April 2018 - Linux-stable-mirror

FAILED: patch "[PATCH] drm/amdgpu: fix prime teardown order" failed to apply to 4.16-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.16-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 7f8fb919655ef3011b01bb114fef42107957860a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= <christian.koenig(a)amd.com> Date: Fri, 9 Mar 2018 14:42:54 +0100 Subject: [PATCH] drm/amdgpu: fix prime teardown order MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We unmapped imported DMA-bufs when the GEM handle was dropped, not when the hardware was done with the buffere. Signed-off-by: Christian König <christian.koenig(a)amd.com> Reviewed-by: Michel Dänzer <michel.daenzer(a)amd.com> CC: stable(a)vger.kernel.org Signed-off-by: Alex Deucher <alexander.deucher(a)amd.com> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index bb9b21266b67..46b9ea4e6103 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -36,8 +36,6 @@ void amdgpu_gem_object_free(struct drm_gem_object *gobj) struct amdgpu_bo *robj = gem_to_amdgpu_bo(gobj); if (robj) { - if (robj->gem_base.import_attach) - drm_prime_gem_destroy(&robj->gem_base, robj->tbo.sg); amdgpu_mn_unregister(robj); amdgpu_bo_unref(&robj); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index ec6ec1f8a085..fc720603b970 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -60,6 +60,8 @@ static void amdgpu_ttm_bo_destroy(struct ttm_buffer_object *tbo) amdgpu_bo_kunmap(bo); + if (bo->gem_base.import_attach) + drm_prime_gem_destroy(&bo->gem_base, bo->tbo.sg); drm_gem_object_release(&bo->gem_base); amdgpu_bo_unref(&bo->parent); if (!list_empty(&bo->shadow_list)) {

6 years, 9 months

1
0
0 0

FAILED: patch "[PATCH] drm/amdgpu: fix KV harvesting" failed to apply to 4.16-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.16-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From a97fc4e4524cda15e0176ee8853038f6b3920a9a Mon Sep 17 00:00:00 2001 From: Alex Deucher <alexander.deucher(a)amd.com> Date: Thu, 1 Mar 2018 11:05:31 -0500 Subject: [PATCH] drm/amdgpu: fix KV harvesting MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Always set the graphics values to the max for the asic type. E.g., some 1 RB chips are actually 1 RB chips, others are actually harvested 2 RB chips. Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=99353 Reviewed-by: Christian König <christian.koenig(a)amd.com> Signed-off-by: Alex Deucher <alexander.deucher(a)amd.com> Cc: stable(a)vger.kernel.org diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 972d421caada..e13d9d83767b 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -4358,34 +4358,8 @@ static void gfx_v7_0_gpu_early_init(struct amdgpu_device *adev) case CHIP_KAVERI: adev->gfx.config.max_shader_engines = 1; adev->gfx.config.max_tile_pipes = 4; - if ((adev->pdev->device == 0x1304) || - (adev->pdev->device == 0x1305) || - (adev->pdev->device == 0x130C) || - (adev->pdev->device == 0x130F) || - (adev->pdev->device == 0x1310) || - (adev->pdev->device == 0x1311) || - (adev->pdev->device == 0x131C)) { - adev->gfx.config.max_cu_per_sh = 8; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1309) || - (adev->pdev->device == 0x130A) || - (adev->pdev->device == 0x130D) || - (adev->pdev->device == 0x1313) || - (adev->pdev->device == 0x131D)) { - adev->gfx.config.max_cu_per_sh = 6; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1306) || - (adev->pdev->device == 0x1307) || - (adev->pdev->device == 0x130B) || - (adev->pdev->device == 0x130E) || - (adev->pdev->device == 0x1315) || - (adev->pdev->device == 0x131B)) { - adev->gfx.config.max_cu_per_sh = 4; - adev->gfx.config.max_backends_per_se = 1; - } else { - adev->gfx.config.max_cu_per_sh = 3; - adev->gfx.config.max_backends_per_se = 1; - } + adev->gfx.config.max_cu_per_sh = 8; + adev->gfx.config.max_backends_per_se = 2; adev->gfx.config.max_sh_per_se = 1; adev->gfx.config.max_texture_channel_caches = 4; adev->gfx.config.max_gprs = 256;

6 years, 9 months

1
0
0 0

ext4 patches for 4.4.y

by Nathan Chancellor

Hi Greg and Ted, I've been looking at ext4 history the past couple of days seeing if the patches that you attempted to apply from 4.17-rc1 were relevant and I noticed a couple from previous 4.x versions that seem like they should be applied here, as they are clean picks and tagged for stable: 74dae4278546 ("ext4: fix crashes in dioread_nolock mode") c755e251357a ("ext4: fix deadlock between inline_data and ext4_expand_extra_isize_ea()") I've been running them for the day and not had any issues (though I didn't have any before). If there are any objections, let me know! Thanks! Nathan

6 years, 9 months

2
2
0 0

[PATCH v3 3.18.y 0/3] 4.17-rc1 stable tagged ext4 patches

by Harsh Shandilya

These are all the ext4 patches that were tagged for -stable and failed to apply to 3.18.y. Patch e40ff2138985 ("ext4: force revalidation of directory pointer after seekdir(2)") was Cc'd to stable as well but it requires commmit ae5e165d855d ("fs: new API for handling inode->i_version") to be applied as well which is neither a stable candidate nor under 100 lines so I've skipped e40ff2138985. If somebody can suggest a backport of the commit which doesn't require ae5e165d855d, I'll be glad. Theodore Ts'o (3): ext4: add validity checks for bitmap block numbers ext4: fail ext4_iget for root directory if unallocated ext4: don't allow r/w mounts if metadata blocks overlap the superblock fs/ext4/balloc.c | 16 ++++++++++++++-- fs/ext4/ialloc.c | 8 +++++++- fs/ext4/inode.c | 6 ++++++ fs/ext4/super.c | 6 ++++++ 4 files changed, 33 insertions(+), 3 deletions(-) -- 2.15.0.2308.g658a28aa74af

6 years, 9 months

2
4
0 0

Please apply commit cf0d53ba4947 and 523184972b28 to v4.4.y, v4.9.y, v4.14.y

by Sinan Kaya

Hi Greg, Upstream commit cf0d53ba4947 ("vfio/pci: Virtualize Maximum Read Request Size") and commit 523184972b28 ("vfio/pci: Virtualize Maximum Payload Size") fixes nasty PCIe virtualization issues for platforms that support Maximum Payload Size bigger than 128. Issue shows up when a device is assigned to the guest machine as a passthrough. Guest machine configures the MPS/MRRS settings to values that are incompatible with the parent bridge device. This causes PCIe transaction timeouts and AER errors to be spilled in the host kernel. Please apply commit cf0d53ba4947 and 523184972b28 to all affected releases to fix the resulting regression. Thanks, Sinan -- Sinan Kaya Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc. Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.

