qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Harsh Prateek Bora <harshpb@linux.ibm.com>
To: Aditya Gupta <adityag@linux.ibm.com>, qemu-devel@nongnu.org
Cc: qemu-ppc@nongnu.org, Nicholas Piggin <npiggin@gmail.com>,
	Daniel Henrique Barboza <danielhb413@gmail.com>,
	Sourabh Jain <sourabhjain@linux.ibm.com>,
	Mahesh J Salgaonkar <mahesh@linux.ibm.com>,
	Hari Bathini <hbathini@linux.ibm.com>
Subject: Re: [PATCH 3/6] hw/ppc: Preserve memory regions registered for fadump
Date: Wed, 5 Mar 2025 12:10:35 +0530	[thread overview]
Message-ID: <f5494f6f-599f-427b-8c37-42cc5396c35d@linux.ibm.com> (raw)
In-Reply-To: <20250217071711.83735-4-adityag@linux.ibm.com>



On 2/17/25 12:47, Aditya Gupta wrote:
> While the first kernel boots, it registers memory regions for fadump
> such as:
>      * CPU state data  (has to be populated by the platform)
>      * HPTE state data (has to be populated by the platform)
>      * Real Mode Regions (platform should copy it to requested
>        destination addresses)
>      * OS defined regions (such as parameter save area)
> 
> Platform is also expected to modify the 'bytes_dumped' to the length of
> data preserved/copied by platform (ideally same as the source length
> passed by kernel).
> 
> The kernel passes source address and length for the memory regions, and
> a destination address to where the memory is to be copied.
> 
> Implement the preserving/copying of the Real Mode Regions and the
> Parameter Save Area in QEMU Pseries
> 
> Signed-off-by: Aditya Gupta <adityag@linux.ibm.com>
> ---
>   hw/ppc/spapr_rtas.c    | 117 ++++++++++++++++++++++++++++++++++++++++-
>   include/hw/ppc/spapr.h |  27 +++++++++-
>   2 files changed, 142 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/ppc/spapr_rtas.c b/hw/ppc/spapr_rtas.c
> index 01c82375f03d..9b29cadab2c9 100644
> --- a/hw/ppc/spapr_rtas.c
> +++ b/hw/ppc/spapr_rtas.c
> @@ -344,6 +344,120 @@ static void rtas_ibm_set_system_parameter(PowerPCCPU *cpu,
>   struct fadump_metadata fadump_metadata;
>   bool is_next_boot_fadump;
>   
> +/* Preserve the memory locations registered for fadump */
> +static bool fadump_preserve_mem(void)
> +{
> +    struct rtas_fadump_mem_struct *fdm = &fadump_metadata.registered_fdm;
> +    uint64_t next_section_addr;
> +    int dump_num_sections, data_type;
> +    uint64_t src_addr, src_len, dest_addr;
> +    void *copy_buffer;
> +
> +    assert(fadump_metadata.fadump_registered);
> +    assert(fadump_metadata.fdm_addr != -1);
> +
> +    /* Read the fadump header passed during fadump registration */
> +    cpu_physical_memory_read(fadump_metadata.fdm_addr,
> +            &fdm->header, sizeof(fdm->header));
> +
> +    /* Verify that we understand the fadump header version */
> +    if (fdm->header.dump_format_version != cpu_to_be32(FADUMP_VERSION)) {
> +        /*
> +         * Dump format version is unknown and likely changed from the time
> +         * of fadump registration. Back out now.
> +         */
> +        return false;
> +    }
> +
> +    dump_num_sections = be16_to_cpu(fdm->header.dump_num_sections);
> +
> +    if (dump_num_sections > FADUMP_MAX_SECTIONS) {
> +        qemu_log_mask(LOG_GUEST_ERROR,
> +            "FADUMP: Too many sections: %d\n", fdm->header.dump_num_sections);
> +        return false;
> +    }
> +
> +    next_section_addr =
> +        fadump_metadata.fdm_addr +
> +        be32_to_cpu(fdm->header.offset_first_dump_section);
> +
> +    /*
> +     * Handle all sections
> +     *
> +     * CPU State Data and HPTE regions are handled in their own cases
> +     *
> +     * RMR regions and any custom OS reserved regions such as parameter
> +     * save area, are handled by simply copying the source region to
> +     * destination address
> +     */
> +    for (int i = 0; i < dump_num_sections; ++i) {
> +        /* Read the fadump section from memory */
> +        cpu_physical_memory_read(next_section_addr,
> +                &fdm->rgn[i], sizeof(fdm->rgn[i]));
> +
> +        next_section_addr += sizeof(fdm->rgn[i]);
> +
> +        data_type = be16_to_cpu(fdm->rgn[i].source_data_type);
> +        src_addr  = be64_to_cpu(fdm->rgn[i].source_address);
> +        src_len   = be64_to_cpu(fdm->rgn[i].source_len);
> +        dest_addr = be64_to_cpu(fdm->rgn[i].destination_address);
> +
> +        /* Reset error_flags & bytes_dumped for now */
> +        fdm->rgn[i].error_flags = 0;
> +        fdm->rgn[i].bytes_dumped = 0;
> +
> +        if (be32_to_cpu(fdm->rgn[i].request_flag) != FADUMP_REQUEST_FLAG) {
> +            qemu_log_mask(LOG_UNIMP,
> +                "FADUMP: Skipping copying region as not requested\n");
> +            continue;
> +        }
> +
> +        switch (data_type) {
> +        case FADUMP_CPU_STATE_DATA:
> +            /* TODO: Add CPU state data */
> +            break;
> +        case FADUMP_HPTE_REGION:
> +            /* TODO: Add hpte state data */
> +            break;
> +        case FADUMP_REAL_MODE_REGION:
> +        case FADUMP_PARAM_AREA:
> +            /* Skip copy if source and destination are same (eg. param area) */
> +            if (src_addr != dest_addr) {
> +                copy_buffer = g_malloc(src_len + 1);
> +                if (copy_buffer == NULL) {
> +                    qemu_log_mask(LOG_GUEST_ERROR,
> +                        "FADUMP: Failed allocating memory for copying reserved memory regions\n");
> +                    fdm->rgn[i].error_flags =
> +                        cpu_to_be16(FADUMP_ERROR_LENGTH_EXCEEDS_SOURCE);
> +
> +                    continue;
> +                }
> +
> +                /* Copy the source region to destination */
> +                cpu_physical_memory_read(src_addr, copy_buffer, src_len);
> +                cpu_physical_memory_write(dest_addr, copy_buffer, src_len);
> +                g_free(copy_buffer);
> +            }
> +
> +            /*
> +             * Considering cpu_physical_memory_write would have copied the
> +             * complete region
> +             */
> +            fdm->rgn[i].bytes_dumped = cpu_to_be64(src_len);

