0000 X-MS-Exchange-Authentication-Results: spf=pass (sender IP is 63.35.35.123) smtp.mailfrom=arm.com; dkim=pass (signature was verified) header.d=armh.onmicrosoft.com;dmarc=pass action=none header.from=arm.com; Received-SPF: Pass (protection.outlook.com: domain of arm.com designates 63.35.35.123 as permitted sender) receiver=protection.outlook.com; client-ip=63.35.35.123; helo=64aa7808-outbound-1.mta.getcheckrecipient.com; pr=C Received: from 64aa7808-outbound-1.mta.getcheckrecipient.com (63.35.35.123) by DBAEUR03FT015.mail.protection.outlook.com (100.127.142.112) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6723.15 via Frontend Transport; Tue, 22 Aug 2023 11:21:46 +0000 Received: ("Tessian outbound c99fbc01d472:v175"); Tue, 22 Aug 2023 11:21:46 +0000 X-CheckRecipientChecked: true X-CR-MTA-CID: c6590858adc2e791 X-CR-MTA-TID: 64aa7808 Received: from e68900f8b7f5.1 by 64aa7808-outbound-1.mta.getcheckrecipient.com id 6A7FC4E3-BB36-4C5F-AE0C-D87129C60FFB.1; Tue, 22 Aug 2023 11:21:39 +0000 Received: from EUR05-AM6-obe.outbound.protection.outlook.com by 64aa7808-outbound-1.mta.getcheckrecipient.com with ESMTPS id e68900f8b7f5.1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384); Tue, 22 Aug 2023 11:21:39 +0000 ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=XbhH9iJiWTcbmYutGHXLy6Elz8kQS/5nJWkte+EXZUyO7uNSC2EfI4xgsg6bpTDxt0KiUksHO+sW7At028Gum14RMjIStg9uze0jIHY57W+/wLWFqkIcDsmKwtjUMSe9YxlxwM8hXVyHlvG2igo+SPf/mdn6LrdFUUhRb7a9d7Fu9URqyPgywJcMp29/9jNuUfOfejCM1BLTMP1uK8xJdvzuZk7xwQlf2yoG0iIOxliLzd/5jCFBOje5/cWmi1QSzN1WriVJPSneBT2R98ztHtMwQbUBOZIMZkG3I+fl00u4c4JJcn04/ylbnYV67tInDkzdVPGcWQDBjjLqfWCjDQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=P2ZBfnwiYwZZhA+Ye5n5liGl62yApGtAVaiZ+6AA0fg=; b=RjJK5BfVu3k6q24jKjVhZVj31JeCkwlH9eCRur+SdLsB0WXDWqpkQczbkObeRGi2htszOVn0p6mCX8dx8R/IU2D1FLyoFFePN/N/quxm0+HP9QI7p9Tddkf/YEzFe04gtXF38F70rSEydXCSqd/XqnNf2A/jg6ZDXSmoQ0SryPSu9vwNGhx9BTZlU92ZUG/WjWuQ9Bir7QJ5hqgA1IsCQqQGE3IzwcmxkkWtTkIjFKf2b4x1kD8fZxRpnjSqey7lkF8KOuTh4JT01LFFebsbbbDqOXwPVppFbXZpqIyKJwQ1eHiEX/KGYYoBuvuW2yBjnY07+DlyyzixfP1f85FFnA== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass (sender ip is 40.67.248.234) smtp.rcpttodomain=sourceware.org smtp.mailfrom=arm.com; dmarc=pass (p=none sp=none pct=100) action=none header.from=arm.com; dkim=none (message not signed); arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=armh.onmicrosoft.com; s=selector2-armh-onmicrosoft-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=P2ZBfnwiYwZZhA+Ye5n5liGl62yApGtAVaiZ+6AA0fg=; b=wnJ0WOaXsZEJULcFuD3FaihG2HzvQcDA0AFZltlgGSt8SHR8vvRRwfSsCMLPkUOvcxBtqPWHhDdPr5bUsfDL5KfTy8HjFBYsyqsJ4gIfA7wpCbwLm6x7sqF6tp0aA+rmrPw1TfOzIDIHajc37+PdXxJBN6keyHwFgd+NTbWMLUM= Received: from AS9PR07CA0003.eurprd07.prod.outlook.com (2603:10a6:20b:46c::16) by AM8PR08MB5569.eurprd08.prod.outlook.com (2603:10a6:20b:1de::19) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6699.24; Tue, 22 Aug 2023 11:21:37 +0000 Received: from AM7EUR03FT027.eop-EUR03.prod.protection.outlook.com (2603:10a6:20b:46c:cafe::ff) by AS9PR07CA0003.outlook.office365.com (2603:10a6:20b:46c::16) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.6699.15 via Frontend Transport; Tue, 22 Aug 2023 11:21:37 +0000 X-MS-Exchange-Authentication-Results: spf=pass (sender