6 years, 9 months

3
7
0 0

FAILED: patch "[PATCH] ALSA: pcm: Avoid potential races between OSS ioctls and" failed to apply to 4.14-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 4.14-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From 02a5d6925cd34c3b774bdb8eefb057c40a30e870 Mon Sep 17 00:00:00 2001 From: Takashi Iwai <tiwai(a)suse.de> Date: Thu, 22 Mar 2018 18:10:14 +0100 Subject: [PATCH] ALSA: pcm: Avoid potential races between OSS ioctls and read/write Although we apply the params_lock mutex to the whole read and write operations as well as snd_pcm_oss_change_params(), we may still face some races. First off, the params_lock is taken inside the read and write loop. This is intentional for avoiding the too long locking, but it allows the in-between parameter change, which might lead to invalid pointers. We check the readiness of the stream and set up via snd_pcm_oss_make_ready() at the beginning of read and write, but it's called only once, by assuming that it remains ready in the rest. Second, many ioctls that may change the actual parameters (i.e. setting runtime->oss.params=1) aren't protected, hence they can be processed in a half-baked state. This patch is an attempt to plug these holes. The stream readiness check is moved inside the read/write inner loop, so that the stream is always set up in a proper state before further processing. Also, each ioctl that may change the parameter is wrapped with the params_lock for avoiding the races. The issues were triggered by syzkaller in a few different scenarios, particularly the one below appearing as GPF in loopback_pos_update. Reported-by: syzbot+c4227aec125487ec3efa(a)syzkaller.appspotmail.com Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> diff --git a/sound/core/oss/pcm_oss.c b/sound/core/oss/pcm_oss.c index 02298c9c6020..f8bfee8022e0 100644 --- a/sound/core/oss/pcm_oss.c +++ b/sound/core/oss/pcm_oss.c @@ -823,8 +823,8 @@ static int choose_rate(struct snd_pcm_substream *substream, return snd_pcm_hw_param_near(substream, params, SNDRV_PCM_HW_PARAM_RATE, best_rate, NULL); } -static int snd_pcm_oss_change_params(struct snd_pcm_substream *substream, - bool trylock) +/* call with params_lock held */ +static int snd_pcm_oss_change_params_locked(struct snd_pcm_substream *substream) { struct snd_pcm_runtime *runtime = substream->runtime; struct snd_pcm_hw_params *params, *sparams; @@ -838,11 +838,8 @@ static int snd_pcm_oss_change_params(struct snd_pcm_substream *substream, const struct snd_mask *sformat_mask; struct snd_mask mask; - if (trylock) { - if (!(mutex_trylock(&runtime->oss.params_lock))) - return -EAGAIN; - } else if (mutex_lock_interruptible(&runtime->oss.params_lock)) - return -ERESTARTSYS; + if (!runtime->oss.params) + return 0; sw_params = kzalloc(sizeof(*sw_params), GFP_KERNEL); params = kmalloc(sizeof(*params), GFP_KERNEL); sparams = kmalloc(sizeof(*sparams), GFP_KERNEL); @@ -1068,6 +1065,23 @@ static int snd_pcm_oss_change_params(struct snd_pcm_substream *substream, kfree(sw_params); kfree(params); kfree(sparams); + return err; +} + +/* this one takes the lock by itself */ +static int snd_pcm_oss_change_params(struct snd_pcm_substream *substream, + bool trylock) +{ + struct snd_pcm_runtime *runtime = substream->runtime; + int err; + + if (trylock) { + if (!(mutex_trylock(&runtime->oss.params_lock))) + return -EAGAIN; + } else if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; + + err = snd_pcm_oss_change_params_locked(substream); mutex_unlock(&runtime->oss.params_lock); return err; } @@ -1096,11 +1110,14 @@ static int snd_pcm_oss_get_active_substream(struct snd_pcm_oss_file *pcm_oss_fil return 0; } +/* call with params_lock held */ static int snd_pcm_oss_prepare(struct snd_pcm_substream *substream) { int err; struct snd_pcm_runtime *runtime = substream->runtime; + if (!runtime->oss.prepare) + return 0; err = snd_pcm_kernel_ioctl(substream, SNDRV_PCM_IOCTL_PREPARE, NULL); if (err < 0) { pcm_dbg(substream->pcm, @@ -1120,14 +1137,35 @@ static int snd_pcm_oss_make_ready(struct snd_pcm_substream *substream) struct snd_pcm_runtime *runtime; int err; - if (substream == NULL) - return 0; runtime = substream->runtime; if (runtime->oss.params) { err = snd_pcm_oss_change_params(substream, false); if (err < 0) return err; } + if (runtime->oss.prepare) { + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; + err = snd_pcm_oss_prepare(substream); + mutex_unlock(&runtime->oss.params_lock); + if (err < 0) + return err; + } + return 0; +} + +/* call with params_lock held */ +static int snd_pcm_oss_make_ready_locked(struct snd_pcm_substream *substream) +{ + struct snd_pcm_runtime *runtime; + int err; + + runtime = substream->runtime; + if (runtime->oss.params) { + err = snd_pcm_oss_change_params_locked(substream); + if (err < 0) + return err; + } if (runtime->oss.prepare) { err = snd_pcm_oss_prepare(substream); if (err < 0) @@ -1332,13 +1370,14 @@ static ssize_t snd_pcm_oss_write1(struct snd_pcm_substream *substream, const cha if (atomic_read(&substream->mmap_count)) return -ENXIO; - if ((tmp = snd_pcm_oss_make_ready(substream)) < 0) - return tmp; while (bytes > 0) { if (mutex_lock_interruptible(&runtime->oss.params_lock)) { tmp = -ERESTARTSYS; break; } + tmp = snd_pcm_oss_make_ready_locked(substream); + if (tmp < 0) + goto err; if (bytes < runtime->oss.period_bytes || runtime->oss.buffer_used > 0) { tmp = bytes; if (tmp + runtime->oss.buffer_used > runtime->oss.period_bytes) @@ -1439,13 +1478,14 @@ static ssize_t snd_pcm_oss_read1(struct snd_pcm_substream *substream, char __use if (atomic_read(&substream->mmap_count)) return -ENXIO; - if ((tmp = snd_pcm_oss_make_ready(substream)) < 0) - return tmp; while (bytes > 0) { if (mutex_lock_interruptible(&runtime->oss.params_lock)) { tmp = -ERESTARTSYS; break; } + tmp = snd_pcm_oss_make_ready_locked(substream); + if (tmp < 0) + goto err; if (bytes < runtime->oss.period_bytes || runtime->oss.buffer_used > 0) { if (runtime->oss.buffer_used == 0) { tmp = snd_pcm_oss_read2(substream, runtime->oss.buffer, runtime->oss.period_bytes, 1); @@ -1501,10 +1541,12 @@ static int snd_pcm_oss_reset(struct snd_pcm_oss_file *pcm_oss_file) continue; runtime = substream->runtime; snd_pcm_kernel_ioctl(substream, SNDRV_PCM_IOCTL_DROP, NULL); + mutex_lock(&runtime->oss.params_lock); runtime->oss.prepare = 1; runtime->oss.buffer_used = 0; runtime->oss.prev_hw_ptr_period = 0; runtime->oss.period_ptr = 0; + mutex_unlock(&runtime->oss.params_lock); } return 0; } @@ -1590,9 +1632,10 @@ static int snd_pcm_oss_sync(struct snd_pcm_oss_file *pcm_oss_file) goto __direct; if ((err = snd_pcm_oss_make_ready(substream)) < 0) return err; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; format = snd_pcm_oss_format_from(runtime->oss.format); width = snd_pcm_format_physical_width(format); - mutex_lock(&runtime->oss.params_lock); if (runtime->oss.buffer_used > 0) { #ifdef OSS_DEBUG pcm_dbg(substream->pcm, "sync: buffer_used\n"); @@ -1643,7 +1686,9 @@ static int snd_pcm_oss_sync(struct snd_pcm_oss_file *pcm_oss_file) substream->f_flags = saved_f_flags; if (err < 0) return err; + mutex_lock(&runtime->oss.params_lock); runtime->oss.prepare = 1; + mutex_unlock(&runtime->oss.params_lock); } substream = pcm_oss_file->streams[SNDRV_PCM_STREAM_CAPTURE]; @@ -1654,8 +1699,10 @@ static int snd_pcm_oss_sync(struct snd_pcm_oss_file *pcm_oss_file) err = snd_pcm_kernel_ioctl(substream, SNDRV_PCM_IOCTL_DROP, NULL); if (err < 0) return err; + mutex_lock(&runtime->oss.params_lock); runtime->oss.buffer_used = 0; runtime->oss.prepare = 1; + mutex_unlock(&runtime->oss.params_lock); } return 0; } @@ -1674,10 +1721,13 @@ static int snd_pcm_oss_set_rate(struct snd_pcm_oss_file *pcm_oss_file, int rate) rate = 1000; else if (rate > 192000) rate = 192000; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; if (runtime->oss.rate != rate) { runtime->oss.params = 1; runtime->oss.rate = rate; } + mutex_unlock(&runtime->oss.params_lock); } return snd_pcm_oss_get_rate(pcm_oss_file); } @@ -1705,10 +1755,13 @@ static int snd_pcm_oss_set_channels(struct snd_pcm_oss_file *pcm_oss_file, unsig if (substream == NULL) continue; runtime = substream->runtime; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; if (runtime->oss.channels != channels) { runtime->oss.params = 1; runtime->oss.channels = channels; } + mutex_unlock(&runtime->oss.params_lock); } return snd_pcm_oss_get_channels(pcm_oss_file); } @@ -1794,10 +1847,13 @@ static int snd_pcm_oss_set_format(struct snd_pcm_oss_file *pcm_oss_file, int for if (substream == NULL) continue; runtime = substream->runtime; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; if (runtime->oss.format != format) { runtime->oss.params = 1; runtime->oss.format = format; } + mutex_unlock(&runtime->oss.params_lock); } } return snd_pcm_oss_get_format(pcm_oss_file); @@ -1817,8 +1873,6 @@ static int snd_pcm_oss_set_subdivide1(struct snd_pcm_substream *substream, int s { struct snd_pcm_runtime *runtime; - if (substream == NULL) - return 0; runtime = substream->runtime; if (subdivide == 0) { subdivide = runtime->oss.subdivision; @@ -1842,9 +1896,16 @@ static int snd_pcm_oss_set_subdivide(struct snd_pcm_oss_file *pcm_oss_file, int for (idx = 1; idx >= 0; --idx) { struct snd_pcm_substream *substream = pcm_oss_file->streams[idx]; + struct snd_pcm_runtime *runtime; + if (substream == NULL) continue; - if ((err = snd_pcm_oss_set_subdivide1(substream, subdivide)) < 0) + runtime = substream->runtime; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; + err = snd_pcm_oss_set_subdivide1(substream, subdivide); + mutex_unlock(&runtime->oss.params_lock); + if (err < 0) return err; } return err; @@ -1854,8 +1915,6 @@ static int snd_pcm_oss_set_fragment1(struct snd_pcm_substream *substream, unsign { struct snd_pcm_runtime *runtime; - if (substream == NULL) - return 0; runtime = substream->runtime; if (runtime->oss.subdivision || runtime->oss.fragshift) return -EINVAL; @@ -1875,9 +1934,16 @@ static int snd_pcm_oss_set_fragment(struct snd_pcm_oss_file *pcm_oss_file, unsig for (idx = 1; idx >= 0; --idx) { struct snd_pcm_substream *substream = pcm_oss_file->streams[idx]; + struct snd_pcm_runtime *runtime; + if (substream == NULL) continue; - if ((err = snd_pcm_oss_set_fragment1(substream, val)) < 0) + runtime = substream->runtime; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; + err = snd_pcm_oss_set_fragment1(substream, val); + mutex_unlock(&runtime->oss.params_lock); + if (err < 0) return err; } return err; @@ -1961,6 +2027,9 @@ static int snd_pcm_oss_set_trigger(struct snd_pcm_oss_file *pcm_oss_file, int tr } if (psubstream) { runtime = psubstream->runtime; + cmd = 0; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; if (trigger & PCM_ENABLE_OUTPUT) { if (runtime->oss.trigger) goto _skip1; @@ -1978,13 +2047,19 @@ static int snd_pcm_oss_set_trigger(struct snd_pcm_oss_file *pcm_oss_file, int tr cmd = SNDRV_PCM_IOCTL_DROP; runtime->oss.prepare = 1; } - err = snd_pcm_kernel_ioctl(psubstream, cmd, NULL); - if (err < 0) - return err; - } _skip1: + mutex_unlock(&runtime->oss.params_lock); + if (cmd) { + err = snd_pcm_kernel_ioctl(psubstream, cmd, NULL); + if (err < 0) + return err; + } + } if (csubstream) { runtime = csubstream->runtime; + cmd = 0; + if (mutex_lock_interruptible(&runtime->oss.params_lock)) + return -ERESTARTSYS; if (trigger & PCM_ENABLE_INPUT) { if (runtime->oss.trigger) goto _skip2; @@ -1999,11 +2074,14 @@ static int snd_pcm_oss_set_trigger(struct snd_pcm_oss_file *pcm_oss_file, int tr cmd = SNDRV_PCM_IOCTL_DROP; runtime->oss.prepare = 1; } - err = snd_pcm_kernel_ioctl(csubstream, cmd, NULL); - if (err < 0) - return err; - } _skip2: + mutex_unlock(&runtime->oss.params_lock); + if (cmd) { + err = snd_pcm_kernel_ioctl(csubstream, cmd, NULL); + if (err < 0) + return err; + } + } return 0; }