Is this really valid for FADUMP_PARAM_AREA where we intend to skip copy?

> +
> +            break;
> +        default:
> +            qemu_log_mask(LOG_GUEST_ERROR,
> +                "FADUMP: Skipping unknown source data type: %d\n", data_type);
> +
> +            fdm->rgn[i].error_flags =
> +                cpu_to_be16(FADUMP_ERROR_INVALID_DATA_TYPE);
> +        }
> +    }
> +
> +    return true;
> +}
> +
>   static void trigger_fadump_boot(target_ulong spapr_retcode)
>   {
>       /*
> @@ -353,7 +467,8 @@ static void trigger_fadump_boot(target_ulong spapr_retcode)
>        */
>       pause_all_vcpus();
>   
> -    if (true /* TODO: Preserve memory registered for fadump */) {
> +    /* Preserve the memory locations registered for fadump */
> +    if (!fadump_preserve_mem()) {

This change can be avoided as suggested in previous patch.

>           /* Failed to preserve the registered memory regions */
>           rtas_st(spapr_retcode, 0, RTAS_OUT_HW_ERROR);
>   
> diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
> index efa2f891a8a7..a80704187583 100644
> --- a/include/hw/ppc/spapr.h
> +++ b/include/hw/ppc/spapr.h
> @@ -776,7 +776,32 @@ void push_sregs_to_kvm_pr(SpaprMachineState *spapr);
>   #define FADUMP_CMD_UNREGISTER          2
>   #define FADUMP_CMD_INVALIDATE          3
>   
> -#define FADUMP_VERSION    1
> +#define FADUMP_VERSION                 1

This change can be avoided if taken care initially.