IP is 40.67.248.234) smtp.mailfrom=arm.com; dkim=none (message not signed) header.d=none;dmarc=pass action=none header.from=arm.com; Received-SPF: Pass (protection.outlook.com: domain of arm.com designates 40.67.248.234 as permitted sender) receiver=protection.outlook.com; client-ip=40.67.248.234; helo=nebula.arm.com; pr=C Received: from nebula.arm.com (40.67.248.234) by AM7EUR03FT027.mail.protection.outlook.com (100.127.140.124) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.20.6723.15 via Frontend Transport; Tue, 22 Aug 2023 11:21:37 +0000 Received: from AZ-NEU-EX04.Arm.com (10.251.24.32) by AZ-NEU-EX03.Arm.com (10.251.24.31) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2507.27; Tue, 22 Aug 2023 11:21:36 +0000 Received: from e129171.cambridge.arm.com (10.1.34.58) by mail.arm.com (10.251.24.32) with Microsoft SMTP Server id 15.1.2507.27 via Frontend Transport; Tue, 22 Aug 2023 11:21:36 +0000 From: Luis Machado To: CC: Subject: [PATCH v4 03/16] [gdb/gdbserver] refactor: Simplify SVE interface to read/write registers Date: Tue, 22 Aug 2023 12:21:17 +0100 Message-ID: <20230822112130.1513216-4-luis.machado@arm.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20230822112130.1513216-1-luis.machado@arm.com> References: <20230822112130.1513216-1-luis.machado@arm.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain X-EOPAttributedMessage: 1 X-MS-TrafficTypeDiagnostic: AM7EUR03FT027:EE_|AM8PR08MB5569:EE_|DBAEUR03FT015:EE_|VI1PR08MB5343:EE_ X-MS-Office365-Filtering-Correlation-Id: ff35fb19-0b02-4041-9b40-08dba301f8ec x-checkrecipientrouted: true NoDisclaimer: true X-MS-Exchange-SenderADCheck: 1 X-MS-Exchange-AntiSpam-Relay: 0 X-Microsoft-Antispam-Untrusted: BCL:0; X-Microsoft-Antispam-Message-Info-Original: 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 X-Forefront-Antispam-Report-Untrusted: CIP:40.67.248.234;CTRY:IE;LANG:en;SCL:1;SRV:;IPV:NLI;SFV:NSPM;H:nebula.arm.com;PTR:InfoDomainNonexistent;CAT:NONE;SFS:(13230031)(4636009)(136003)(376002)(396003)(39860400002)(346002)(1800799009)(186009)(451199024)(82310400011)(36840700001)(40470700004)(46966006)(6916009)(316002)(70586007)(70206006)(8676002)(8936002)(2616005)(4326008)(36756003)(40460700003)(41300700001)(1076003)(82740400003)(356005)(81166007)(478600001)(6666004)(40480700001)(83380400001)(30864003)(2906002)(86362001)(47076005)(7696005)(36860700001)(336012)(44832011)(426003)(5660300002)(26005)(36900700001);DIR:OUT;SFP:1101; X-MS-Exchange-Transport-CrossTenantHeadersStamped: AM8PR08MB5569 X-MS-Exchange-Transport-CrossTenantHeadersStripped: DBAEUR03FT015.eop-EUR03.prod.protection.outlook.com X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id-Prvs: 2d0adfda-47ff-4159-cdbe-08dba301f329 X-Microsoft-Antispam: BCL:0; X-Microsoft-Antispam-Message-Info: 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 X-Forefront-Antispam-Report: CIP:63.35.35.123;CTRY:IE;LANG:en;SCL:1;SRV:;IPV:CAL;SFV:NSPM;H:64aa7808-outbound-1.mta.getcheckrecipient.com;PTR:ec2-63-35-35-123.eu-west-1.compute.amazonaws.com;CAT:NONE;SFS:(13230031)(4636009)(136003)(376002)(396003)(39850400004)(346002)(1800799009)(186009)(451199024)(82310400011)(36840700001)(40470700004)(46966006)(6916009)(316002)(70586007)(70206006)(8676002)(8936002)(2616005)(107886003)(4326008)(36756003)(40460700003)(41300700001)(1076003)(82740400003)(81166007)(478600001)(6666004)(40480700001)(83380400001)(30864003)(2906002)(86362001)(47076005)(7696005)(36860700001)(336012)(44832011)(426003)(5660300002)(26005);DIR:OUT;SFP:1101; X-OriginatorOrg: arm.