6 years, 9 months

3
2
0 0

[PATCH 0/4] 4.17-rc1 stable tagged ext4 patches for 3.18.y

by Harsh Shandilya

These are all the ext4 patches that were tagged for -stable and failed to apply to 3.18.y. Side note: Patch e15dc99dbb9c ("ALSA: pcm: Fix endless loop for XRUN recovery in OSS emulation") which was tagged for -stable is not required on 3.18.y so I have skipped the backport. Theodore Ts'o (4): ext4: add validity checks for bitmap block numbers ext4: fail ext4_iget for root directory if unallocated ext4: don't allow r/w mounts if metadata blocks overlap the superblock ext4: force revalidation of directory pointer after seekdir(2) fs/ext4/balloc.c | 16 ++++++++++++++-- fs/ext4/dir.c | 8 +++++--- fs/ext4/ialloc.c | 7 +++++++ fs/ext4/inode.c | 6 ++++++ fs/ext4/super.c | 6 ++++++ 5 files changed, 38 insertions(+), 5 deletions(-) -- 2.15.0.2308.g658a28aa74af

6 years, 9 months

3
13
0 0

[PATCH v6] blk-mq: Avoid that a completion can be ignored for BLK_EH_RESET_TIMER

by Bart Van Assche

The blk-mq timeout handling code ignores completions that occur after blk_mq_check_expired() has been called and before blk_mq_rq_timed_out() has reset rq->aborted_gstate. If a block driver timeout handler always returns BLK_EH_RESET_TIMER then the result will be that the request never terminates. Fix this race as follows: - Use the deadline instead of the request generation to detect whether or not a request timer fired after reinitialization of a request. - Store the request state in the lowest two bits of the deadline instead of the lowest two bits of 'gstate'. - Rename MQ_RQ_STATE_MASK into RQ_STATE_MASK and change it from an enumeration member into a #define such that its type can be changed into unsigned long. That allows to write & ~RQ_STATE_MASK instead of ~(unsigned long)RQ_STATE_MASK. - Remove all request member variables that became superfluous due to this change: gstate, gstate_seq and aborted_gstate_sync. - Remove the request state information that became superfluous due to this patch, namely RQF_MQ_TIMEOUT_EXPIRED. - Remove the code that became superfluous due to this change, namely the RCU lock and unlock statements in blk_mq_complete_request() and also the synchronize_rcu() call in the timeout handler. Signed-off-by: Bart Van Assche <bart.vanassche(a)wdc.com> Cc: Tejun Heo <tj(a)kernel.org> Cc: Christoph Hellwig <hch(a)lst.de> Cc: Ming Lei <ming.lei(a)redhat.com> Cc: Sagi Grimberg <sagi(a)grimberg.me> Cc: Israel Rukshin <israelr(a)mellanox.com>, Cc: Max Gurtovoy <maxg(a)mellanox.com> Cc: <stable(a)vger.kernel.org> # v4.16 --- Changes compared to v5: - Restored the synchronize_rcu() call between marking a request for timeout handling and the actual timeout handling to avoid that timeout handling starts while .queue_rq() is still in progress if the timeout is very short. - Only use cmpxchg() if another context could attempt to change the request state concurrently. Use WRITE_ONCE() otherwise. Changes compared to v4: - Addressed multiple review comments from Christoph. The most important are that atomic_long_cmpxchg() has been changed into cmpxchg() and also that there is now a nice and clean split between the legacy and blk-mq versions of blk_add_timer(). - Changed the patch name and modified the patch description because there is disagreement about whether or not the v4.16 blk-mq core can complete a single request twice. Kept the "Cc: stable" tag because of https://bugzilla.kernel.org/show_bug.cgi?id=199077. Changes compared to v3 (see also https://www.mail-archive.com/linux-block@vger.kernel.org/msg20073.html): - Removed the spinlock again that was introduced to protect the request state. v4 uses atomic_long_cmpxchg() instead. - Split __deadline into two variables - one for the legacy block layer and one for blk-mq. Changes compared to v2 (https://www.mail-archive.com/linux-block@vger.kernel.org/msg18338.html): - Rebased and retested on top of kernel v4.16. Changes compared to v1 (https://www.mail-archive.com/linux-block@vger.kernel.org/msg18089.html): - Removed the gstate and aborted_gstate members of struct request and used the __deadline member to encode both the generation and state information. block/blk-core.c | 6 -- block/blk-mq-debugfs.c | 1 - block/blk-mq.c | 158 ++++++++++--------------------------------------- block/blk-mq.h | 85 +++++++++++++++++--------- block/blk-timeout.c | 89 ++++++++++++++++------------ block/blk.h | 13 ++-- include/linux/blkdev.h | 29 +++------ 7 files changed, 154 insertions(+), 227 deletions(-) diff --git a/block/blk-core.c b/block/blk-core.c index de90ecab61cd..730a8e3be7ce 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -200,12 +200,6 @@ void blk_rq_init(struct request_queue *q, struct request *rq) rq->start_time = jiffies; set_start_time_ns(rq); rq->part = NULL; - seqcount_init(&rq->gstate_seq); - u64_stats_init(&rq->aborted_gstate_sync); - /* - * See comment of blk_mq_init_request - */ - WRITE_ONCE(rq->gstate, MQ_RQ_GEN_INC); } EXPORT_SYMBOL(blk_rq_init); diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c index adb8d6f00098..529383841b3b 100644 --- a/block/blk-mq-debugfs.c +++ b/block/blk-mq-debugfs.c @@ -346,7 +346,6 @@ static const char *const rqf_name[] = { RQF_NAME(STATS), RQF_NAME(SPECIAL_PAYLOAD), RQF_NAME(ZONE_WRITE_LOCKED), - RQF_NAME(MQ_TIMEOUT_EXPIRED), RQF_NAME(MQ_POLL_SLEPT), }; #undef RQF_NAME diff --git a/block/blk-mq.c b/block/blk-mq.c index bb7f59d319fa..6f20845827f4 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -481,7 +481,8 @@ void blk_mq_free_request(struct request *rq) if (blk_rq_rl(rq)) blk_put_rl(blk_rq_rl(rq)); - blk_mq_rq_update_state(rq, MQ_RQ_IDLE); + if (!blk_mq_change_rq_state(rq, blk_mq_rq_state(rq), MQ_RQ_IDLE)) + WARN_ON_ONCE(true); if (rq->tag != -1) blk_mq_put_tag(hctx, hctx->tags, ctx, rq->tag); if (sched_tag != -1) @@ -527,8 +528,7 @@ static void __blk_mq_complete_request(struct request *rq) bool shared = false; int cpu; - WARN_ON_ONCE(blk_mq_rq_state(rq) != MQ_RQ_IN_FLIGHT); - blk_mq_rq_update_state(rq, MQ_RQ_COMPLETE); + WARN_ON_ONCE(blk_mq_rq_state(rq) != MQ_RQ_COMPLETE); if (rq->internal_tag != -1) blk_mq_sched_completed_request(rq); @@ -577,36 +577,6 @@ static void hctx_lock(struct blk_mq_hw_ctx *hctx, int *srcu_idx) *srcu_idx = srcu_read_lock(hctx->srcu); } -static void