On 3/19/22 03:06, Christian Lamparter wrote: > On 18/03/2022 15:16, Andy Shevchenko wrote: >> On Fri, Mar 18, 2022 at 10:03:11AM +0100, Christian Lamparter wrote: >>> the driver uses libata's "tag" values from in various arrays. >>> Since the mentioned patch bumped the ATA_TAG_INTERNAL to 32, >>> the value of the SATA_DWC_QCMD_MAX needs to be bumped to 33. >>> >>> Otherwise ATA_TAG_INTERNAL cause a crash like this: >>> >>> | BUG: Kernel NULL pointer dereference at 0x00000000 >>> | Faulting instruction address: 0xc03ed4b8 >>> | Oops: Kernel access of bad area, sig: 11 [#1] >>> | BE PAGE_SIZE=4K PowerPC 44x Platform >>> | CPU: 0 PID: 362 Comm: scsi_eh_1 Not tainted 5.4.163 #0 >>> | NIP: c03ed4b8 LR: c03d27e8 CTR: c03ed36c >>> | REGS: cfa59950 TRAP: 0300 Not tainted (5.4.163) >>> | MSR: 00021000 <CE,ME> CR: 42000222 XER: 00000000 >>> | DEAR: 00000000 ESR: 00000000 >>> | GPR00: c03d27e8 cfa59a08 cfa55fe0 00000000 0fa46bc0 [...] >>> | [..] >>> | NIP [c03ed4b8] sata_dwc_qc_issue+0x14c/0x254 >>> | LR [c03d27e8] ata_qc_issue+0x1c8/0x2dc >>> | Call Trace: >>> | [cfa59a08] [c003f4e0] __cancel_work_timer+0x124/0x194 (unreliable) >>> | [cfa59a78] [c03d27e8] ata_qc_issue+0x1c8/0x2dc >>> | [cfa59a98] [c03d2b3c] ata_exec_internal_sg+0x240/0x524 >>> | [cfa59b08] [c03d2e98] ata_exec_internal+0x78/0xe0 >>> | [cfa59b58] [c03d30fc] ata_read_log_page.part.38+0x1dc/0x204 >>> | [cfa59bc8] [c03d324c] ata_identify_page_supported+0x68/0x130 >>> | [...] >>> >>> this is because sata_dwc_dma_xfer_complete() NULLs the >>> dma_pending's next neighbour "chan" (a *dma_chan struct) in >>> this '32' case right here (line ~735): >>>> hsdevp->dma_pending[tag] = SATA_DWC_DMA_PENDING_NONE; >>> >>> Then the next time, a dma gets issued; dma_dwc_xfer_setup() passes >>> the NULL'd hsdevp->chan to the dmaengine_slave_config() which then >>> causes the crash. >> >> ... >> >>> ticerex said when I've asked him about his real name+email for the patches: >>> "Please use my forum nick." >>> <https://forum.openwrt.org/t/my-book-live-duo-reboot-loop/122464/14> >>> (I know checkpatch.pl complains about that. But what can you do?) >> >> I think Reported-by: is fine to have any kind of reference to the reporter. >> I can consider it false positive. >> >> ... > > K, I've reported this back to the reporter ;). > > Documentation/process/maintainer-tip.rst and process/5.Posting.rst > provided some hints: > > "Please note that if the bug was reported in private, then ask for > permission first before using the Reported-by tag." > > and maintainer-tip.rst, the format should be: > ``Reported-by: ``Reporter <reporter@mail>`` > > (My goal here is to get "a fix" merged, so conformance is key. ;-)) No worries, a fix will go in :) Since the person did not want to give his/her/they real name & email, we could consider this as that person not accepting to be mentioned in a reported-by tag. So dropping the tag is I think preferred. Keeping it as you had it is fine too. No big deal. > >>> -#define SATA_DWC_QCMD_MAX 32 >>> +#define SATA_DWC_QCMD_MAX 33 >> >> Can't we use >> >> #define SATA_DWC_QCMD_MAX (ATA_TAG_INTERNAL + 1) >> > > I've looked around a bit. > > include/linux/libata.h itself has the following related definitions: > > | enum { > | ATA_MAX_QUEUE = 32, > | ATA_TAG_INTERNAL = ATA_MAX_QUEUE, > | [..] > > | struct ata_port { > | [...] > | struct ata_queued_cmd qcmd[ATA_MAX_QUEUE + 1]; > | > | } > > ATA_MAX_QUEUE seems to be the "size" value. Whereas ATA_TAG_INTERNAL > is used as (tag == ATA_TAG_INTERNAL) more often. > > > I came up with a viable compromise: > > Would it be "OK" to define a "new" ATA_MAX_TAG? > > This could be either set ATA_MAX_TAG = ATA_MAX_QUEUE + 1 > > or let it be assigned automatically, if it's placed after ATA_TAG_INTERNAL > in the libata.h's enum like this: (I prefer this, but it being "33" is not > obvious if you don't dabble in C all the time) > > | enum { > | [...] > | ATA_MAX_QUEUE = 32, > | ATA_TAG_INTERNAL = ATA_MAX_QUEUE, > | /* > | * ATA_MAX_TAG accounts for ATA_MAX_QUEUE TAGs + 1 special TAG/slot > | * for ATA_TAG_INTERNAL (libata's internal commands/DMA management). > | * This needs to be it in this location. > | */ > | ATA_MAX_TAG, > | [...] > > This ATA_MAX_TAG could then be used for ata_port's qcmd (from above) and > sata_dwc_460ex's SATA_DWC_QCMD_MAX. > > For good measure: > > BUILD_BUG_ON(ATA_TAG_INTERNAL >= ATA_MAX_TAG); > > could be added into libata.h's __ata_qc_from_tag(). > It access the qcmd array, so anyone using libata's accessors will catch > future updates. > > Is this fine or getting to close to being overbuild? > > Thank you both for your insights, > Christian -- Damien Le Moal Western Digital Research