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 22 Aug 2023 11:21:46.8982 (UTC) X-MS-Exchange-CrossTenant-Network-Message-Id: ff35fb19-0b02-4041-9b40-08dba301f8ec X-MS-Exchange-CrossTenant-Id: f34e5979-57d9-4aaa-ad4d-b122a662184d X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=f34e5979-57d9-4aaa-ad4d-b122a662184d;Ip=[63.35.35.123];Helo=[64aa7808-outbound-1.mta.getcheckrecipient.com] X-MS-Exchange-CrossTenant-AuthSource: DBAEUR03FT015.eop-EUR03.prod.protection.outlook.com X-MS-Exchange-CrossTenant-AuthAs: Anonymous X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem X-MS-Exchange-Transport-CrossTenantHeadersStamped: VI1PR08MB5343 X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00,DKIM_SIGNED,DKIM_VALID,FORGED_SPF_HELO,GIT_PATCH_0,KAM_DMARC_NONE,RCVD_IN_DNSWL_NONE,RCVD_IN_MSPIKE_H2,SPF_HELO_PASS,SPF_NONE,TXREP,UNPARSEABLE_RELAY 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: Updates on v4: - Removed unused return value. - Fixed buffer overrun in a for loop. --- This is a patch in preparation to upcoming patches enabling SME support. It attempts to simplify the gdb/gdbserver shared interface used to read/write SVE registers. Where the current code makes use of unique_ptr, allocating a new buffer by hand and passing a buffer around, this patch makes that code use gdb::byte_vector and passes a reference to this byte vector to the functions, allowing the functions to have ready access to the size of the buffer. It also shares a bit more code between gdb and gdbserver, in particular around handling of ptrace get/set requests for SVE. I think gdbserver could be refactored to handle register reads/writes more like gdb's native layer as opposed to letting the generic linux-low layer do the ptrace calls. This is not very flexible and assumes one size for the responses. If you have something like NT_ARM_SVE, where you can have either FPSIMD or SVE contents, it doesn't work that well. I didn't want to change that interface right now as it is a bit too much work and touches all the targets, some of which I can't easily test. Hence the reason why the buffer the generic linux-now passes down to linux-aarch64-low is unused or ignored. No user-visible changes should happen as part of this refactor other than a slightly reworded warning message. While doing the refactor, I also noticed what seems to be a mistake in checking if the register cache contains active (non-zero) SVE data. For instance, the original code did something like this in aarch64_sve_regs_copy_from_reg_buf: has_sve_state |= reg_buf->raw_compare (AARCH64_SVE_Z0_REGNUM + i reg, sizeof (__int128_t)); "reg" is a zeroed-out buffer that we compare the Z register contents past the first 128 bits. The problem here is that raw_compare returns 1 if the contents compare the same, which means has_sve_state will be true. But if we compared the Z register contents to 0, it means we *do not* have SVE state, and therefore has_sve_state should be false. The consequence of this mistake is that we convert the initial