blk_mq_rq_update_aborted_gstate(struct request *rq, u64 gstate) -{ - unsigned long flags; - - /* - * blk_mq_rq_aborted_gstate() is used from the completion path and - * can thus be called from irq context. u64_stats_fetch in the - * middle of update on the same CPU leads to lockup. Disable irq - * while updating. - */ - local_irq_save(flags); - u64_stats_update_begin(&rq->aborted_gstate_sync); - rq->aborted_gstate = gstate; - u64_stats_update_end(&rq->aborted_gstate_sync); - local_irq_restore(flags); -} - -static u64 blk_mq_rq_aborted_gstate(struct request *rq) -{ - unsigned int start; - u64 aborted_gstate; - - do { - start = u64_stats_fetch_begin(&rq->aborted_gstate_sync); - aborted_gstate = rq->aborted_gstate; - } while (u64_stats_fetch_retry(&rq->aborted_gstate_sync, start)); - - return aborted_gstate; -} - /** * blk_mq_complete_request - end I/O on a request * @rq: the request being processed @@ -618,27 +588,12 @@ static u64 blk_mq_rq_aborted_gstate(struct request *rq) void blk_mq_complete_request(struct request *rq) { struct request_queue *q = rq->q; - struct blk_mq_hw_ctx *hctx = blk_mq_map_queue(q, rq->mq_ctx->cpu); - int srcu_idx; if (unlikely(blk_should_fake_timeout(q))) return; - /* - * If @rq->aborted_gstate equals the current instance, timeout is - * claiming @rq and we lost. This is synchronized through - * hctx_lock(). See blk_mq_timeout_work() for details. - * - * Completion path never blocks and we can directly use RCU here - * instead of hctx_lock() which can be either RCU or SRCU. - * However, that would complicate paths which want to synchronize - * against us. Let stay in sync with the issue path so that - * hctx_lock() covers both issue and completion paths. - */ - hctx_lock(hctx, &srcu_idx); - if (blk_mq_rq_aborted_gstate(rq) != rq->gstate) + if (blk_mq_change_rq_state(rq, MQ_RQ_IN_FLIGHT, MQ_RQ_COMPLETE)) __blk_mq_complete_request(rq); - hctx_unlock(hctx, srcu_idx); } EXPORT_SYMBOL(blk_mq_complete_request); @@ -662,27 +617,7 @@ void blk_mq_start_request(struct request *rq) wbt_issue(q->rq_wb, &rq->issue_stat); } - WARN_ON_ONCE(blk_mq_rq_state(rq) != MQ_RQ_IDLE); - - /* - * Mark @rq in-flight which also advances the generation number, - * and register for timeout. Protect with a seqcount to allow the - * timeout path to read both @rq->gstate and @rq->deadline - * coherently. - * - * This is the only place where a request is marked in-flight. If - * the timeout path reads an in-flight @rq->gstate, the - * @rq->deadline it reads together under @rq->gstate_seq is - * guaranteed to be the matching one. - */ - preempt_disable(); - write_seqcount_begin(&rq->gstate_seq); - - blk_mq_rq_update_state(rq, MQ_RQ_IN_FLIGHT); - blk_add_timer(rq); - - write_seqcount_end(&rq->gstate_seq); - preempt_enable(); + blk_mq_add_timer(rq, MQ_RQ_IDLE, MQ_RQ_IN_FLIGHT); if (q->dma_drain_size && blk_rq_bytes(rq)) { /* @@ -695,22 +630,19 @@ void blk_mq_start_request(struct request *rq) } EXPORT_SYMBOL(blk_mq_start_request); -/* - * When we reach here because queue is busy, it's safe to change the state - * to IDLE without checking @rq->aborted_gstate because we should still be - * holding the RCU read lock and thus protected against timeout. - */ static void __blk_mq_requeue_request(struct request *rq) { struct request_queue *q = rq->q; + enum mq_rq_state old_state = blk_mq_rq_state(rq); blk_mq_put_driver_tag(rq); trace_block_rq_requeue(q, rq); wbt_requeue(q->rq_wb, &rq->issue_stat); - if (blk_mq_rq_state(rq) != MQ_RQ_IDLE) { - blk_mq_rq_update_state(rq, MQ_RQ_IDLE); + if (old_state != MQ_RQ_IDLE) { + if (!blk_mq_change_rq_state(rq, old_state, MQ_RQ_IDLE)) + WARN_ON_ONCE(true); if (q->dma_drain_size && blk_rq_bytes(rq)) rq->nr_phys_segments--; } @@ -819,8 +751,6 @@ static void blk_mq_rq_timed_out(struct request *req, bool reserved) const struct blk_mq_ops *ops = req->q->mq_ops; enum blk_eh_timer_return ret = BLK_EH_RESET_TIMER; - req->rq_flags |= RQF_MQ_TIMEOUT_EXPIRED; - if (ops->timeout) ret = ops->timeout(req, reserved); @@ -829,13 +759,7 @@ static void blk_mq_rq_timed_out(struct request *req, bool reserved) __blk_mq_complete_request(req); break; case BLK_EH_RESET_TIMER: - /* - * As nothing prevents from completion happening while - * ->aborted_gstate is set, this may lead to ignored - * completions and further spurious timeouts. - */ - blk_mq_rq_update_aborted_gstate(req, 0); - blk_add_timer(req); + blk_mq_add_timer(req, MQ_RQ_COMPLETE, MQ_RQ_IN_FLIGHT); break; case BLK_EH_NOT_HANDLED: break; @@ -849,48 +773,35 @@ static void blk_mq_check_expired(struct blk_mq_hw_ctx *hctx, struct request *rq, void *priv, bool reserved) { struct blk_mq_timeout_data *data = priv; - unsigned long gstate, deadline; - int start; - - might_sleep(); + unsigned long __deadline = READ_ONCE(rq->__deadline); + unsigned long deadline = __deadline & ~RQ_STATE_MASK; + enum mq_rq_state rq_state = __deadline & RQ_STATE_MASK; - if (rq->rq_flags & RQF_MQ_TIMEOUT_EXPIRED) - return; - - /* read coherent snapshots of @rq->state_gen and @rq->deadline */ - while (true) { - start = read_seqcount_begin(&rq->gstate_seq); - gstate = READ_ONCE(rq->gstate); - deadline = blk_rq_deadline(rq); - if (!read_seqcount_retry(&rq->gstate_seq, start)) - break; - cond_resched(); - } - - /* if in-flight && overdue, mark for abortion */ - if ((gstate & MQ_RQ_STATE_MASK) == MQ_RQ_IN_FLIGHT && - time_after_eq(jiffies, deadline)) { - blk_mq_rq_update_aborted_gstate(rq, gstate); + rq->aborted_gstate = __deadline ^ (1ULL << 63); + if (time_after_eq(jiffies, deadline) && rq_state == MQ_RQ_IN_FLIGHT) { + rq->aborted_gstate = __deadline; data->nr_expired++; hctx->nr_expired++; } else if (!data->next_set || time_after(data->next, deadline)) { data->next = deadline; data->next_set = 1; } + } static void blk_mq_terminate_expired(struct blk_mq_hw_ctx *hctx, struct request *rq, void *priv, bool reserved) { + unsigned long old_val = rq->aborted_gstate; + unsigned long new_val = (rq->aborted_gstate & ~RQ_STATE_MASK) | + MQ_RQ_COMPLETE; + /* - * We marked @rq->aborted_gstate and waited for RCU. If there were - * completions that we lost to, they would have finished and - * updated @rq->gstate by now; otherwise, the completion path is - * now guaranteed to see @rq->aborted_gstate and yield. If - * @rq->aborted_gstate still matches @rq->gstate, @rq is ours. + * We marked @rq->aborted_gstate and waited for ongoing .queue_rq() + * calls. If rq->__deadline has not changed that means that it is + * now safe to change the request state and to handle the timeout. */ - if (!