From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.6 required=3.0 tests=DKIM_INVALID,DKIM_SIGNED, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E99FEC3A59E for ; Sat, 24 Aug 2019 21:39:11 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id A1D682133F for ; Sat, 24 Aug 2019 21:39:11 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (2048-bit key) header.d=linaro.org header.i=@linaro.org header.b="rgeFwQ2v" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org A1D682133F Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linaro.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:39524 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1i1dkc-0004W5-Op for qemu-devel@archiver.kernel.org; Sat, 24 Aug 2019 17:39:10 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:36365) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1i1dgg-0001M2-Hq for qemu-devel@nongnu.org; Sat, 24 Aug 2019 17:35:08 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1i1dge-00035e-Hm for qemu-devel@nongnu.org; Sat, 24 Aug 2019 17:35:06 -0400 Received: from mail-pf1-x443.google.com ([2607:f8b0:4864:20::443]:39704) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1i1dgc-00032p-HK for qemu-devel@nongnu.org; Sat, 24 Aug 2019 17:35:04 -0400 Received: by mail-pf1-x443.google.com with SMTP id y200so1357593pfb.6 for ; Sat, 24 Aug 2019 14:35:01 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=vNEsJcB2kukOuHiyOwZ4ZNcbnZh0P4uvd68CJElB0UM=; b=rgeFwQ2v34ZzDuPDQgHGQsCpSDKPRgIiGSdMipmsbSX+aheUfiumGIqtXKt/j1LaSf W2HHQO/ZHzIZfwYv8VewLyISpxPAcKFuEMH/x692np2eGFG2C2+jm/lquf+SeePGzDiX wKNAChlY4dPRgW9SnrtiMCb00HORwYNqBIbua7c2ZJReJte7u7S9Hvh9L+2+wzrwGVNV ol3Kd+Hltt5amD3xjyCDy18eKAZ3hXPdaNVfpHnjRSvzaYOT/epR51Ww9lw1nQPvvrkq 2a++6bmxMAdWdQSjBa3mHW4KZ6r3Mzuskn575G/4NLJ6DDzBdsS8rq6vv78/K6JkZ1cX UWOQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references; bh=vNEsJcB2kukOuHiyOwZ4ZNcbnZh0P4uvd68CJElB0UM=; b=tajVj3ra7S7ZRNsrXe+Bu6KeUfKRyQZ30i2nJAHNCmlzteY3psDXT4V6G61HPhjhvb 7YPMgBGVMxNk+J0s9pKd/snjps3n9WFvddJuC33Nz9g0xzj0daTvE77GN1QyYoK0IeHo zrqR2OfbbqA2Fiv7qY6bXBMUpNxI6PHGdRQgxsfA6pShMUikWgUdY+RimaP1Lr41fCFN tJPipw8AEdy1+a48DGrzgtZH+ilvfvQCexZNkIjjr+wEUXq2uovRl+KCSm7NPUEfGn6c UfLr3DN+AZ/7gGKfkwAZ6hgQDftNvz+onVrKILDNrcdONBrFKVOLaYAKqwYZoj6wn8kU 5IXw== X-Gm-Message-State: APjAAAW0bxC9iUZBej2Wn83220e/4Ki9AxhW/r8f3EA6xJwuc++7+sBq oGXIh289bL1JQxHn6FEsgKaefWFYr5o= X-Google-Smtp-Source: APXvYqx8451f+uKW+ygQwXO/bVu2qRkRBYiGVddMPsbSjnqsm26zphtmbh8+mu97O89ph1zrqVY5nw== X-Received: by 2002:a17:90a:3266:: with SMTP id k93mr12250768pjb.46.1566682500442; Sat, 24 Aug 2019 14:35:00 -0700 (PDT) Received: from localhost.localdomain (97-113-7-119.tukw.qwest.net. [97.113.7.119]) by smtp.gmail.com with ESMTPSA id 81sm6325753pfx.111.2019.08.24.14.34.59 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Sat, 24 Aug 2019 14:34:59 -0700 (PDT) From: Richard Henderson To: qemu-devel@nongnu.org Date: Sat, 24 Aug 2019 14:34:50 -0700 Message-Id: <20190824213451.31118-6-richard.henderson@linaro.org> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190824213451.31118-1-richard.henderson@linaro.org> References: <20190824213451.31118-1-richard.henderson@linaro.org> X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 2607:f8b0:4864:20::443 Subject: [Qemu-devel] [PATCH 5/6] cputlb: Handle watchpoints via TLB_WATCHPOINT X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: peter.maydell@linaro.org, david@redhat.com Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" The raising of exceptions from check_watchpoint, buried inside of the I/O subsystem, is fundamentally broken. We do not have the helper return address with which we can unwind guest state. Replace PHYS_SECTION_WATCH and io_mem_watch with TLB_WATCHPOINT. Move the call to cpu_check_watchpoint into the cputlb helpers where we do have the helper return address. This also allows us to handle watchpoints on RAM to bypass the full i/o access path. Signed-off-by: Richard Henderson --- include/exec/cpu-all.h | 5 +- accel/tcg/cputlb.c | 83 +++++++++++++++++++++++++++--- exec.c | 114 +++-------------------------------------- 3 files changed, 87 insertions(+), 115 deletions(-) diff --git a/include/exec/cpu-all.h b/include/exec/cpu-all.h index 8d07ae23a5..d2d443c4f9 100644 --- a/include/exec/cpu-all.h +++ b/include/exec/cpu-all.h @@ -329,11 +329,14 @@ CPUArchState *cpu_copy(CPUArchState *env); #define TLB_NOTDIRTY (1 << (TARGET_PAGE_BITS - 2)) /* Set if TLB entry is an IO callback. */ #define TLB_MMIO (1 << (TARGET_PAGE_BITS - 3)) +/* Set if TLB entry contains a watchpoint. */ +#define TLB_WATCHPOINT (1 << (TARGET_PAGE_BITS - 4)) /* Use this mask to check interception with an alignment mask * in a TCG backend. */ -#define TLB_FLAGS_MASK (TLB_INVALID_MASK | TLB_NOTDIRTY | TLB_MMIO) +#define TLB_FLAGS_MASK \ + (TLB_INVALID_MASK | TLB_NOTDIRTY | TLB_MMIO | TLB_WATCHPOINT) /** * tlb_hit_page: return true if page aligned @addr is a hit against the diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index c9576bebcf..f7a414a131 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -710,6 +710,7 @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, hwaddr iotlb, xlat, sz, paddr_page; target_ulong vaddr_page; int asidx = cpu_asidx_from_attrs(cpu, attrs); + int wp_flags; assert_cpu_is_self(cpu); @@ -752,6 +753,8 @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, code_address = address; iotlb = memory_region_section_get_iotlb(cpu, section, vaddr_page, paddr_page, xlat, prot, &address); + wp_flags = cpu_watchpoint_address_matches(cpu, vaddr_page, + TARGET_PAGE_SIZE); index = tlb_index(env, mmu_idx, vaddr_page); te = tlb_entry(env, mmu_idx, vaddr_page); @@ -805,6 +808,9 @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, tn.addend = addend - vaddr_page; if (prot & PAGE_READ) { tn.addr_read = address; + if (wp_flags & BP_MEM_READ) { + tn.addr_read |= TLB_WATCHPOINT; + } } else { tn.addr_read = -1; } @@ -831,6 +837,9 @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, if (prot & PAGE_WRITE_INV) { tn.addr_write |= TLB_INVALID_MASK; } + if (wp_flags & BP_MEM_WRITE) { + tn.addr_write |= TLB_WATCHPOINT; + } } copy_tlb_helper_locked(te, &tn); @@ -1264,13 +1273,33 @@ load_helper(CPUArchState *env, target_ulong addr, TCGMemOpIdx oi, tlb_addr &= ~TLB_INVALID_MASK; } - /* Handle an IO access. */ + /* Handle anything that isn't just a straight memory access. */ if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) { + CPUIOTLBEntry *iotlbentry; + + /* For anything that is unaligned, recurse through full_load. */ if ((addr & (size - 1)) != 0) { goto do_unaligned_access; } - return io_readx(env, &env_tlb(env)->d[mmu_idx].iotlb[index], - mmu_idx, addr, retaddr, access_type, op); + + iotlbentry = &env_tlb(env)->d[mmu_idx].iotlb[index]; + + /* Handle watchpoints. */ + if (unlikely(tlb_addr & TLB_WATCHPOINT)) { + /* On watchpoint hit, this will longjmp out. */ + cpu_check_watchpoint(env_cpu(env), addr, size, + iotlbentry->attrs, BP_MEM_READ, retaddr); + + /* The backing page may or may not require I/O. */ + tlb_addr &= ~TLB_WATCHPOINT; + if ((tlb_addr & ~TARGET_PAGE_MASK) == 0) { + goto do_aligned_access; + } + } + + /* Handle I/O access. */ + return io_readx(env, iotlbentry, mmu_idx, addr, + retaddr, access_type, op); } /* Handle slow unaligned access (it spans two pages or IO). */ @@ -1297,6 +1326,7 @@ load_helper(CPUArchState *env, target_ulong addr, TCGMemOpIdx oi, return res & MAKE_64BIT_MASK(0, size * 8); } + do_aligned_access: haddr = (void *)((uintptr_t)addr + entry->addend); switch (op) { case MO_UB: @@ -1486,13 +1516,32 @@ store_helper(CPUArchState *env, target_ulong addr, uint64_t val, tlb_addr = tlb_addr_write(entry) & ~TLB_INVALID_MASK; } - /* Handle an IO access. */ + /* Handle anything that isn't just a straight memory access. */ if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) { + CPUIOTLBEntry *iotlbentry; + + /* For anything that is unaligned, recurse through byte stores. */ if ((addr & (size - 1)) != 0) { goto do_unaligned_access; } - io_writex(env, &env_tlb(env)->d[mmu_idx].iotlb[index], mmu_idx, - val, addr, retaddr, op); + + iotlbentry = &env_tlb(env)->d[mmu_idx].iotlb[index]; + + /* Handle watchpoints. */ + if (unlikely(tlb_addr & TLB_WATCHPOINT)) { + /* On watchpoint hit, this will longjmp out. */ + cpu_check_watchpoint(env_cpu(env), addr, size, + iotlbentry->attrs, BP_MEM_WRITE, retaddr); + + /* The backing page may or may not require I/O. */ + tlb_addr &= ~TLB_WATCHPOINT; + if ((tlb_addr & ~TARGET_PAGE_MASK) == 0) { + goto do_aligned_access; + } + } + + /* Handle I/O access. */ + io_writex(env, iotlbentry, mmu_idx, val, addr, retaddr, op); return; } @@ -1504,6 +1553,8 @@ store_helper(CPUArchState *env, target_ulong addr, uint64_t val, uintptr_t index2; CPUTLBEntry *entry2; target_ulong page2, tlb_addr2; + size_t size2; + do_unaligned_access: /* * Ensure the second page is in the TLB. Note that the first page @@ -1511,16 +1562,33 @@ store_helper(CPUArchState *env, target_ulong addr, uint64_t val, * cannot evict the first. */ page2 = (addr + size) & TARGET_PAGE_MASK; + size2 = (addr + size) & ~TARGET_PAGE_MASK; index2 = tlb_index(env, mmu_idx, page2); entry2 = tlb_entry(env, mmu_idx, page2); tlb_addr2 = tlb_addr_write(entry2); if (!tlb_hit_page(tlb_addr2, page2) && !victim_tlb_hit(env, mmu_idx, index2, tlb_off, page2 & TARGET_PAGE_MASK)) { - tlb_fill(env_cpu(env), page2, size, MMU_DATA_STORE, + tlb_fill(env_cpu(env), page2, size2, MMU_DATA_STORE, mmu_idx, retaddr); } + /* + * Handle watchpoints. Since this may trap, all checks + * must happen before any store. + */ + if (unlikely(tlb_addr & TLB_WATCHPOINT)) { + cpu_check_watchpoint(env_cpu(env), addr, + -(addr | TARGET_PAGE_MASK), + env_tlb(env)->d[mmu_idx].iotlb[index].attrs, + BP_MEM_WRITE, retaddr); + } + if (unlikely(tlb_addr2 & TLB_WATCHPOINT)) { + cpu_check_watchpoint(env_cpu(env), page2, size2, + env_tlb(env)->d[mmu_idx].iotlb[index2].attrs, + BP_MEM_WRITE, retaddr); + } + /* * XXX: not efficient, but simple. * This loop must go in the forward direction to avoid issues @@ -1540,6 +1608,7 @@ store_helper(CPUArchState *env, target_ulong addr, uint64_t val, return; } + do_aligned_access: haddr = (void *)((uintptr_t)addr + entry->addend); switch (op) { case MO_UB: diff --git a/exec.c b/exec.c index 8575ce51ad..ad0f4a598f 100644 --- a/exec.c +++ b/exec.c @@ -193,15 +193,12 @@ typedef struct subpage_t { #define PHYS_SECTION_UNASSIGNED 0 #define PHYS_SECTION_NOTDIRTY 1 #define PHYS_SECTION_ROM 2 -#define PHYS_SECTION_WATCH 3 static void io_mem_init(void); static void memory_map_init(void); static void tcg_log_global_after_sync(MemoryListener *listener); static void tcg_commit(MemoryListener *listener); -static MemoryRegion io_mem_watch; - /** * CPUAddressSpace: all the information a CPU needs about an AddressSpace * @cpu: the CPU whose AddressSpace this is @@ -1472,7 +1469,6 @@ hwaddr memory_region_section_get_iotlb(CPUState *cpu, target_ulong *address) { hwaddr iotlb; - int flags, match; if (memory_region_is_ram(section->mr)) { /* Normal RAM. */ @@ -1490,19 +1486,6 @@ hwaddr memory_region_section_get_iotlb(CPUState *cpu, iotlb += xlat; } - /* Avoid trapping reads of pages with a write breakpoint. */ - match = (prot & PAGE_READ ? BP_MEM_READ : 0) - | (prot & PAGE_WRITE ? BP_MEM_WRITE : 0); - flags = cpu_watchpoint_address_matches(cpu, vaddr, TARGET_PAGE_SIZE); - if (flags & match) { - /* - * Make accesses to pages with watchpoints go via the - * watchpoint trap routines. - */ - iotlb = PHYS_SECTION_WATCH + paddr; - *address |= TLB_MMIO; - } - return iotlb; } #endif /* defined(CONFIG_USER_ONLY) */ @@ -2810,10 +2793,14 @@ void cpu_check_watchpoint(CPUState *cpu, vaddr addr, vaddr len, assert(tcg_enabled()); if (cpu->watchpoint_hit) { - /* We re-entered the check after replacing the TB. Now raise - * the debug interrupt so that is will trigger after the - * current instruction. */ + /* + * We re-entered the check after replacing the TB. + * Now raise the debug interrupt so that it will + * trigger after the current instruction. + */ + qemu_mutex_lock_iothread(); cpu_interrupt(cpu, CPU_INTERRUPT_DEBUG); + qemu_mutex_unlock_iothread(); return; } @@ -2858,88 +2845,6 @@ void