FPSIMD-formatted data we get from ptrace for the NT_ARM_SVE register set to a SVE-formatted one. In the end, this doesn't cause user-visible differences because the values of both the Z and V registers will still be the same. But the logic is not correct. I used the opportunity to fix this, and it gets tested later on by the additional SME tests. I do plan on submitting some SVE-specific tests to make sure we have a bit more coverage in GDB's testsuite. Regression-tested on aarch64-linux Ubuntu 22.04/20.04. --- gdb/aarch64-linux-nat.c | 30 +--- gdb/nat/aarch64-scalable-linux-ptrace.c | 202 +++++++++++++++--------- gdb/nat/aarch64-scalable-linux-ptrace.h | 27 ++-- gdbserver/linux-aarch64-low.cc | 24 ++- 4 files changed, 173 insertions(+), 110 deletions(-) diff --git a/gdb/aarch64-linux-nat.c b/gdb/aarch64-linux-nat.c index 9f32279c0ac..267a1ca0e79 100644 --- a/gdb/aarch64-linux-nat.c +++ b/gdb/aarch64-linux-nat.c @@ -319,9 +319,8 @@ store_fpregs_to_thread (const struct regcache *regcache) static void fetch_sveregs_from_thread (struct regcache *regcache) { - std::unique_ptr base - = aarch64_sve_get_sveregs (regcache->ptid ().lwp ()); - aarch64_sve_regs_copy_to_reg_buf (regcache, base.get ()); + /* Fetch SVE state from the thread and copy it into the register cache. */ + aarch64_sve_regs_copy_to_reg_buf (regcache->ptid ().lwp (), regcache); } /* Store to the current thread the valid sve register @@ -330,28 +329,9 @@ fetch_sveregs_from_thread (struct regcache *regcache) static void store_sveregs_to_thread (struct regcache *regcache) { - int ret; - struct iovec iovec; - int tid = regcache->ptid ().lwp (); - - /* First store vector length to the thread. This is done first to ensure the - ptrace buffers read from the kernel are the correct size. */ - if (!aarch64_sve_set_vq (tid, regcache)) - perror_with_name (_("Unable to set VG register")); - - /* Obtain a dump of SVE registers from ptrace. */ - std::unique_ptr base = aarch64_sve_get_sveregs (tid); - - /* Overwrite with regcache state. */ - aarch64_sve_regs_copy_from_reg_buf (regcache, base.get ()); - - /* Write back to the kernel. */ - iovec.iov_base = base.get (); - iovec.iov_len = ((struct user_sve_header *) base.get ())->size; - ret = ptrace (PTRACE_SETREGSET, tid, NT_ARM_SVE, &iovec); - - if (ret < 0) - perror_with_name (_("Unable to store sve registers")); + /* Fetch SVE state from the register cache and update the thread TID with + it. */ + aarch64_sve_regs_copy_from_reg_buf (regcache->ptid ().lwp (), regcache); } /* Fill GDB's register array with the pointer authentication mask values from diff --git a/gdb/nat/aarch64-scalable-linux-ptrace.c b/gdb/nat/aarch64-scalable-linux-ptrace.c index cc43f510892..6eea9104cef 100644 --- a/gdb/nat/aarch64-scalable-linux-ptrace.c +++ b/gdb/nat/aarch64-scalable-linux-ptrace.c @@ -120,28 +120,43 @@ aarch64_sve_set_vq (int tid, struct reg_buffer_common *reg_buf) /* See nat/aarch64-scalable-linux-ptrace.h. */ -std::unique_ptr -aarch64_sve_get_sveregs (int tid) +gdb::byte_vector +aarch64_fetch_sve_regset (int tid) { - struct iovec iovec; uint64_t vq = aarch64_sve_get_vq (tid); if (vq == 0) - perror_with_name (_("Unable to fetch SVE register header")); + perror_with_name (_("Unable to fetch SVE vector length")); /* A ptrace call with