(rq->rq_flags & RQF_MQ_TIMEOUT_EXPIRED) && - READ_ONCE(rq->gstate) == rq->aborted_gstate) + if (cmpxchg(&rq->__deadline, old_val, new_val) == old_val) blk_mq_rq_timed_out(rq, reserved); } @@ -929,10 +840,10 @@ static void blk_mq_timeout_work(struct work_struct *work) bool has_rcu = false; /* - * Wait till everyone sees ->aborted_gstate. The - * sequential waits for SRCUs aren't ideal. If this ever - * becomes a problem, we can add per-hw_ctx rcu_head and - * wait in parallel. + * For very short timeouts it can happen that + * blk_mq_check_expired() modifies the state of a request + * while .queue_rq() is still in progress. Hence wait until + * these .queue_rq() calls have finished. */ queue_for_each_hw_ctx(q, hctx, i) { if (!hctx->nr_expired) @@ -948,7 +859,7 @@ static void blk_mq_timeout_work(struct work_struct *work) if (has_rcu) synchronize_rcu(); - /* terminate the ones we won */ + /* Terminate the requests marked by blk_mq_check_expired(). */ blk_mq_queue_tag_busy_iter(q, blk_mq_terminate_expired, NULL); } @@ -2060,15 +1971,6 @@ static int blk_mq_init_request(struct blk_mq_tag_set *set, struct request *rq, return ret; } - seqcount_init(&rq->gstate_seq); - u64_stats_init(&rq->aborted_gstate_sync); - /* - * start gstate with gen 1 instead of 0, otherwise it will be equal - * to aborted_gstate, and be identified timed out by - * blk_mq_terminate_expired. - */ - WRITE_ONCE(rq->gstate, MQ_RQ_GEN_INC); - return 0; } diff --git a/block/blk-mq.h b/block/blk-mq.h index 88c558f71819..66efc8a3988b 100644 --- a/block/blk-mq.h +++ b/block/blk-mq.h @@ -27,18 +27,11 @@ struct blk_mq_ctx { struct kobject kobj; } ____cacheline_aligned_in_smp; -/* - * Bits for request->gstate. The lower two bits carry MQ_RQ_* state value - * and the upper bits the generation number. - */ +/* Lowest two bits of request->__deadline. */ enum mq_rq_state { MQ_RQ_IDLE = 0, MQ_RQ_IN_FLIGHT = 1, MQ_RQ_COMPLETE = 2, - - MQ_RQ_STATE_BITS = 2, - MQ_RQ_STATE_MASK = (1 << MQ_RQ_STATE_BITS) - 1, - MQ_RQ_GEN_INC = 1 << MQ_RQ_STATE_BITS, }; void blk_mq_freeze_queue(struct request_queue *q); @@ -100,37 +93,73 @@ extern void blk_mq_hctx_kobj_init(struct blk_mq_hw_ctx *hctx); void blk_mq_release(struct request_queue *q); +/* + * If the state of request @rq equals @old_state, update deadline and request + * state atomically to @time and @new_state. blk-mq only. cmpxchg() is only + * used if there could be a concurrent update attempt from another context. + */ +static inline bool blk_mq_rq_set_deadline(struct request *rq, + unsigned long new_time, + enum mq_rq_state old_state, + enum mq_rq_state new_state) +{ + unsigned long old_val, new_val; + + if (old_state != MQ_RQ_IN_FLIGHT) { + old_val = READ_ONCE(rq->__deadline); + if ((old_val & RQ_STATE_MASK) != old_state) + return false; + new_val = (new_time & ~RQ_STATE_MASK) | + (new_state & RQ_STATE_MASK); + WRITE_ONCE(rq->__deadline, new_val); + return true; + } + + do { + old_val = READ_ONCE(rq->__deadline); + if ((old_val & RQ_STATE_MASK) != old_state) + return false; + new_val = (new_time & ~RQ_STATE_MASK) | + (new_state & RQ_STATE_MASK); + } while (cmpxchg(&rq->__deadline, old_val, new_val) != old_val); + + return true; +} + /** * blk_mq_rq_state() - read the current MQ_RQ_* state of a request * @rq: target request. */ -static inline int blk_mq_rq_state(struct request *rq) +static inline enum mq_rq_state blk_mq_rq_state(struct request *rq) { - return READ_ONCE(rq->gstate) & MQ_RQ_STATE_MASK; + return READ_ONCE(rq->__deadline) & RQ_STATE_MASK; } /** - * blk_mq_rq_update_state() - set the current MQ_RQ_* state of a request - * @rq: target request. - * @state: new state to set. + * blk_mq_change_rq_state - atomically test and set request state + * @rq: Request pointer. + * @old_state: Old request state. + * @new_state: New request state. * - * Set @rq's state to @state. The caller is responsible for ensuring that - * there are no other updaters. A request can transition into IN_FLIGHT - * only from IDLE and doing so increments the generation number. + * Returns %true if and only if the old state was @old and if the state has + * been changed into @new. */ -static inline void blk_mq_rq_update_state(struct request *rq, - enum mq_rq_state state) +static inline bool blk_mq_change_rq_state(struct request *rq, + enum mq_rq_state old_state, + enum mq_rq_state new_state) { - u64 old_val = READ_ONCE(rq->gstate); - u64 new_val = (old_val & ~MQ_RQ_STATE_MASK) | state; - - if (state == MQ_RQ_IN_FLIGHT) { - WARN_ON_ONCE((old_val & MQ_RQ_STATE_MASK) != MQ_RQ_IDLE); - new_val += MQ_RQ_GEN_INC; - } - - /* avoid exposing interim values */ - WRITE_ONCE(rq->gstate, new_val); + unsigned long old_val = (READ_ONCE(rq->__deadline) & ~RQ_STATE_MASK) | + old_state; + unsigned long new_val = (old_val & ~RQ_STATE_MASK) | new_state; + + /* + * For transitions from state in-flight to another state cmpxchg() must + * be used. For other state transitions it is safe to use WRITE_ONCE(). + */ + if (old_state == MQ_RQ_IN_FLIGHT) + return cmpxchg(&rq->__deadline, old_val, new_val) == old_val; + WRITE_ONCE(rq->__deadline, new_val); + return true; } static inline struct blk_mq_ctx *__blk_mq_get_ctx(struct request_queue *q, diff --git a/block/blk-timeout.c b/block/blk-timeout.c index 50a191720055..e98da6db7d4b 100644 --- a/block/blk-timeout.c +++ b/block/blk-timeout.c @@ -165,8 +165,9 @@ void blk_abort_request(struct request *req) * immediately and that scan sees the new timeout value. * No need for fancy synchronizations. */ - blk_rq_set_deadline(req, jiffies); - kblockd_schedule_work(&req->q->timeout_work); + if (blk_mq_rq_set_deadline(req, jiffies, MQ_RQ_IN_FLIGHT, + MQ_RQ_IN_FLIGHT)) + kblockd_schedule_work(&req->q->timeout_work); } else { if (blk_mark_rq_complete(req)) return; @@ -187,52 +188,17 @@ unsigned long blk_rq_timeout(unsigned long timeout) return timeout; } -/** - * blk_add_timer - Start timeout timer for a single request - * @req: request that is about to start running. - * - * Notes: - * Each request has its own timer, and as it is added to the queue, we - * set up the timer. When