Thanks
Harsh
> +
> +/*
> + * The Firmware Assisted Dump Memory structure supports a maximum of 10 sections
> + * in the dump memory structure. Presently, three sections are used for
> + * CPU state data, HPTE & Parameters area, while the remaining seven sections
> + * can be used for boot memory regions.
> + */
> +#define FADUMP_MAX_SECTIONS            10
> +#define RTAS_FADUMP_MAX_BOOT_MEM_REGS  7
> +
> +/* Firmware provided dump sections */
> +#define FADUMP_CPU_STATE_DATA   0x0001
> +#define FADUMP_HPTE_REGION      0x0002
> +#define FADUMP_REAL_MODE_REGION 0x0011
> +
> +/* OS defined sections */
> +#define FADUMP_PARAM_AREA       0x0100
> +
> +/* Dump request flag */
> +#define FADUMP_REQUEST_FLAG     0x00000001
> +
> +/* Dump status flag */
> +#define FADUMP_ERROR_INVALID_DATA_TYPE          0x8000
> +#define FADUMP_ERROR_INVALID_SOURCE_ADDR        0x4000
> +#define FADUMP_ERROR_LENGTH_EXCEEDS_SOURCE      0x2000
>   
>   /*
>    * The Firmware Assisted Dump Memory structure supports a maximum of 10 sections


  reply	other threads:[~2025-03-05  6:41 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-17  7:17 [PATCH 0/6] Implement Firmware Assisted Dump for PSeries Aditya Gupta
2025-02-17  7:17 ` [PATCH 1/6] hw/ppc: Implement skeleton code for fadump in PSeries Aditya Gupta
2025-02-27  3:07   ` Nicholas Piggin
2025-02-27  6:49     ` Aditya Gupta
2025-02-27  8:48       ` Nicholas Piggin
2025-02-27 12:15         ` Aditya Gupta
2025-03-04  9:01   ` Harsh Prateek Bora
2025-03-06  4:08     ` Aditya Gupta
2025-02-17  7:17 ` [PATCH 2/6] hw/ppc: Trigger Fadump boot if fadump is registered Aditya Gupta
2025-02-27  3:14   ` Nicholas Piggin
2025-02-27  6:56     ` Aditya Gupta
2025-03-04  9:21   ` Harsh Prateek Bora
2025-03-06  4:11     ` Aditya Gupta
2025-02-17  7:17 ` [PATCH 3/6] hw/ppc: Preserve memory regions registered for fadump Aditya Gupta
2025-03-05  6:40   ` Harsh Prateek Bora [this message]
2025-03-06  4:16     ` Aditya Gupta
2025-02-17  7:17 ` [PATCH 4/6] hw/ppc: Implement saving CPU state in Fadump Aditya Gupta
2025-02-27  3:27   ` Nicholas Piggin
2025-02-27  7:01     ` Aditya Gupta
2025-03-05  7:23   ` Harsh Prateek Bora
2025-03-06  4:22     ` Aditya Gupta
2025-02-17  7:17 ` [PATCH 5/6] hw/ppc: Pass device tree properties for Fadump Aditya Gupta
2025-02-27  3:28   ` Nicholas Piggin
2025-02-27  7:02     ` Aditya Gupta
2025-03-05  7:34     ` Harsh Prateek Bora
2025-02-17  7:17 ` [PATCH 6/6] hw/ppc: Enable Fadump for PSeries Aditya Gupta
2025-02-27  3:33   ` Nicholas Piggin
2025-02-27  7:07     ` Aditya Gupta
2025-02-27  8:52       ` Nicholas Piggin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f5494f6f-599f-427b-8c37-42cc5396c35d@linux.ibm.com \
    --to=harshpb@linux.ibm.com \
    --cc=adityag@linux.ibm.com \
    --cc=danielhb413@gmail.com \
    --cc=hbathini@linux.ibm.com \
    --cc=mahesh@linux.ibm.com \
    --cc=npiggin@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    --cc=sourabhjain@linux.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).