NT_ARM_SVE will return a header followed by either a dump of all the SVE and FP registers, or an fpsimd structure (identical to the one returned by NT_FPREGSET) if the kernel has not yet executed any SVE code. Make sure we allocate enough space for a full SVE dump. */ - iovec.iov_len = SVE_PT_SIZE (vq, SVE_PT_REGS_SVE); - std::unique_ptr buf (new gdb_byte[iovec.iov_len]); - iovec.iov_base = buf.get (); + gdb::byte_vector sve_state (SVE_PT_SIZE (vq, SVE_PT_REGS_SVE), 0); + + struct iovec iovec; + iovec.iov_base = sve_state.data (); + iovec.iov_len = sve_state.size (); if (ptrace (PTRACE_GETREGSET, tid, NT_ARM_SVE, &iovec) < 0) perror_with_name (_("Unable to fetch SVE registers")); - return buf; + return sve_state; +} + +/* See nat/aarch64-scalable-linux-ptrace.h. */ + +void +aarch64_store_sve_regset (int tid, const gdb::byte_vector &sve_state) +{ + struct iovec iovec; + /* We need to cast from (const void *) here. */ + iovec.iov_base = (void *) sve_state.data (); + iovec.iov_len = sve_state.size (); + + if (ptrace (PTRACE_SETREGSET, tid, NT_ARM_SVE, &iovec) < 0) + perror_with_name (_("Unable to store SVE registers")); } /* If we are running in BE mode, byteswap the contents @@ -165,11 +180,13 @@ aarch64_maybe_swab128 (gdb_byte *dst, const gdb_byte *src, size_t size) /* See nat/aarch64-scalable-linux-ptrace.h. */ void -aarch64_sve_regs_copy_to_reg_buf (struct reg_buffer_common *reg_buf, - const void *buf) +aarch64_sve_regs_copy_to_reg_buf (int tid, struct reg_buffer_common *reg_buf) { - char *base = (char *) buf; - struct user_sve_header *header = (struct user_sve_header *) buf; + gdb::byte_vector sve_state = aarch64_fetch_sve_regset (tid); + + char *base = (char *) sve_state.data (); + struct user_sve_header *header + = (struct user_sve_header *) sve_state.data (); uint64_t vq = sve_vq_from_vl (header->vl); uint64_t vg = sve_vg_from_vl (header->vl); @@ -249,18 +266,33 @@ aarch64_sve_regs_copy_to_reg_buf (struct reg_buffer_common *reg_buf, reg_buf->raw_supply (AARCH64_SVE_FFR_REGNUM, reg); } + + /* At this point we have updated the register cache with the contents of + the NT_ARM_SVE register set. */ } /* See nat/aarch64-scalable-linux-ptrace.h. */ void -aarch64_sve_regs_copy_from_reg_buf (const struct reg_buffer_common *reg_buf, - void *buf) +aarch64_sve_regs_copy_from_reg_buf (int tid, + struct reg_buffer_common *reg_buf) { - struct user_sve_header *header = (struct user_sve_header *) buf; - char *base = (char *) buf; + /* First store the vector length to the thread. This is done first to + ensure the ptrace buffers read from the kernel are the correct size. */ + if (!aarch64_sve_set_vq (tid, reg_buf)) + perror_with_name (_("Unable to set VG register")); + + /* Obtain a dump of SVE registers from ptrace. */ + gdb::byte_vector sve_state = aarch64_fetch_sve_regset (tid); + + struct user_sve_header *header = (struct user_sve_header *) sve_state.data (); uint64_t vq = sve_vq_from_vl (header->vl); + gdb::byte_vector new_state (SVE_PT_SIZE (32, SVE_PT_REGS_SVE), 0); + memcpy (new_state.data (), sve_state.data (), sve_state.size ()); + header = (struct user_sve_header *) new_state.data (); + char *base = (char *) new_state.data (); + /* Sanity check the data in the header. */ if (!sve_vl_valid (header->vl) || SVE_PT_SIZE (vq, header->flags) != header->size) @@ -275,36 +307,40 @@ aarch64_sve_regs_copy_from_reg_buf (const struct reg_buffer_common *reg_buf, resulting in the initialization of SVE state written back to the kernel, which is why we try to avoid it. */ - bool has_sve_state = false; - gdb_byte *reg = (gdb_byte *) alloca (SVE_PT_SVE_ZREG_SIZE (vq)); - struct user_fpsimd_state *fpsimd - = (struct user_fpsimd_state *)(base + SVE_PT_FPSIMD_OFFSET); - - memset (reg, 0, SVE_PT_SVE_ZREG_SIZE (vq)); + /* Buffer (using the maximum size a Z register) used to look for zeroed + out sve state. */ + gdb_byte reg[256]; + memset (reg, 0, sizeof (reg)); /* Check in the reg_buf if any of the Z registers are set after the first 128 bits, or if any of the other SVE registers are set. */ - + bool has_sve_state = false; for (int i = 0; i < AARCH64_SVE_Z_REGS_NUM; i++) { - has_sve_state |= reg_buf->raw_compare (AARCH64_SVE_Z0_REGNUM + i, - reg, sizeof (__int128_t)); - if (has_sve_state) - break; + if (!reg_buf->raw_compare (AARCH64_SVE_Z0_REGNUM + i, reg, + V_REGISTER_SIZE)) + { + has_sve_state = true; + break; + } } if (!has_sve_state) for (int i = 0; i < AARCH64_SVE_P_REGS_NUM; i++) { - has_sve_state |= reg_buf->raw_compare (AARCH64_SVE_P0_REGNUM + i, - reg, 0); - if (has_sve_state) - break; + if (!reg_buf->raw_compare (AARCH64_SVE_P0_REGNUM + i, reg, 0)) + { + has_sve_state = true; + break; + } } if (!has_sve_state) - has_sve_state |= reg_buf->raw_compare (AARCH64_SVE_FFR_REGNUM, - reg, 0); + has_sve_state + = !reg_buf->raw_compare (AARCH64_SVE_FFR_REGNUM, reg, 0); + + struct user_fpsimd_state *fpsimd + = (struct user_fpsimd_state *)(base + SVE_PT_FPSIMD_OFFSET); /* If no SVE state exists, then use the existing fpsimd structure to write out state and return. */ @@ -344,50 +380,74 @@ aarch64_sve_regs_copy_from_reg_buf (const struct reg_buffer_common *reg_buf, if (REG_VALID == reg_buf->get_register_status (AARCH64_FPCR_REGNUM)) reg_buf->raw_collect (AARCH64_FPCR_REGNUM, &fpsimd->fpcr); - return; - } - - /* Otherwise, reformat the fpsimd structure into a full SVE set, by - expanding the V registers (working backwards so we don't splat - registers before they are copied) and using null for everything else. - Note that enough space for a full SVE dump was originally allocated - for base. */ + /* At this point we have collected all the data from the register + cache and we are ready to update the FPSIMD register content + of the thread. */ - header->flags |= SVE_PT_REGS_SVE; - header->size = SVE_PT_SIZE (vq, SVE_PT_REGS_SVE); + /* Fall through so we can update the thread's contents with the + FPSIMD register cache values. */ + } + else + { + /* Otherwise, reformat the fpsimd structure into a full SVE set, by + expanding the V registers (working backwards so we don't splat + registers before they are copied) and using zero for everything + else. + Note that enough space for a full SVE dump was originally allocated + for base. */ + + header->flags |= SVE_PT_REGS_SVE; + header->size = SVE_PT_SIZE (vq, SVE_PT_REGS_SVE); + + memcpy (base + SVE_PT_SVE_FPSR_OFFSET (vq), &fpsimd->fpsr, + sizeof (uint32_t)); + memcpy (base + SVE_PT_SVE_FPCR_OFFSET (vq), &fpsimd->fpcr, + sizeof (uint32_t)); + + for (int i = AARCH64_SVE_Z_REGS_NUM - 1; i >= 0 ; i--) + { + memcpy (base + SVE_PT_SVE_ZREG_OFFSET (vq, i), &fpsimd->vregs[i], + sizeof (__int128_t)); + } - memcpy (base + SVE_PT_SVE_FPSR_OFFSET (vq), &fpsimd->fpsr, - sizeof (uint32_t)); - memcpy (base + SVE_PT_SVE_FPCR_OFFSET (vq), &fpsimd->fpcr, - sizeof (uint32_t)); + /* At this point we have converted the FPSIMD layout to an SVE + layout and copied the register data. - for (int i = AARCH64_SVE_Z_REGS_NUM; i >= 0 ; i--) - { - memcpy (base + SVE_PT_SVE_ZREG_OFFSET (vq, i), &fpsimd->vregs[i], - sizeof (__int128_t)); + Fall through so we can update the thread's contents with the SVE + register cache values. */ } } + else + { + /* We already have SVE state for this thread, so we just need to update + the values of the registers. */ + for (int i = 0; i < AARCH64_SVE_Z_REGS_NUM; i++) + if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_Z0_REGNUM + + i)) + reg_buf->raw_collect (AARCH64_SVE_Z0_REGNUM + i, + base + SVE_PT_SVE_ZREG_OFFSET (vq, i)); - /* Replace the kernel values with those from reg_buf. */ - - for (int i = 0; i < AARCH64_SVE_Z_REGS_NUM; i++) - if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_Z0_REGNUM + i)) - reg_buf->raw_collect (AARCH64_SVE_Z0_REGNUM + i, - base + SVE_PT_SVE_ZREG_OFFSET (vq, i)); - - for (int i = 0; i < AARCH64_SVE_P_REGS_NUM; i++) - if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_P0_REGNUM + i)) - reg_buf->raw_collect (AARCH64_SVE_P0_REGNUM + i, - base + SVE_PT_SVE_PREG_OFFSET (vq, i)); + for (int i = 0; i < AARCH64_SVE_P_REGS_NUM; i++) + if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_P0_REGNUM + + i)) + reg_buf->raw_collect (AARCH64_SVE_P0_REGNUM + i, + base + SVE_PT_SVE_PREG_OFFSET (vq, i)); + + if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_FFR_REGNUM)) + reg_buf->raw_collect (AARCH64_SVE_FFR_REGNUM, + base + SVE_PT_SVE_FFR_OFFSET (vq)); + if (REG_VALID == reg_buf->get_register_status (AARCH64_FPSR_REGNUM)) + reg_buf->raw_collect (AARCH64_FPSR_REGNUM, + base + SVE_PT_SVE_FPSR_OFFSET (vq)); + if (REG_VALID == reg_buf->get_register_status (AARCH64_FPCR_REGNUM)) + reg_buf->raw_collect (AARCH64_FPCR_REGNUM, + base + SVE_PT_SVE_FPCR_OFFSET (vq)); + } - if (REG_VALID == reg_buf->get_register_status (AARCH64_SVE_FFR_REGNUM)) - reg_buf->raw_collect (AARCH64_SVE_FFR_REGNUM, - base + SVE_PT_SVE_FFR_OFFSET (vq)); - if (REG_VALID == reg_buf->get_register_status (AARCH64_FPSR_REGNUM)) - reg_buf->raw_collect (AARCH64_FPSR_REGNUM, - base + SVE_PT_SVE_FPSR_OFFSET (vq)); - if (REG_VALID == reg_buf->get_register_status (AARCH64_FPCR_REGNUM)) - reg_buf->raw_collect (AARCH64_FPCR_REGNUM, - base + SVE_PT_SVE_FPCR_OFFSET (vq)); + /* At this point we have collected all the data from the register cache and + we are ready to update the SVE/FPSIMD register contents of the thread. + sve_state should contain all the data in the correct format, ready to be + passed on to ptrace. */ + aarch64_store_sve_regset (tid, new_state); } diff --git a/gdb/nat/aarch64-scalable-linux-ptrace.h b/gdb/nat/aarch64-scalable-linux-ptrace.h index 2847c4e0263..167782c493d 100644 --- a/gdb/nat/aarch64-scalable-linux-ptrace.h +++ b/gdb/nat/aarch64-scalable-linux-ptrace.h @@ -52,22 +52,27 @@ uint64_t aarch64_sve_get_vq (int tid); bool aarch64_sve_set_vq (int tid, uint64_t vq); bool aarch64_sve_set_vq (int tid, struct