the request completes, we cancel the timer. - */ -void blk_add_timer(struct request *req) +static void __blk_add_timer(struct request *req) { struct request_queue *q = req->q; unsigned long expiry; - if (!q->mq_ops) - lockdep_assert_held(q->queue_lock); - - /* blk-mq has its own handler, so we don't need ->rq_timed_out_fn */ - if (!q->mq_ops && !q->rq_timed_out_fn) - return; - - BUG_ON(!list_empty(&req->timeout_list)); - - /* - * Some LLDs, like scsi, peek at the timeout to prevent a - * command from being retried forever. - */ - if (!req->timeout) - req->timeout = q->rq_timeout; - - blk_rq_set_deadline(req, jiffies + req->timeout); - req->rq_flags &= ~RQF_MQ_TIMEOUT_EXPIRED; - - /* - * Only the non-mq case needs to add the request to a protected list. - * For the mq case we simply scan the tag map. - */ - if (!q->mq_ops) - list_add_tail(&req->timeout_list, &req->q->timeout_list); - /* * If the timer isn't already pending or this timeout is earlier * than an existing one, modify the timer. Round up to next nearest * second. */ expiry = blk_rq_timeout(round_jiffies_up(blk_rq_deadline(req))); - if (!timer_pending(&q->timeout) || time_before(expiry, q->timeout.expires)) { unsigned long diff = q->timeout.expires - expiry; @@ -247,5 +213,52 @@ void blk_add_timer(struct request *req) if (!timer_pending(&q->timeout) || (diff >= HZ / 2)) mod_timer(&q->timeout, expiry); } +} + +/** + * blk_add_timer - Start timeout timer for a single request + * @req: request that is about to start running. + * + * Notes: + * Each request has its own timer, and as it is added to the queue, we + * set up the timer. When the request completes, we cancel the timer. + */ +void blk_add_timer(struct request *req) +{ + struct request_queue *q = req->q; + + lockdep_assert_held(q->queue_lock); + if (!q->rq_timed_out_fn) + return; + if (!req->timeout) + req->timeout = q->rq_timeout; + + blk_rq_set_deadline(req, jiffies + req->timeout); + list_add_tail(&req->timeout_list, &req->q->timeout_list); + + return __blk_add_timer(req); +} + +/** + * blk_mq_add_timer - set the deadline for a single request + * @req: request for which to set the deadline. + * @old: current request state. + * @new: new request state. + * + * Sets the deadline of a request if and only if it has state @old and + * at the same time changes the request state from @old into @new. The caller + * must guarantee that the request state won't be modified while this function + * is in progress. + */ +void blk_mq_add_timer(struct request *req, enum mq_rq_state old, + enum mq_rq_state new) +{ + struct request_queue *q = req->q; + + if (!req->timeout) + req->timeout = q->rq_timeout; + if (!blk_mq_rq_set_deadline(req, jiffies + req->timeout, old, new)) + WARN_ON_ONCE(true); + return __blk_add_timer(req); } diff --git a/block/blk.h b/block/blk.h index b034fd2460c4..7cd64f533a46 100644 --- a/block/blk.h +++ b/block/blk.h @@ -170,6 +170,8 @@ static inline bool bio_integrity_endio(struct bio *bio) void blk_timeout_work(struct work_struct *work); unsigned long blk_rq_timeout(unsigned long timeout); void blk_add_timer(struct request *req); +void blk_mq_add_timer(struct request *req, enum mq_rq_state old, + enum mq_rq_state new); void blk_delete_timer(struct request *); @@ -308,18 +310,19 @@ static inline void req_set_nomerge(struct request_queue *q, struct request *req) } /* - * Steal a bit from this field for legacy IO path atomic IO marking. Note that - * setting the deadline clears the bottom bit, potentially clearing the - * completed bit. The user has to be OK with this (current ones are fine). + * Steal two bits from this field. The legacy IO path uses the lowest bit for + * atomic IO marking. Note that setting the deadline clears the bottom bit, + * potentially clearing the completed bit. The current legacy block layer is + * fine with that. Must be called with the request queue lock held. */ static inline void blk_rq_set_deadline(struct request *rq, unsigned long time) { - rq->__deadline = time & ~0x1UL; + rq->__deadline = time & RQ_STATE_MASK; } static inline unsigned long blk_rq_deadline(struct request *rq) { - return rq->__deadline & ~0x1UL; + return rq->__deadline & ~RQ_STATE_MASK; } /* diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index b7681f3ee793..51cd69f14537 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -27,8 +27,6 @@ #include <linux/percpu-refcount.h> #include <linux/scatterlist.h> #include <linux/blkzoned.h> -#include <linux/seqlock.h> -#include <linux/u64_stats_sync.h> struct module; struct scsi_ioctl_command; @@ -125,10 +123,8 @@ typedef __u32 __bitwise req_flags_t; #define RQF_SPECIAL_PAYLOAD ((__force req_flags_t)(1 << 18)) /* The per-zone write lock is held for this request */ #define RQF_ZONE_WRITE_LOCKED ((__force req_flags_t)(1 << 19)) -/* timeout is expired */ -#define RQF_MQ_TIMEOUT_EXPIRED ((__force req_flags_t)(1 << 20)) /* already slept for hybrid poll */ -#define RQF_MQ_POLL_SLEPT ((__force req_flags_t)(1 << 21)) +#define RQF_MQ_POLL_SLEPT ((__force req_flags_t)(1 << 20)) /* flags that prevent us from merging requests: */ #define RQF_NOMERGE_FLAGS \ @@ -225,28 +221,19 @@ struct request { unsigned int extra_len; /* length of alignment and padding */ - /* - * On blk-mq, the lower bits of ->gstate (generation number and - * state) carry the MQ_RQ_* state value and the upper bits the - * generation number which is monotonically incremented and used to - * distinguish the reuse instances. - * - * ->gstate_seq allows updates to ->gstate and other fields - * (currently ->deadline) during request start to be read - * atomically from the timeout path, so that it can operate on a - * coherent set of information. - */ - seqcount_t gstate_seq; - u64 gstate; - /* * ->aborted_gstate is used by the timeout to claim a specific * recycle instance of this request. See blk_mq_timeout_work(). */ - struct u64_stats_sync aborted_gstate_sync; u64 aborted_gstate; - /* access through blk_rq_set_deadline, blk_rq_deadline */ + /* + * Access through blk_rq_deadline() and blk_rq_set_deadline(), + * blk_mark_rq_complete(), blk_clear_rq_complete() and + * blk_rq_is_complete() for legacy queues or blk_mq_rq_state() for + * blk-mq queues. + */ +#define RQ_STATE_MASK 0x3UL unsigned long __deadline; struct list_head timeout_list; -- 2.16.3