cpu_check_watchpoint(CPUState *cpu, vaddr addr, vaddr len, } } -static void check_watchpoint(int offset, int len, MemTxAttrs attrs, int flags) -{ - CPUState *cpu = current_cpu; - vaddr addr = (cpu->mem_io_vaddr & TARGET_PAGE_MASK) + offset; - - cpu_check_watchpoint(cpu, addr, len, attrs, flags, 0); -} - -/* Watchpoint access routines. Watchpoints are inserted using TLB tricks, - so these check for a hit then pass through to the normal out-of-line - phys routines. */ -static MemTxResult watch_mem_read(void *opaque, hwaddr addr, uint64_t *pdata, - unsigned size, MemTxAttrs attrs) -{ - MemTxResult res; - uint64_t data; - int asidx = cpu_asidx_from_attrs(current_cpu, attrs); - AddressSpace *as = current_cpu->cpu_ases[asidx].as; - - check_watchpoint(addr & ~TARGET_PAGE_MASK, size, attrs, BP_MEM_READ); - switch (size) { - case 1: - data = address_space_ldub(as, addr, attrs, &res); - break; - case 2: - data = address_space_lduw(as, addr, attrs, &res); - break; - case 4: - data = address_space_ldl(as, addr, attrs, &res); - break; - case 8: - data = address_space_ldq(as, addr, attrs, &res); - break; - default: abort(); - } - *pdata = data; - return res; -} - -static MemTxResult watch_mem_write(void *opaque, hwaddr addr, - uint64_t val, unsigned size, - MemTxAttrs attrs) -{ - MemTxResult res; - int asidx = cpu_asidx_from_attrs(current_cpu, attrs); - AddressSpace *as = current_cpu->cpu_ases[asidx].as; - - check_watchpoint(addr & ~TARGET_PAGE_MASK, size, attrs, BP_MEM_WRITE); - switch (size) { - case 1: - address_space_stb(as, addr, val, attrs, &res); - break; - case 2: - address_space_stw(as, addr, val, attrs, &res); - break; - case 4: - address_space_stl(as, addr, val, attrs, &res); - break; - case 8: - address_space_stq(as, addr, val, attrs, &res); - break; - default: abort(); - } - return res; -} - -static const MemoryRegionOps watch_mem_ops = { - .read_with_attrs = watch_mem_read, - .write_with_attrs = watch_mem_write, - .endianness = DEVICE_NATIVE_ENDIAN, - .valid = { - .min_access_size = 1, - .max_access_size = 8, - .unaligned = false, - }, - .impl = { - .min_access_size = 1, - .max_access_size = 8, - .unaligned = false, - }, -}; - static MemTxResult flatview_read(FlatView *fv, hwaddr addr, MemTxAttrs attrs, uint8_t *buf, hwaddr len); static MemTxResult flatview_write(FlatView *fv, hwaddr addr, MemTxAttrs attrs, @@ -3115,9 +3020,6 @@ static void io_mem_init(void) memory_region_init_io(&io_mem_notdirty, NULL, ¬dirty_mem_ops, NULL, NULL, UINT64_MAX); memory_region_clear_global_locking(&io_mem_notdirty); - - memory_region_init_io(&io_mem_watch, NULL, &watch_mem_ops, NULL, - NULL, UINT64_MAX); } AddressSpaceDispatch *address_space_dispatch_new(FlatView *fv) @@ -3131,8 +3033,6 @@ AddressSpaceDispatch *address_space_dispatch_new(FlatView *fv) assert(n == PHYS_SECTION_NOTDIRTY); n = dummy_section(&d->map, fv, &io_mem_rom); assert(n == PHYS_SECTION_ROM); - n = dummy_section(&d->map, fv, &io_mem_watch); - assert(n == PHYS_SECTION_WATCH); d->phys_map = (PhysPageEntry) { .ptr = PHYS_MAP_NODE_NIL, .skip = 1 }; -- 2.17.1