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 Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by smtp.lore.kernel.org (Postfix) with ESMTP id 7878BCD98F6 for ; Thu, 18 Jun 2026 15:56:13 +0000 (UTC) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id C341C4027F; Thu, 18 Jun 2026 17:56:12 +0200 (CEST) Received: from mail-dl1-f54.google.com (mail-dl1-f54.google.com [74.125.82.54]) by mails.dpdk.org (Postfix) with ESMTP id EF3B240268 for ; Thu, 18 Jun 2026 17:56:11 +0200 (CEST) Received: by mail-dl1-f54.google.com with SMTP id a92af1059eb24-13721dfd471so1753495c88.1 for ; Thu, 18 Jun 2026 08:56:11 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=networkplumber-org.20251104.gappssmtp.com; s=20251104; t=1781798171; x=1782402971; darn=dpdk.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:subject:cc:to:from:date:from:to:cc:subject:date :message-id:reply-to; bh=P6Cvf9dYg3jgZ6bMSQ8ZzwbCNqjTjVCzUiMksQZQt3Y=; b=E2n0DzJdSuxurGPCMPoFQ+FiEgW8oG4JGXBeKxUDd1Apw0dFH/JAI6B4oNdhoVeWUZ XaP7V3PgdA5vkOo8p58KLR0zuyGaENuaTFPgYRcx3v+3a/DjobX6jfiaGR4ylkaxqbat oRItVQu2CQbzSejRov5db3enAnhw30GE/YvrMletI1QyfmBGtAr2SPpHtGTEigI6FmE3 mEv673oHO76gA2jvDrhv0oVV4hGmtEKylhqhdG90nybPzT+r+PiEUpiGomPhJKcnOKZP XTx1wkS9PxcXkwg8BcEPLzOtEY72UvhID8FLHxy6V94Fih2uGvFlGn8XqtKXlMfqTKUQ 2scg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1781798171; x=1782402971; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:subject:cc:to:from:date:x-gm-gg:x-gm-message-state:from :to:cc:subject:date:message-id:reply-to; bh=P6Cvf9dYg3jgZ6bMSQ8ZzwbCNqjTjVCzUiMksQZQt3Y=; b=OeTMzrp6kH7lvc1dS6HFkdFzlpu3qnbgXBxaqEYtWNr4FJ0w164gbYO2SlK/q/kRgN h0NF3M3fzwUdjGECTS4X4X3WlK8pVt0T3ISRkFccQgWOpPGx/lVfaXwrDna9q1NKTCf5 ej23Yr7nZJp8GhAUfbrUha3W3IITPdW1Lll8af0qWPtLdr1lOPX9mU9mXuxOVzkkDefQ BEoiG/wl6ZmAV2WfpBYiaDj9y0U6spmBGYhlvJmCaPldtWGb+9MHdCAkNA7okJ9AOLBw ghw061+TM9LJ6pEYxE6Fv08G3oscqIeN6g2vLY2B8HMNnjBtkrdQsBlBnHNk5afwmfzk iLNQ== X-Forwarded-Encrypted: i=1; AFNElJ/eG+2I07fLyfIyAklVvJwZ1mgRCHN1IXzLbNmRP9ExkX+9NKyAG6mt/WQ1kdZj+qWnL2U=@dpdk.org X-Gm-Message-State: AOJu0YzO9U0eWLJCR7GhV+iqXUSx902yFLrHH68+r85F7jNL6RvUWEh1 PXyVLgTwtfdOMnoNmwaMEFRjYVF7bu88cFgxLoB/J/fqcamJmrRY7BB2SZ53FWac5O4= X-Gm-Gg: AfdE7cnSTnN25kzh+feVBI8RJTWJnFj3Rt9cBYLZzC02FqXAlycG7o5uvUgEL+/qRZ3 PK8jb9abQ+YunsSDMr6DHpBLI3AlqdCgtxQ3RoUqmAwPCTAOiEMKiwwvnln5B2VMTZO9R7Bzels /L/JEOen215Fz1H9ooAO4DQiu1JPQC0BLrgKzCEDmp/b/kAyKAOJj5GVFQfV/b81UNBoXNAsWiH cxmPsC2OqknOiSnxyw2jvYrw6vbUhYsCjvkZVXOgksJadopPUNiKh24ORPYdA2nHd2V/FwVsFm3 3uS4Vlz0xwqQgBlAWOQw6ATB9CSNo+wNZ3ExNHAoKekAG0kzRh1C/nKdT+ri3+yo+TFE+/aPlDx BnF3WYK+aiWYNjulWYGdYQxN9pSxPdd2DV4ONFnLzNj1yEupwSlhqxlw/9951WiKCSgnWOH5YwG w+fMC5hH49J7JgKJVqmn2JAjaj48gg0NMyh+iH0HF/UOxKce6aCy2G9w== X-Received: by 2002:a05:7022:e23:b0:138:12fa:377e with SMTP id a92af1059eb24-139a210e309mr138920c88.20.1781798170760; Thu, 18 Jun 2026 08:56:10 -0700 (PDT) Received: from phoenix.local (204-195-96-226.wavecable.com. [204.195.96.226]) by smtp.gmail.com with ESMTPSA id 5a478bee46e88-30c02b17238sm355529eec.9.2026.06.18.08.56.10 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 18 Jun 2026 08:56:10 -0700 (PDT) Date: Thu, 18 Jun 2026 08:56:06 -0700 From: Stephen Hemminger To: Hemant Agrawal Cc: david.marchand@redhat.com, dev@dpdk.org Subject: Re: [PATCH v1 00/17] net/dpaa: bug fixes for bus, net and fmlib drivers Message-ID: <20260618085606.46b295ba@phoenix.local> In-Reply-To: <20260618141151.3990283-1-hemant.agrawal@nxp.com> References: <20260618141151.3990283-1-hemant.agrawal@nxp.