6 years, 9 months

4
8
0 0

GREETING FROM MISS QADESA,,F

by Miss Qadesa AbdulAziz

Friend, My name is Miss Qadesa AbdulAziz and I am 17 years old girl from Syria. There is serious war crisis here in Syria, and I have lost my parents and my two brothers in this war. I want you to help me and receive ($7.md) which my late father deposited with my name in a bank in London. I want to come to your country and start a new life and invest with you, because am the only survival in my family. I wait to hear from you, Please do not let me die here, I begging you the name of Almighty. please respond here my pirate email qadesa(a)protonmail.com Regards Miss Qadesa AbdulAziz

6 years, 9 months

1
0
0 0

[PATCH v2] target: Fix Fortify_panic kernel exception

by Bryant G. Ly

The bug exists in the memcmp in which the length passed in must be guaranteed to be 1. This bug currently exists because the second pointer passed in, can be smaller than the cmd->data_length, which causes a fortify_panic. The fix is to use memchr_inv instead to find whether or not a 0 exists instead of using memcmp. This way you dont have to worry about buffer overflow which is the reason for the fortify_panic. The bug was found by running a block backstore via LIO. [ 496.212958] Call Trace: [ 496.212960] [c0000007e58e3800] [c000000000cbbefc] fortify_panic+0x24/0x38 (unreliable) [ 496.212965] [c0000007e58e3860] [d00000000f150c28] iblock_execute_write_same+0x3b8/0x3c0 [target_core_iblock] [ 496.212976] [c0000007e58e3910] [d000000006c737d4] __target_execute_cmd+0x54/0x150 [target_core_mod] [ 496.212982] [c0000007e58e3940] [d000000006d32ce4] ibmvscsis_write_pending+0x74/0xe0 [ibmvscsis] [ 496.212991] [c0000007e58e39b0] [d000000006c74fc8] transport_generic_new_cmd+0x318/0x370 [target_core_mod] [ 496.213001] [c0000007e58e3a30] [d000000006c75084] transport_handle_cdb_direct+0x64/0xd0 [target_core_mod] [ 496.213011] [c0000007e58e3aa0] [d000000006c75298] target_submit_cmd_map_sgls+0x1a8/0x320 [target_core_mod] [ 496.213021] [c0000007e58e3b30] [d000000006c75458] target_submit_cmd+0x48/0x60 [target_core_mod] [ 496.213026] [c0000007e58e3bd0] [d000000006d34c20] ibmvscsis_scheduler+0x370/0x600 [ibmvscsis] [ 496.213031] [c0000007e58e3c90] [c00000000013135c] process_one_work+0x1ec/0x580 [ 496.213035] [c0000007e58e3d20] [c000000000131798] worker_thread+0xa8/0x600 [ 496.213039] [c0000007e58e3dc0] [c00000000013a468] kthread+0x168/0x1b0 [ 496.213044] [c0000007e58e3e30] [c00000000000b528] ret_from_kernel_thread+0x5c/0xb4 Fixes: 2237498f0b5c ("target/iblock: Convert WRITE_SAME to blkdev_issue_zeroout") Signed-off-by: Bryant G. Ly <bryantly(a)linux.vnet.ibm.com> Reviewed-by: Steven Royer <seroyer(a)linux.vnet.ibm.com> Tested-by: Taylor Jakobson <tjakobs(a)us.ibm.com> Cc: Christoph Hellwig <hch(a)lst.de> Cc: Nicholas Bellinger <nab(a)linux-iscsi.org> Cc: <stable(a)vger.kernel.org> --- drivers/target/target_core_iblock.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/target/target_core_iblock.c b/drivers/target/target_core_iblock.c index 07c814c..6042901 100644 --- a/drivers/target/target_core_iblock.c +++ b/drivers/target/target_core_iblock.c @@ -427,8 +427,8 @@ iblock_execute_zero_out(struct block_device *bdev, struct se_cmd *cmd) { struct se_device *dev = cmd->se_dev; struct scatterlist *sg = &cmd->t_data_sg[0]; - unsigned char *buf, zero = 0x00, *p = &zero; - int rc, ret; + unsigned char *buf, *not_zero; + int ret; buf = kmap(sg_page(sg)) + sg->offset; if (!buf) @@ -437,10 +437,10 @@ iblock_execute_zero_out(struct block_device *bdev, struct se_cmd *cmd) * Fall back to block_execute_write_same() slow-path if * incoming WRITE_SAME payload does not contain zeros. */ - rc = memcmp(buf, p, cmd->data_length); + not_zero = memchr_inv(buf, 0x00, cmd->data_length); kunmap(sg_page(sg)); - if (rc) + if (not_zero) return TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE; ret = blkdev_issue_zeroout(bdev, -- 2.7.2

6 years, 9 months

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror April 2018