reg_buffer_common *reg_buf); -/* Read the current SVE register set using ptrace, allocating space as - required. */ +/* Read the current SVE register set from thread TID and return its data + through a byte vector. */ -extern std::unique_ptr aarch64_sve_get_sveregs (int tid); +extern gdb::byte_vector aarch64_fetch_sve_regset (int tid); -/* Put the registers from linux structure buf into register buffer. Assumes the - vector lengths in the register buffer match the size in the kernel. */ +/* Write the SVE contents from SVE_STATE to thread TID. */ -extern void aarch64_sve_regs_copy_to_reg_buf (struct reg_buffer_common *reg_buf, - const void *buf); +extern void +aarch64_store_sve_regset (int tid, const gdb::byte_vector &sve_state); + +/* Given a thread id TID and a register buffer REG_BUF, update the register + buffer with the SVE state from thread TID. */ + +extern void +aarch64_sve_regs_copy_to_reg_buf (int tid, struct reg_buffer_common *reg_buf); -/* Put the registers from register buffer into linux structure buf. Assumes the - vector lengths in the register buffer match the size in the kernel. */ +/* Given a thread id TID and a register buffer REG_BUF containing SVE + register data, write the SVE data to thread TID. */ extern void -aarch64_sve_regs_copy_from_reg_buf (const struct reg_buffer_common *reg_buf, - void *buf); +aarch64_sve_regs_copy_from_reg_buf (int tid, + struct reg_buffer_common *reg_buf); #endif /* NAT_AARCH64_SCALABLE_LINUX_PTRACE_H */ diff --git a/gdbserver/linux-aarch64-low.cc b/gdbserver/linux-aarch64-low.cc index 8b22f1979d6..7c633c212f3 100644 --- a/gdbserver/linux-aarch64-low.cc +++ b/gdbserver/linux-aarch64-low.cc @@ -719,9 +719,18 @@ aarch64_target::low_new_fork (process_info *parent, /* Wrapper for aarch64_sve_regs_copy_to_reg_buf. */ static void -aarch64_sve_regs_copy_to_regcache (struct regcache *regcache, const void *buf) +aarch64_sve_regs_copy_to_regcache (struct regcache *regcache, + ATTRIBUTE_UNUSED const void *buf) { - return aarch64_sve_regs_copy_to_reg_buf (regcache, buf); + /* BUF is unused here since we collect the data straight from a ptrace + request in aarch64_sve_regs_copy_to_reg_buf, therefore bypassing + gdbserver's own call to ptrace. */ + + int tid = lwpid_of (current_thread); + + /* Update the register cache. aarch64_sve_regs_copy_to_reg_buf handles + fetching the NT_ARM_SVE state from thread TID. */ + aarch64_sve_regs_copy_to_reg_buf (tid, regcache); } /* Wrapper for aarch64_sve_regs_copy_from_reg_buf. */ @@ -729,7 +738,16 @@ aarch64_sve_regs_copy_to_regcache (struct regcache *regcache, const void *buf) static void aarch64_sve_regs_copy_from_regcache (struct regcache *regcache, void *buf) { - return aarch64_sve_regs_copy_from_reg_buf (regcache, buf); + int tid = lwpid_of (current_thread); + + /* Update the thread SVE state. aarch64_sve_regs_copy_from_reg_buf + handles writing the SVE/FPSIMD state back to thread TID. */ + aarch64_sve_regs_copy_from_reg_buf (tid, regcache); + + /* We need to return the expected data in BUF, so copy whatever the kernel + already has to BUF. */ + gdb::byte_vector sve_state = aarch64_fetch_sve_regset (tid); + memcpy (buf, sve_state.data (), sve_state.size ()); } /* Array containing all the possible register sets for AArch64/Linux. During -- 2.25.1