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org On Thu, 18 Jun 2026 19:41:34 +0530 Hemant Agrawal wrote: > This series contains bug fixes for the DPAA PMD (bus/dpaa, net/dpaa, > net/dpaa/fmlib and dma/dpaa): > > - Fix error handling in qman_create_fq and qman_query > - Fix fqid endianness in qman_fq_flow_control > - Fix CGR index usage in dpaa_modify_cgr > - Add null check in fmlib scheme delete > - Fix BMI RX stats register offset > - Fix file descriptor leak after CCSR mmap > - Fix device probe regression on LS1043A > - Fix double-close in device remove path > - Fix incorrect condition in interrupt unregister > - Fix Coverity-reported issues in dpaa_flow and dpaa_qdma > - Fix xstat name for tx undersized counter > - Fix xstat string typos in BMI stats table > - Remove duplicate ptype entries > - Fix wrong buffer in xstats get by id > - Fix null l3_len check in checksum offload > - Fix mbuf leak in SG fd creation > > All patches are bug fixes tagged with Fixes: and Cc: stable@dpdk.org. > > Gagandeep Singh (3): > bus/dpaa: fix device probe issue > net/dpaa: fix device remove > net/dpaa: fix invalid check on interrupt unregister > > Hemant Agrawal (11): > bus/dpaa: fix error handling of qman_create_fq > bus/dpaa: fix fqid endianness > bus/dpaa: fix error handling in qman_query > net/dpaa: fix modify cgr to use index > bus/dpaa: fix fd leak for ccsr mmap > net/dpaa: fix xstat name for tx undersized counter > net/dpaa: fix xstat string typos in BMI stats table > net/dpaa: remove duplicate ptype entries > net/dpaa: fix wrong buffer in xstats get by id > net/dpaa: fix null l3_len check in checksum offload > net/dpaa: fix mbuf leak in SG fd creation > > Jun Yang (1): > bus/dpaa: fix BMI RX stats register offset > > Prashant Gupta (1): > net/dpaa/fmlib: add null check in scheme delete > > Vanshika Shukla (1): > net/dpaa: fix coverity reported issues > > drivers/bus/dpaa/base/qbman/bman_driver.c | 3 ++- > drivers/bus/dpaa/base/qbman/qman.c | 11 +++++---- > drivers/bus/dpaa/base/qbman/qman_driver.c | 6 ++--- > drivers/bus/dpaa/dpaa_bus.c | 6 ++--- > drivers/bus/dpaa/include/fman.h | 6 ++--- > drivers/dma/dpaa/dpaa_qdma.c | 7 +++++- > drivers/net/dpaa/dpaa_ethdev.c | 27 +++++++++++------------ > drivers/net/dpaa/dpaa_flow.c | 4 ++++ > drivers/net/dpaa/dpaa_rxtx.c | 3 +++ > drivers/net/dpaa/fmlib/fm_lib.c | 3 +++ > 10 files changed, 46 insertions(+), 30 deletions(-) > Looks good but there are some warnings from more detailed AI review that need addressing. Review of [PATCH v1 00/17] bus/dpaa, net/dpaa fixes Reviewed against current main. No errors found; the series is a solid set of bug fixes. A few warnings and notes below, mostly about undocumented side effects and Fixes: accuracy. Patches not listed (01, 02, 03, 04, 06, 07, 12, 14, 15, 16, 17) look correct. [PATCH v1 05/17] net/dpaa/fmlib: add null check in scheme delete Info: the commit body refers to FM_PCD_MatchTableSchemeDelete(), but the function actually changed is fm_pcd_kg_scheme_delete(). The NULL guard and E_NO_DEVICE return are correct and match the sibling functions in fm_lib.c; only the message names the wrong API. [PATCH v1 08/17] bus/dpaa: fix device probe issue The early "return 0" was indeed skipping device-list creation, so removing it fixes probe on LS1043A. Good. Warning: that early return did two things -- it forced max_push_rxq_num = 0 AND skipped the DPAA_PUSH_QUEUES_NUMBER override that follows. With the return gone, execution now falls through to: penv = getenv("DPAA_PUSH_QUEUES_NUMBER"); if (penv) dpaa_bus.max_push_rxq_num = atoi(penv); so on LS1043A the env var can now raise the push-queue count back above zero, which the original code deliberately prevented (the comment is "Disabling the default push mode for LS1043A", and the SoC has the FMAN push-mode errata handled in dpaa_rxtx.c). If LS1043A must keep push mode disabled regardless of the env var, guard the override, e.g.: if (dpaa_bus.svr_ver == SVR_LS1043A_FAMILY) { dpaa_bus.max_push_rxq_num = 0; } else { penv = getenv("DPAA_PUSH_QUEUES_NUMBER"); ... } If the override is intended to apply to LS1043A, please say so in the commit message. [PATCH v1 09/17] net/dpaa: fix device remove The RTE_ETH_DEV_UNUSED guard against double close/release is correct, and "int ret = 0" is now required because the assignment is conditional. Warning: the patch also drops the dpaa_finish() call, which the commit message does not mention. dpaa_finish() is registered as a destructor (RTE_FINI_PRIO(dpaa_finish, 103)), so it still runs at process exit, but the previous explicit call ran at last-device remove (!dpaa_valid_dev). Removing it moves the global teardown (dpaa_fm_term, per-queue portal close, is_global_init = 0) from last-remove time to exit time. For run-then-exit this is equivalent, but for remove-all-then-continue (e.g. re-probe in a running process) is_global_init now stays 1 and portals stay open until exit. Please call this change out in the commit message and confirm re-probe still behaves. [PATCH v1 10/17] net/dpaa: fix invalid check on interrupt unregister The fix is correct: rte_intr_callback_unregister() returns the number of callbacks removed (>=1) on success and a negative value on failure, so "if (ret)" logged a spurious warning on every successful unregister; "if (ret < 0)" is right. Warning: the Fixes: tag points at 9c99878aa1 ("log: introduce logtype register macro") which is unrelated to interrupt unregistration and looks like a copy/paste error. An incorrect Fixes: will misdirect the stable backport -- please point it at the commit that introduced the "if (ret)" check. [PATCH v1 11/17] net/dpaa: fix coverity reported issues Both fixes are correct. The dpaa_qdma.c bound (num == 0 || num > FSL_QDMA_SG_MAX_ENTRY) prevents the desc_ssge[num - 1] underflow when pending_num is 0. The dpaa_flow.c port_handle close fixes the continue-path leak without introducing a double-close: the success path nulls port_handle inside dpaa_fm_deconfig(), and the error/continue paths legitimately leave it set for the new guard to close. Info: this bundles two unrelated Coverity fixes across drivers/dma/dpaa and drivers/net/dpaa under a single net/dpaa Fixes: tag (e7665de896). The qdma OOB almost certainly has a different origin commit. Splitting into two patches (or at least carrying the correct Fixes: for the qdma change) would make the stable backport cleaner. [PATCH v1 13/17] net/dpaa: fix xstat string typos in BMI stats table Both typo fixes are correct. Info: the very next entry on the same table is also misspelled -- "rx_buf_diallocate" (fmbm_rbdc, "Rx Buffers Deallocate Counter") should be "rx_buf_deallocate". Worth fixing in the same pass since this patch is specifically cleaning up these names.